Ultrasoungraphy of The Eye

Ultrasonography of the Eye and Orbit
Authors
Front Matter
Preface
Dedication
Acknowledgments
1 - Physics of Ultrasound
2 - Ultrasonic Systems
3 - Ocular Diagnosis
4 - Very High Frequency Digital Ultrasound Scanning in LASIK and Phakic Intraocular Lenses
5 - Orbital Diagnosis
Appendices
Color Plates
Authors: Coleman, D. Jackson; Silverman, Ronald H.; Lizzi, Frederic L.; Lloyd, Harriet; Rondeau, Mark J.; Reinstein, Dan Z.; Daly, Suzanne W.
Title: Ultrasonography of the Eye and Orbit, 2nd Edition
Copyright ©2006 Lippincott Williams & Wilkins
> Front of Book > Authors
Authors
D. Jackson Coleman MD, FACS

The John Milton McLean Professor of Ophthalmology
Director, Margaret M. Dyson Vision
Research Institute
Chairman, Department of Ophthalmology
Weill Medical College of Cornell University
New York Presbyterian Hospital
Senior Research Physician
Riverside Research Institute
New York, New York
Ronald H. Silverman PhD

Professor of Computer Science in Ophthalmology
Research Director, Bioacoustic Research Facility
Margaret M. Dyson Vision Research Institute
Department of Ophthalmology
Member of Research Staff
Biomedical Engineering Directorate
New York, New York
Frederic L. Lizzi EngScD*

Research Director
Biomedical Engineering Directorate
Adjunct Professor of Ophthalmic Physics in
Ophthalmology
New York, New York
*Deceased
Harriet Lloyd MS
Research Associate in Ophthalmology
New York, New York
Mark J. Rondeau
Research Associate in Ophthalmology
Associate Director, Bioacoustic Research Facility
Margaret M. Dyson Vision Research Institute
New York, New York
Dan Z. Reinstein MD, FRCSC, DABO

Clinical Assistant Professor of Ophthalmology
New York, New York
Medical Director
London Vision Clinic
London, UK
Suzanne W. Daly BSN, RDMS, CRNO

Senior Lecturer in Ophthalmology
New York, New York
P.xii
CONTRIBUTORS FOR DVD
Walter Cronkite
Introduction
C. P. Wilkinson MD
Professor and Chairman
Department of Ophthalmology Greater Baltimore Medical Center
Vitreous Hemorrhage
Mario Stirpe MD
GB Bietti Eye Foundation
Subhyaloid Hemorrhage
Mark Blumenkranz MD
Professor and Chairman
Department of Ophthalmology Stanford University
Vitreous Membrane
George Blankenship MD
Vitreo-retinal Surgeon Hershey, PA
Retinal Detachment
Charles Pavlin MD
Professor
University of Toronto
Endophthalmitis
Donald J. D'Amico MD
Professor of Ophthalmology
Director of Diabetic Retinopathy Unit Harvard Medical School
Retinal Detachment
Stanley Chang MD
Edward S. Harkness Professor of Ophthalmology Chairman of the Department of Ophthalmology Columbia Presbyterian Medical Center
Residual Perfluorocarbon Bubbles
Evangelos Gragoudas MD
Director of Retina Service Massachusetts Eye and Ear Infirmary
Choroidal Melanoma
H. Culver Boldt MD
Department of Ophthalmology and Visual Sciences University of Iowa
Ocular Melanoma
Thomas C. Lee MD
Associate Professor of Ophthalmology Weill Medical College of Cornell University
Subretinal Hemorrhage
Yale Fisher MD
Director of the Surgical Retinal Service Manhattan Eye, Ear & Throat Hospital
Orbital Cyst
> Front of Book > Front Matter
Front Matter
Four things come not back: the spoken word, the spent arrow, time past, the neglected opportunity.
Omar Ibn Al-Halif

> Front of Book > Preface
Preface
It has been nearly a score and 10 years since we prepared the first edition of this book. These past 30 years have seen the development of power spectrum
analysis, 3-D scans, very high frequency or UBM, arc scans, wavelets, Doppler and digital processing, and swept scans. On the horizon are contrast agents, and
linear and phased transducer arrays as well as computational power that will further facilitate ultrasonic diagnosis of the eye. The past is merely an adumbration of
future developments.
We are saddened that our long-time friend and colleague, Fred Lizzi, is no longer with us and will not join us in the next edition. It is possible, however, that the
physics will not change and that Chapter 1 might remain the same. In any event, treasure this chapter. He was a genius, with humor and insight.
Our ultrasound group here at Weill Cornell has worked together for a generation—researching, collecting data, and trying to improve the diagnostic millieux. In this
second edition, we have retained the introduction of basic physical principles, which remains unchanged, but have attempted to illuminate the technological
advances that provide improved definition, resolution, and diagnostic capability to ultrasound imaging. We have not attempted, as we did in the first edition, to
provide complete references of all ophthalmic ultrasound publications. Instead, we acknowledge the work of many other investigators. Ultrasound is clearly an
established and vital part of the ophthalmic diagnostic armamentarium. In the ocular diagnosis section, we have stressed the advantages of multi-frequency and
digital processing techniques. In the orbital section, we have stressed the complementary nature of ultrasound and other imaging modalities, such as computed
tomography and magnetic resonance, as these imaging techniques have also improved.
We have added a DVD to this edition to stress the real-time nature of ultrasound diagnosis, and have asked outstanding surgeon colleagues to describe some of
the features best seen in real time.
We thank the National Institutes of Health for supporting our research, The Dyson Foundation, the St. Giles Foundation, the Whitaker Foundation, and Research to
Prevent Blindness for their trust and support.
And above all—we thank our families for letting us work for love….
D. Jackson Coleman
> Front of Book > Dedication
Dedication
This book is dedicated to
Frederic L. Lizzi, EngScD
1942-2005
Our treasured colleague for 38 years. A brilliant biophysicist, innovator, and researcher. He played a major role in our ultrasound research and will be sorely
missed.
> Front of Book > Acknowledgments
Acknowledgments
We are grateful to numerous colleagues for their referral of patients and for allowing us to reprint some of their images, such as the foreign bodies of unusual
character.
Harriet Lloyd deserves very special thanks for her research skills and particularly for her help in writing, as well as editing and preparing the entire manuscript.
We thank Sue Daly for her hours of scan time and recovery of patient data.
We thank George Simoni, BSEE, who has provided innumerable hours of engineering expertise and many innovative and critical ideas for our research.
We thank our colleagues at Riverside Research Institute for their help with our research and the inestimable resource their Biomedical Staff provide.
We would like to thank the many medical students, residents, and fellows who have provided enthusiasm that keeps our work exciting.
We would like to thank Lisa Kairis and Jonathan Pine of Lippincott Williams & Wilkins, and Rebecca Dodson and Bridget Nelson of Schawk Publishing Solutions,
whose skill, assistance, and encouragement have brought this book to publication.
> Table of Contents > 1 - Physics of Ultrasound
1
Physics of Ultrasound
Since its first ocular application (1) in 1956, ultrasound has had a broad impact on the practice of ophthalmology. It is now a standard clinical modality for
measuring ocular dimensions, diagnosing and monitoring ocular diseases, and providing information regarding orbital diseases. Modern ultrasound systems
provide real-time, highly detailed images of ocular structures in a rapid, noninvasive manner, posing no significant threat of tissue damage. Ultrasonic biometry
quantifies ocular dimensions needed to plan and evaluate sight restoration and improvement by intraocular lens implants and corneal surgery. Real-time ultrasound
images, unaffected by optical opacities, have significantly advanced the diagnosis and management of virtually all ocular diseases and abnormalities. Ultrasonic
imaging of orbital disease and blood-flow patterns complements data obtained by other imaging modalities, such as magnetic resonance imaging (MRI).
The effective use of ophthalmic ultrasound requires a basic knowledge of its physical nature and the phenomena associated with its propagation and scattering.
This understanding is important for proper interpretation of clinical results and avoidance of misleading artifacts that can arise in ocular examinations. It is also
important for evaluating emerging techniques that promise to extend the scope of ultrasonic examinations in the future as well as to best use other techniques for
complementary diagnostic value.
Ultrasound is an acoustic wave comprising compressions and rarefactions that propagate within fluid and solid substances (2, 3, 4). By definition, ultrasonic waves
exhibit frequencies above 20 kHz,1 and they differ from sound waves because these high frequencies render them inaudible. Because it is a wave, ultrasound can
be directed, focused, and reflected according to the same general principles that govern these phenomena with other waves, such as light. The high frequencies
(typically 10 MHz) and small wavelengths (e.g., 150 µm) available with ultrasound can provide the detailed resolution required for ocular examinations. Newer
techniques use even higher frequencies (e.g., 50 MHz) to obtain wavelengths near 30 µm for very fine resolution within the anterior chamber (5,6).
Ultrasonic examinations of soft tissues use reflective (“pulse-echo”) systems analogous to those used in radar and sonar. This approach allows examination within
a thin “slice” through tissue structures. A piezoelectric transducer serves as the ultrasonic transmitter and receiver. It generates a short burst of ultrasonic energy
that propagates through the eye and undergoes partial reflection at tissue boundaries that exhibit abrupt changes in mechanical properties, including density and
rigidity. These reflections, or echoes, return to the transducer where they are electronically detected. A-mode, or A-scan, systems graphically display these echoes
as a function of time on a video monitor. B-mode systems generate cross-sectional gray-scale images (the gray scale corresponds to the A-scan amplitude) by
scanning the transducer to address a series of lines through the eye; the amplitudes of received echoes control the brightness (or gray scale) along corresponding
lines of a video image (B-scan). The terms A-scan and B-scan as well as C-scan and M-scan derive from early radar terms, using pulse position indicator (PPI)
display.
Subsequent chapters describe how these A- and B-mode results are interpreted for diagnostic purposes. Proper interpretation requires an understanding of how A-
and B-mode signals are related to underlying tissue properties and how they are affected by the characteristics of the ultrasonic system and transducer. The
principles involved with ultrasonic imaging differ from those encountered with other imaging modalities. Computed
P.2
tomography (CT) measures the partial absorption of xradiation transmitted through the body, and MRI senses molecular phenomena elicited within tissue. Optical
coherence tomography (OCT) senses light that is backscattered by local changes in optical refractive indices rather than mechanical properties (7). OCT can
produce high-resolution cross-sectional images of ocular tissues, such as the retina, but, as with other optical techniques, its depth of penetration is limited by
optically opaque media, such as the sclera and intravitreal substances.
The physical principles of ultrasonic imaging are reviewed in this chapter, which describes how ultrasound is generated and detected, how it is reflected and
absorbed in tissue, and how various factors influence the resolution that can be achieved in examining the eye and orbit. (References 2, 3, 4 are comprehensive
texts treating the physics of ultrasound.)
GENERATION AND DETECTION OF ULTRASOUND

The key element in any ultrasonic system is a piezoelectric transducer, which is used to generate an ultrasonic wave from an applied voltage signal and to detect
ultrasonic echoes returning from within the eye. A typical transducer unit (Figure 1.1) consists of a thin disk of piezoelectric material, such as lead zirconate titanate
(PZT), a backing section, and an acoustic lens, which focuses the generated ultrasonic beam. The entire unit is commonly referred to as the transducer, although
this term applies most correctly to only the piezoelectric element; common usage is adopted in this text. In most clinical systems, the transducer is coated with a
thin layer of coupling gel and held in contact with the globe or lid. The gel affords a transmission path for ultrasound, which is rapidly absorbed in air. Coupling can
also be provided by fluid solutions confined in small chambers or surgical drape.
Generation and detection of ultrasound take place in the piezoelectric material. The molecular configuration of a simple piezoelectric crystal is shown schematically
in Figure 1.2. The molecules exhibit net charge polarizations that are forced into alignment by the crystalline structure so that effective positive charge centers are
oriented along the same direction. In the transmission mode, an ultrasonic pulse is generated by applying a voltage pulse across external electrodes plated on the
crystal surfaces. The molecules tend to stretch or contract, depending on whether the voltage polarity causes attraction or repulsion of the charge centers. These
molecular effects alter the overall crystal thickness in proportion to the amplitude of the applied voltage. When the polarity of the applied voltage is rapidly varied,
the crystal executes corresponding rapid expansions and contractions, which constitute ultrasonic vibrations.
Figure 1.1. Cutaway view of transducer.
In the receive mode, the crystal is compressed and expanded by an impinging ultrasonic echo pulse; the concomitant changes in molecular charge separation
induce an output voltage whose amplitude and waveform depend upon the echo pulse. The voltage is readily measured as a function of time, enabling ultrasonic
echoes to be detected as they return from the eye.
In the past, piezoelectric transducers were often fabricated from precisely oriented cuts of quartz crystals, which require large excitation voltages. Now, most
transducers are fabricated from more sensitive materials, including lithium sulfate, ceramics (such as PZT), composite materials, and, for high frequencies,
polyvinylidene fluoride (PVDF) membranes (8). Modern transducer materials can detect small ultrasonic signals, containing only microwatts of power. Some of
these materials must be “poled” before they can be used in transducers. In this process, piezoelectric domains are brought into alignment by applying large,
constant voltages at elevated temperatures. Once this alignment is achieved, these materials can generate and detect ultrasound in the manner described
previously.
A piezoelectric transducer responds most actively to voltage signals and ultrasonic pulses that have frequencies near its resonant frequency. This frequency is
determined by the material's thickness, increasing as it is made thinner. Resonance effects can lead to prolonged series of ultrasonic vibrations, which are
suppressed by using backing sections to achieve high resolution, as discussed in a subsequent section.
P.3
Figure 1.2. Schematic representations of molecular configuration in a piezoelectric material illustrating

contraction induced by an applied voltage.
PROPAGATION OF ULTRASOUND
When a piezoelectric transducer is immersed in a fluid and electrically excited, its thickness vibrations generate an ultrasonic wave of compression and rarefaction
that propagates through the fluid. These waves, termed longitudinal or compressional ultrasonic waves, are the type used for tissue visualization. They propagate
through soft tissues in the same manner as they propagate through fluids.
Figure 1.3. Generation and propagation of a compressional ultrasonic wave. The wave is generated by a small
extension of a transducer surface into a fluid.
Ultrasonic propagation is illustrated in Figure 1.3, where a voltage pulse causes a piezoelectric transducer
P.4
to undergo a small, rapid expansion. Extension of the front transducer surface initially compresses the adjacent fluid layer, elevating its density and pressure.
Increased molecular collisions in this compressed region eventually couple the elevated density and pressure to the next fluid layer, while the initially compressed
region returns to its original state. Thus, the compression passes from the first layer to a second region and, in the same manner, continually propagates to more
distant regions in the fluid. Similar phenomena occur when the transducer contracts rather than expands. In this case, rarefaction characterized by lowered fluid
pressure and density propagates away from the transducer.
Induced compression and rarefaction disturbances travel through a substance at a speed (velocity of propagation) that is determined by the density and
compressibility of that substance. In materials with low compressibilities, such as metals, compression passes rapidly from layer to layer, and large propagation
velocities are encountered (e.g., 6,000 m/sec). In contrast, materials that are more readily compressed, such as fluids and tissues, exhibit lower velocities (e.g.,
1,524 m/sec in water). As shown in Table 1.1, ocular tissues exhibit propagation velocities close to those of water (9, 10, 11, 12, 13, 14, 15, 16, 17). The largest
velocity is exhibited by the lens. Propagation velocities are temperature-dependent (18). Near 37°C, a 1°C-temperature rise typically increases velocities by about 1
to 2 m/sec, except in fat where an opposite trend occurs.
TABLE 1.1 Reported Mean Velocities of Ultrasound in Ocular Tissues

Tissue Accepted Velocity Velocity (m/sec) Temperature (°C) Frequency (MHz) Investigator (Reference)
Cornea 1,639 m/sec 1,632 22 4 Chivers9
1,550 22 4 Oksala10
1,553 22 10 Thijssen11
1,572 20 20 De Korte13
1,575 37 60 Ye14
Sclera 1,744 22 4 Chivers9
1,630 22 4 De Oksala10
1,583 22 10 Thijssen11
1,597 20 20 Korte13
1,622 37 60 Ye14
Vitreous 1,532 m/sec 1,508 22 4 Chivers9
1,495 22 4 Oksala10
1,532 37 4 Jannson12
1,506 22 10 Thijssen11
1,514 20 20 De Korte13
Lens 1,641 m/sec 1,548 22 4 Chivers9
1,650 22 4 Oksala10
1,641 37 4 Jannson12
1,620 22 10 Thijssen11
1,659 37 15 Coleman15
1,590 20 20 De Korte13
Cataractous lens 1,629 37 — Coleman15
Water 1,524 37 — Willard 16
Aqueous humor 1,532 37 — Jannson12
Fat 1,476 24 — Frucht17

Fat 1,476 24 — Frucht17
In medical systems, brief excitation voltages are used, and the transducer surface vibrates back and forth several times at a rate equal to its resonant frequency
(e.g., 10 MHz). This series of vibrations generates several contiguous regions of compression and rarefaction that propagate with the previously mentioned
velocity, as shown in Figure 1.4. These regions travel together as an ultrasonic pulse and cause, roughly, sinusoidal variations in density and pressure as they
traverse the eye. Clinical systems use only small transducer motions (total excursions under one micron), and imperceptible and innocuous pressure variations are
produced within the eye and orbit.
Sinusoidal ultrasound pulses manifest a wavelength that is an important determinant of many operational parameters, including resolution. The wavelength, ?, is
the spatial distance over which the pressure perturbation undergoes one complete cycle. Wavelength is determined by the frequency, f, of transducer vibrations
and the propagation velocity, c, of the medium:
? = c/f
P.5
Figure 1.4. Propagation of sinusoidal ultrasonic pulse in a fluid medium at two instants of time. Lower plot
shows ultrasonic pressure variation of peak amplitude p. Static pressure level is exhibited outside the region
occupied by the pulse.
This relation is obeyed because the transducer generates the same pressure level every 1/f seconds, and this pressure travels at a velocity equal to c. In water,
10-MHz operation produces a wavelength of 0.15 mm, which is commensurate with retinal thickness; increasing the frequency to 50 MHz decreases the
wavelength to 0.03 mm, and thus increases resolution.
The wavelength describes the spatial distribution of pressure at a single instant of time. However, ultrasonic pulses continually propagate through tissue structures
at high velocities, causing rapid pressure oscillations as they travel through points within the eye. In fact, as shown in Figure 1.5, the pressure at a point in the eye
varies at the same high rate (e.g., 10 MHz) at which the transducer vibrates.
Nonlinear propagation effects, which distort ultrasonic pulses, can become significant at elevated ultrasonic pressure amplitudes (19). These effects occur because
the acoustic propagation velocity is inversely proportional to density. Thus, the high pressure (high density) regions in Figure 1.4 actually travel at a somewhat
slower speed than the low pressure (low density) regions of the pulse. The magnitude of this effect depends upon the nonlinearity property (“B/A” parameter) of
tissue. As the pulse propagates, this speed differential acts in a cumulative manner and can distort the pulse by shortening
P.6
its compression regions and lengthening its rarefaction regions.
Figure 1.5. Pressure variations caused as ultrasonic pulse passes through point A in Figure 1.4. The frequency
of these variations is the same as that of the transducer vibrations.
Figure 1.6. Generation and propagation of shear ultrasonic wave in a solid. The wave is generated by a
shearing force at the transducer surface. Internal displacements are perpendicular to direction of propagation.
Thus far, longitudinal, or compressional, ultrasonic waves have been discussed. Several other types of ultrasonic waves can be generated in certain materials but
are not important in ophthalmic examinations. Surface (Rayleigh) waves and shear (transverse) waves are examples of these. Shear waves are excited in solids
when a transducer surface vibrates within one plane (Figure 1.6). This motion generates shear forces that are transmitted to progressively farther regions within the
solid; the induced particle motion is perpendicular to the propagation direction. Shear waves have not been used for ocular visualization because they do not
couple well into tissue and because they are rapidly dissipated by viscosity.
REFLECTION AND SCATTERING OF ULTRASOUND

Ultrasonic pulses are reflected at boundaries between media that possess differing mechanical characteristics. Figure 1.7 illustrates the extreme case of total
reflection from a rigid planar structure bounding a fluid of depth L. When the incident compression reaches the structure, the expansive forces accompanying
molecular collisions are redirected back into the fluid, and the phenomena described previously cause the pulse to travel through the fluid in the reverse direction.
The reflected pulse arrives back at the transducer after a time interval equal to 2 L/c, and it generates a corresponding output voltage, as shown in the figure.
Observation of this voltage enables the boundary to be detected and permits L (distance) to be calculated if c (velocity) is known.
In the eye, a similar reflection arises whenever a pulse encounters a boundary between most ocular structures. However, ocular tissues exhibit similar mechanical
properties so that only a small fraction of the incident pulse is reflected. Most of the incident energy is transmitted through the boundary, where it undergoes partial
reflections at each successive interface.
The major reflective surfaces in the normal eye are those of the cornea, lens, and rear wall layers. Figure 1.8 shows a schematic diagram of these surfaces with
typical dimensions and also illustrates the transducer echo voltages that arise from each of them. Corresponding clinical results are shown in Figure 1.9. This figure
presents the radiofrequency (RF) echoes measured directly at the transducer as well as their envelopes, or video signals, that are used to modulate brightness or
gray-scale levels in B-mode images.
The time interval between echo voltage pulses can be used to determine the thickness of the corresponding
P.7
tissue segment because the pertinent propagation velocities are known. Specifically, the time interval between successive echoes is equal to 2 L/c, where L is the
corresponding tissue thickness. Thus, typical lens echoes are separated by 5 µsec (microsecond, which is 10-6 seconds; nanosecond [nsec] denotes 10-9 seconds)
because L = 4 mm (a typical lens thickness) and c = 1,641 m/sec. Biometric systems using this approach have achieved a precision (reproducibility) of ± 20 µm in
axial length determinations, and 40-MHz systems have achieved a precision of better than ± 2 µm in measuring the thickness of the corneal epithelium.
Figure 1.7. Total reflection of an ultrasonic pulse from a rigid boundary. The transducer output voltage displays
both the excitation pulse and the echo pulse so that the total transit time can be measured.
The amplitudes of ultrasonic reflections depend upon the characteristic acoustic impedances, Z, of adjacent tissues. The characteristic impedance of a tissue is
equal to the product of its density, p, and propagation velocity:
Z=pc
The pressure reflection coefficient, Rp, is defined as the ratio of the reflected pressure amplitude to that of the incident pressure amplitude. For smooth tissue
interfaces that are perpendicular to the ultrasound beam, Rp is equal to:
where the subscripts refer to the first and second tissue structures (Figure 1.10).
The reflection coefficient is an important characteristic of a tissue boundary; it affects the echo amplitude in A-mode displays and brightness in gray-scale B-mode
images. Because Rp depends on Z2-Z1, its value depends on the acoustic impedances of tissues on both sides of the boundary. Thus, the reflection from a foreign
body (large Z) will be smaller, if it is situated in a dense blood clot (moderate Z), than if it is situated in the vitreous humor (small Z). Within homogeneous
structures, such as the normal vitreous, lens, or even the optic nerve, Z is constant and no reflections arise. In heterogeneous structures (e.g., cataractous lens) Z
can vary from point to point, producing numerous, closely spaced echoes.
P.8
Figure 1.8. Schematic representation of interfaces presented by cornea (C), aqueous humor (A), lens (L),
vitreous humor (V), and retina (R). Transducer voltage demonstrates echoes from anterior and posterior corneal
surfaces (AC and PC), lens surfaces (AL and PL), and retina. Measured time intervals can be used with velocity
data to determine the thickness of each ocular segment.
Figure 1.9. Echo voltages obtained from eye; upper trace shows RF echoes corresponding to those illustrated
in Figure 1.8. Lower trace shows corresponding video waveforms described in Chapter 2. Following ocular
echoes, a complex echo pattern arises from scattering within orbital fat.
P.9
Figure 1.10. Ray diagram showing paths of incident, reflected, and transmitted ultrasonic pulses at a tissue
interface under normal incidence. The characteristic acoustic impedances of both tissues determine the
relations between the amplitudes of these pulses.
Most normal and pathologic structures give rise to small reflections. Normal ocular reflection coefficients range from about 7% at the lens-aqueous humor interface
to approximately 1% at the chorioretinal interface (20). Reflection coefficients exceeding 10% are encountered with rigid foreign bodies. Negative values of Rp
occur at interfaces, such as the cornea-aqueous boundary, where Z1 (cornea) is larger than Z2 (aqueous). Physically, this implies that the compressive component
of the incident wave is reflected as a rarefactive component and vice versa. This acoustic reversal inverts the corresponding RF echo signal but does not alter the
video signal (which is the envelope of the rectified RF signal).
Figure 1.11. Reflections from planar interface at normal (left) and oblique (right) incidence. Oblique incidence
results in lowered echo amplitude and increased duration as a result of the variation in transit times along
different rays.
Several factors modify the echo amplitudes observed in practice. If an ultrasonic beam impinges on a tissue boundary at an oblique angle, the reflected pressure
received at the transducer is less than that described previously. The reflection coefficient, in this case, depends upon the angle of incidence, surface roughness,
and ultrasonic wavelength. As illustrated in Figures 1.11 and 1.12, the echo amplitudes from rough surfaces vary more slowly with angle of incidence than do those
arising from smooth (specular) surfaces, such as those of the cornea and lens. Echo strength can also change, if the tissue boundary is sharply curved because of
focusing and defocusing phenomena, as illustrated in Figure 1.13. Finally, apparent reflectivities are modified because of ultrasonic attenuation in intervening
tissues; as discussed later, attenuation arises from absorption and other factors that diminish the amplitudes of incident and reflected ultrasonic pulses.
Another important source of ultrasonic echoes is scattering, which arises from small, closely spaced reflective surfaces within the ultrasound beam. Scattering
occurs from many tissues, including internal tumor structures (small blood vessels, calcific deposits, cellular aggregates, and so forth), ciliary body parenchyma,
and connective septae in orbital fat. Scattering from distributions of such elements gives rise to complex echo voltages, in which the contributions from many
individual small scatterers are
P.10
superimposed, as shown in the schematic representation of Figure 1.14. The overall echo patterns are determined by a variety of factors, including geometric
scatterer properties (size, shape, and orientation), the spatial distribution of scatterers, acoustic impedances of tissue constituents, absorption, and ultrasonic
wavelength. When scattering occurs, diagnostically useful information is often obtained by examining the “texture” of echo patterns and corresponding B-mode
images and by assessing the rate at which attenuation reduces echo amplitudes with increasing depth. Spectrum analysis techniques have been used to
quantitatively analyze these echo patterns and relate their properties to underlying tissue morphology for tissue identification (21,22).
Figure 1.12. Reflection from rough interface. Surface roughness redirects energy in a variety of directions,
causing decreased echo amplitude and increased duration. Oblique incidence does not affect echo amplitude to
the extent encountered with smooth surfaces.
Figure 1-13. Reflection from curved surface. Beam spreading upon reflection reduces echo amplitudes.
Ultrasonic contrast agents exploit a special case of scattering that occurs when gas bubbles are exposed to ultrasound (23). Contrast agents commonly consist of
encapsulated gas particles several micrometers in diameter. These are usually injected into the venous circulation to increase the scattering from blood and
thereby facilitate detection, tracking, and analysis of blood flow in vascular components. The enhanced scattering from these circulating particles is a result of the
large difference in compressibility between gas and fluids. Scattering can become large when the incident ultrasonic frequency is equal to the resonant frequency
of the bubbles. In water, the resonant frequency of a 1-µm radius bubble is approximately 3 MHz; this frequency increases as the bubble radius decreases.
ATTENUATION OF ULTRASOUND
As an ultrasonic wave propagates through any medium, its energy is progressively attenuated through scattering and absorption. Scattering redirects incident
energy, reducing the levels that reach distal tissues. Absorption
P.11
converts acoustic energy to heat. It arises from factors, such as viscosity, which prevent the density of a medium from responding instantaneously to ultrasonic
pressure variations. In biologic media, absorption arises from many complex cellular and molecular phenomena that are not completely characterized. Temperature
rises as a result of absorptive heating are not significant at the low power levels used in diagnostic systems.
Figure 1.14. Scattering from distributed small inhomogeneities. The echo voltage consists of the superposition
of returns from many scatterers.
The attenuation of ultrasound owing to absorption and scattering is important because it significantly reduces echo amplitudes at deep tissue sites. Because
attenuation increases with increasing depth and frequency, it not only constrains the maximum tissue depth that can be examined, but also limits the highest
frequencies that can be used.
Attenuation increases exponentially with the distance traversed by an ultrasonic wave; as the wave travels a distance x, its pressure amplitude decreases by a
factor e-?x, where ? is the pressure attenuation coefficient of the medium. Attenuation losses are usually denoted in terms of decibels (dB), which are computed as
20 log e-?x. Because ? increases in an approximately linear manner with frequency, it is common to specify attenuation coefficients in terms of dB/cm-MHz. An
attenuation coefficient of 0.5 dB/cm-MHz means that at 10 MHz there will be a loss of 5 dB/cm, corresponding to a 44% pressure reduction in 1 cm of tissue depth.
Over a distance of 2 cm, absorption would lead to a 10-dB loss, corresponding to a pressure reduction of about 70%. These fractional reductions are experienced
by both incident and reflected pulses.
Materials vary widely in their attenuation coefficients. Usually, materials with high propagation velocities (e.g., metals) exhibit low attenuation coefficients, whereas
materials with low velocities have high coefficients. Air has a large coefficient that cannot be used to transmit ultrasonic waves over cm distances at frequencies in
the megahertz range. In the eye, reported attenuation coefficients range from 0.1 dB/cm-MHz in the vitreous humor to 2 dB/cm-MHz in the lens (24,25). More
recent measurements, applicable at 20 MHz, have also been reported (26).
REFRACTION OF ULTRASOUND
An ultrasonic wave is redirected (refracted) whenever it obliquely traverses boundaries between media with different propagation velocities. Refraction is used as
an advantage in acoustic lenses. However, it also leads to defocusing and shifting of ultrasonic beams as they encounter curved tissue surfaces within the eye.
Figure 1.15 illustrates the reflection and refraction that result when an ultrasonic pulse encounters a plane interface at an angle ?, with respect to its normal. In this
situation, a reflected pulse is generated at an angle ?, whereas a transmitted pulse propagates into the second medium at an angle ~, which is determined from
Snell's law:
where c1 and c2 are the propagation velocities in the first and second media, respectively. The physical basis for refraction is illustrated in Figure 1.16, which shows
positions of an ultrasonic pulse at sequential instants of time separated by 1 µsec. In this example, c2 is less than c 1 so that as soon as part of the pulse enters the
second medium it travels a smaller distance during the following 1-µsec interval. At t2, the top portion of the illustrated pulse just encounters the boundary; in the
next microsecond, it will propagate over a relatively small distance, d 2, in the low-velocity medium. On the other hand, during this period, the bottom portion of the
pulse is still propagating in the high-velocity medium and travels a relatively large distance, d1. The difference between d1 and d2 causes the transmitted pulse to be
tilted to the degree noted in Snell's law.
P.12
Figure 1.15. Ray diagram showing reflection and transmission for oblique incidence at boundary between
media with different propagation velocities.
Figure 1.16. Sequential positions of incident and refracted ultrasonic pulses at planar boundary.
P.13
The same type of refractive effects at curved surfaces can cause focusing or defocusing of transmitted pulses. Focusing is accomplished by using a high-velocity
planoconcave lens, as illustrated in Figure 1.17, which again shows the position of a propagating pulse at sequential time intervals. Focusing occurs at the concave
surface because the portion of the pulse emerging from the lens travels more slowly than that portion still within the lens. The emerging pulse lies on concave
contours that tend to converge at the focus of the lens. The focal length, F, is related to the lens radius of curvature, A, by the relation:
Figure 1.17. Refractive effects of curved high-velocity structures. Top: a high-velocity planoconcave lens
focuses a plane ultrasonic wave. Bottom: the ocular lens causes defocusing.
where cL and c refer to the lens and coupling medium propagation velocities, respectively. This relation is the same as that encountered in optics with refractive
indices. (Converging optical lenses use convex surfaces because the propagation velocity of light is lower in lens materials than in air.) The actual focusing of
transducer beams by acoustic lenses is modified by diffraction, which is treated in a subsequent section.
Figure 1.17 also shows how defocusing occurs when an ultrasonic pulse emerges from convex structures, such as the ocular lens, which exhibit high propagation
velocities. When a beam passes through the center of the lens, beam spreading and absorption losses occur. If transmission occurs through peripheral lenticular
segments, refraction leads to a shift in the direction of propagation. These effects can impede examination of tissues posterior to the lens. However, propagation
through the sclera does not alter beam characteristics, unless the beam is almost tangent to the scleral surface (27).
P.14
AXIAL RESOLUTION
The degree to which tissue structures can be resolved with ultrasonic systems is limited in practice by factors such as attenuation and diffraction. This section
discusses how these phenomena affect axial (thickness) resolution. The next section discusses their impact on lateral (width) resolution.
The smallest tissue thickness that can be resolved by an ultrasonic system is termed its axial resolution and is determined by the time duration of the ultrasonic
pulse. Short durations are needed to resolve thin-tissue segments. This fact can be illustrated by considering the conditions needed to resolve the cornea, which is
typically 0.5 mm thick and gives rise to echoes separated by 0.6 µsec. Corneal echoes that would be obtained with three different pulse durations are shown in
Figure 1.18. In the first two cases, the pulse duration is less than 0.6 µsec so that the two corneal echoes can be distinguished, and the cornea is resolved. If the
pulse duration exceeds 0.6 µsec, as in the third case, the echoes overlap, and the corneal surfaces are not resolved.
In general, a pulse duration, T, yields an axial resolution equal to cT/2, where c is the propagation velocity of the relevant tissue structure. It is convenient to
consider an average tissue velocity of 1.5 mm/µsec, resulting in the relation:
axial resolution (mm) = 0.75 T
Figure 1.18. Effects of pulse duration on axial resolution. In the upper two examples, pulse duration is small
enough to resolve both corneal surfaces. In the lower example, the pulse duration is too large for corneal
resolution.
where T is specified in microseconds. This relation provides a convenient basis for determining axial resolution, because T can be determined by measuring the
duration of an echo from a flat test object. For example, if a pulse duration of 0.15 µsec is observed, then tissue surfaces separated by 0.11 mm are resolvable. An
alternative description of axial resolution capabilities specifies the bandwidth B of the transducer; this bandwidth is approximately equal to 1/T so that axial
resolution (mm) is equal to 0.75/B, where B is specified in megahertz. Thus, broadband transducers and electronic systems are required for fine axial resolution.
Conversely, a narrow band transducer system may provide better sensitivity when, for example, visualizing low amplitude echoes in the vitreous body.
Generation of short ultrasonic pulses requires careful transducer design and fabrication. Even with short excitation voltage pulses, ultrasonic pulse durations can
be excessively large because of the “ringing” encountered in undamped transducers. As shown in Figure 1.19, an excited transducer actually generates two
ultrasonic pulses at its front surface. One pulse propagates into the coupling medium, while the other travels in the reverse direction through the transducer. This
internal pulse repeatedly undergoes partial reflection and transmission at the two surfaces of the transducer element. Thus, many individual pulses, staggered in
time, are transmitted into the fluid where they form a composite pulse of long duration. The
P.15
same type of internal reflections occurs during reception of ultrasonic echoes, further lengthening the effective pulse duration.
Figure 1.19. Pulse lengthening as a result of ringing in undamped transducer. Successive multiply reflected
pulses are transmitted sequentially into the coupling medium.
In a damped transducer, internal reflections are suppressed by a backing section as indicated in Figure 1.1. The acoustic impedance of the backing material is
selected to approximate that of the transducer. If these impedances are equal, the reflection coefficient is reduced to 0, eliminating internal reflections. As shown in
Figure 1.20, echoes obtained with an ideal backing have a total duration equal to 1.5/f, where f is the transducer's resonant frequency. Thus, pulse duration can be
minimized by using transducers with high resonant frequencies. Although other electrical and mechanical considerations complicate this simple discussion,
damped transducers with responses close to the ideal case have been fabricated for diagnostic applications (28,29).
Thin matching layers have been used at the front transducer surface to match the acoustic impedance of the transducer to that of tissue; these layers must be
designed for particular frequency ranges. They function in a manner similar to optical anti-reflection coatings and can improve the efficiency of energy transfer into
the coupling medium.
Short pulses commensurate with high-quality axial resolution can be achieved at high frequencies. However, attenuation increases with frequency, and the
attendant diminution of echo amplitude impedes the use of resonant frequencies above 20 MHz for ocular examinations. At this center frequency, the minimum
pulse duration (1.5/f) is 0.075 µsec, and an axial resolution near 0.06 mm is theoretically achievable; in practice, resolution on the order of 0.1 mm has been
achieved. Orbital examinations require deeper tissue penetration, and increased absorption losses hinder the use of frequencies above 10 MHz.
The same design approaches have been used for high frequency transducers (e.g., 40 MHz) fabricated from lead zirconate titanate (PZT) and lithium niobate. In
addition, designers have used materials, such as polyvinylidene fluoride (PVDF), whose low acoustic impedance is closer to that of water and tissues, reducing the
size of internal reflections at their front surface. These transducers are limited to examinations in the anterior chamber because, at 40 MHz, two-way attenuation
accrues at a rate of 40 dB/cm for an attenuation coefficient of 0.5 dB/MHz-cm. The pulse duration attainable with such devices is less than 50 nsec (0.05 µs),
permitting an axial resolution near 30 µm.
P.16
Figure 1.20. Ray diagram and waveforms for ideally backed transducer. During reception, the echo voltage is
lengthened by one-half cycle because of the small transit time through the transducer.
LATERAL RESOLUTION
Detailed examinations of the eye and orbit place stringent requirements not only on axial resolution but also on lateral resolution (sometimes specified in terms of
angular or azimuthal resolution). Lateral resolution is defined by the width of the ultrasonic beam generated by the transducer. As the transducer is scanned, small
beamwidths are needed to measure lateral dimensions accurately to distinguish small objects and to accurately delineate tissue contours.
The importance of lateral resolution is exemplified in the situation where a small reflecting object (e.g., an intraocular foreign body) is being examined. The object
will generate echoes as long as it is situated within the ultrasonic beam, so that its lateral position and size cannot be accurately assessed with a wide beam. In
addition, if several such reflectors are located at the same tissue depth, they cannot be identified as separate entities, unless their lateral spacing exceeds the
beamwidth.
Lateral resolution is also a critical factor for examinations of curved or irregular ocular surfaces. In axial length determinations, for example, a wide ultrasonic beam
will give rise to retinal echoes originating from regions anterior to the point of interest (Figure 1.21). Accordingly, the vitreous chamber appears to have an
erroneously shallow depth: a beamwidth of 1 cm can cause an error of 1 mm. Similar effects can obscure small surface irregularities associated with initial tumor
development and degenerative processes.
Ultrasonic beamwidths could be determined by the ray-tracing techniques used in geometric optics. However,
P.17
this simple approach is inadequate because it does not account for diffraction, which arises from the finite sizes of ultrasonic wavelengths. The following sections
show how diffraction limits the lateral resolution attainable with unfocused and focused transducers.
Figure 1.21. Biometric error as a result of large beamwidth. Initial portion of echo from posterior wall arises from
off-axis points yielding an erroneously short measure of axial length.
Unfocused Transducers
Ultrasonic pulses produced by an unfocused transducer can be considered to arise from Huygens' point sources distributed over the transducer's flat, circular
surface. Huygens' sources describe the radiation of ultrasound waves in the same manner applicable to the radiation of light (2). Figure 1.22 depicts an ultrasonic
pulse arising from the superposition of spherical ultrasonic wavelets from each point source. (For clarity, only three sources are diagrammed.) When a transducer is
excited, each wavelet spreads through the transmission medium, causing large pressure amplitudes at points of constructive interference. On the other hand, no
pressure variations occur at points of destructive interference where the compression component from some point sources are exactly canceled by the
simultaneous rarefaction components from other sources. As the wavelets progressively spread through the medium, several distinctive effects are observed,
which have led to the concepts of near field (e.g., conditions at time t1) and far field (e.g., conditions at t 3).
Figure 1.22. Ultrasonic wave fronts emanating from three Huygens' point sources. Resultant pulse components
are shown at successive time intervals t1 , t2 , t3 .
Near-field (Fresnel's region) conditions apply when the pulse is still near the transducer. The length of the near field is equal to a2/?, where a is the radius of the
transducer rim. Here, little beam spreading occurs, and the beamwidth is equal to the transducer diameter. Pulse amplitudes vary rapidly over small distances
because of the complex interference patterns in this region. The rapid amplitude variations are evident at points situated on the transducer axis, as plotted in Figure
1.23. When the pulse passes from this region it enters the far field and experiences a gradual amplitude reduction as a result of progressive beam spreading.
In the far field (Fraunhofer's region), the beamwidth progressively increases as the pulse travels a greater distance, R, from the transducer. If the beam cross
section is examined, a distinctive lobed pattern is seen (Figures 1.24 and 1.25). The central main lobe contains more than 80% of the ultrasonic energy and
exhibits a gradually tapered pressure amplitude. There is also a series of peripheral side lobes encircling the main lobe. Side lobes are characterized by small
pressure amplitudes and are formed because of weak constructive interference; these low-amplitude regions result in corresponding weak tissue echoes, which
can assume importance, if high-gain electronic amplifiers are used.
The effective beamwidth in the far field is usually measured between those points in the main lobe, where the pressure amplitude falls by 3 dB to 70% of its
maximum value; this width is approximately equal to 0.5 (?/a)R. Often it is more meaningful to specify the angle that describes the 3-dB periphery of the main lobe
(Figure 1.25); this angle (in radians) is approximately equal to 0.5 (?/a).
P.18
Figure 1.23. Relative pressure amplitude as function of distance along tranducer's central axis. Amplitude
decreases monotonically for distances larger than a2 /?.
Figure 1.24. Dependence of far-field pressure amplitude upon angular position. Plot is drawn for points at a
fixed distance, R, from transducer center; abscissa values are proportional to off-axis distance, r.
P.19
Figure 1.25. Beam pattern of an unfocused transducer showing near-field and far-field regions. The 3-dB
angular width of the main lobe is 2 ?.
Ultrasonic wavelength and transducer dimensions determine both near-field and far-field characteristics. As ? decreases (frequency increases), the length of the
near field increases and the far-field beamwidth becomes narrower. As an example, if a is 2.5 mm and ? is 0.1 mm (15 MHz), the near-field length is 62.5 mm, and
the far-field beam angle is 1.2 degrees.
Ocular examinations with unfocused transducers are usually performed within the near field, where the beam is typically several mm wide. This degree of lateral
resolution is too coarse for many ophthalmic examinations, especially B-scans, can be improved by focusing, as discussed later.
Focused Transducers
Transducers are focused by using acoustic lenses, as discussed in connection with refraction. However, the width of the resultant beam is not zero at the focal
point, as diagrammed in Figure 1.17; rather, diffraction causes a small, but finite, beamwidth that depends on ultrasonic wavelength and transducer dimensions.
Figure 1.26. Beam pattern of focused transducer. Dotted line indicates focal plane.
Classic analyses treating Huygens' point sources show that focused beams have profiles, such as those shown in Figure 1.26 (30). In the focal plane, the beam
exhibits a lobed structure of the same type encountered under unfocused far-field conditions. Here, the side lobes straddle the narrow-focused region, and the main
lobe beamwidth is equal to 0.5(?/a)F, where F is the focal length of the transducer and a, again, represents the radius of the transducer rim. Typically, ? is 0.15 mm
(10 MHz), a is 5 mm, and F is 30 mm, resulting in a beamwidth of 0.45 mm. The same transducer dimensions provide smaller beamwidths as frequency is
increased.
In the selection of clinical focused transducers, the desired beamwidth can be obtained by proper selection of wavelength and transducer parameters (a and F).
However, several other factors must be considered when choosing a transducer that will permit adequate tissue examination. Focal lengths must be chosen to
provide focusing in the desired tissue region; focal lengths near 30 mm allow for intraocular focusing, whereas larger focal
P.20
lengths are needed for deeper orbital examinations. Furthermore, high frequencies cannot be used in the orbit because of attenuation. These combined limitations
on focal lengths and frequencies constrain lateral resolution in the orbit more severely than in the eye. In addition, clinical transducers must be weakly focused to
generate narrow beams over relatively long tissue depths. Strong focusing, obtained with large transducer diameters, is not usually desirable; although strong
focusing produces narrow beamwidths, it can do so only over unacceptably shallow tissue depths. The specification of an f-number (equal to F/2a) summarizes
these considerations for ultrasound in the same manner used in optics. Small f-numbers (e.g., 1.5) provide sharp focusing over limited depths; large f-numbers
(e.g., 3-4) provide moderate focusing over longer depths.
The same considerations apply to very high frequency (VHF) transducers used in 40-MHz examinations of anterior chamber structures. For example, transducers
with 3-mm radii and 12-mm focal lengths can produce beamwidths near 50 µm for very fine lateral resolution. However, their depths of focus can be less than 1
mm, so that care must be used in positioning the transducer focal zone over the tissue segment to be examined.
These discussions of unfocused and focused transducer beam patterns are useful in all cases, but they are rigorously applicable only when the ultrasonic pulse
contains several cycles of oscillation. They must be modified, if the transducer generates a very short pulse, such as the single-cycle pulse illustrated in Figure
1.20. Short pulse durations do not allow enough time for standard interference patterns to develop. For a single-cycle pulse, the 3-dB main lobe width is equal to
that quoted previously, but side lobe patterns and near-field characteristics differ, in several respects, from those discussed previously (31).
COMPOSITE RESOLUTION
Composite (axial and lateral) resolution depends on frequency and transducer geometry. These factors, in turn, are influenced by the type of examination to be
made. As a brief summary of preceding sections, anterior segment examinations present an environment where fine resolution can be achieved; ocular
examinations permit fine resolution; orbital examinations must be carried out at lower resolution.
For an ideal transducer, both components of resolution improve as resonant frequency is increased; however, attenuation also increases with frequency, limiting
the values that can be realized in practice. In the anterior segment, frequencies near 40 MHz have provided axial resolution of 30 µm and lateral resolution near 50
µm. Within the eye, frequencies near 20 MHz have been used to attain an axial resolution of 0.1 mm; focal lengths of 30 mm, together with a transducer radius of 5
mm, allow lateral resolution of 0.2 mm. In orbital examinations, increased absorption currently limits frequencies to a maximum near 10 MHz. These lower
frequencies and necessarily long focal lengths (e.g., 60 mm) reduce axial resolution to approximately 0.3 mm and lateral resolution to 0.9 mm. Newer, more
sensitive transducers promise to alleviate some of these restrictions, especially in anterior regions of the orbit (32).
Transducer Arrays
Single-element transducers, as shown in Figure 1.1, are a standard use in ophthalmology, and mechanical scanning is used to generate B-mode images. Other
medical specialties frequently use piezoelectric arrays for electronic focusing and scanning (4,33). Arrays comprise a set of small, discrete elements that transmit
and receive ultrasonic pulses, emulating Huygens' sources that can be individually controlled to electronically focus and steer the overall beam. Electronic beam
control provides rapid, versatile operation, without the need for mechanical scanning. However, arrays increase the complexity and cost of ultrasonic instruments.
For high-quality performance, the sizes and spacings of array elements must be comparable to or smaller than the ultrasonic wavelength; arrays are common in
systems using frequencies of, for example, 7 MHz, but they have not yet found widespread use at the higher frequencies and smaller wavelengths used in ocular
examinations.
There are three basic types of arrays: linear arrays (for linear scanning), phased arrays (for sector scanning), and annular arrays (for controlling focal zones along
single scan lines). Each of these controls the timing of excitation pulses to focus the transmitted pulse and also applies time delays to returned echoes to focus
received signals.
A linear array comprises a set of thin, parallel, rectangular elements on a planar substrate. On transmit, focusing is achieved by exciting each element in a
programmed sequence. This is shown schematically for three array elements in Figure 1.27. Elements at the edge of the desired beam are excited first, and the
central element is excited last. The timing of the excitations is adjusted so that the pulses from all elements arrive simultaneously and in phase at the desired focal
point, producing a large focal-point pressure pulse (at time t2 in the figure). In the receive mode, echo signals from the elements are time-shifted in a similar
manner; the maximum time delay is applied to the central element, so that echoes from the focal point are aligned in time. The shifted RF echoes are then summed
to obtain the desired receive focusing. Arrays permit dynamic focusing, which extends the effective depth of focus on receive. In this mode, the applied echo
time-shifts are continually adjusted to maintain an effective focus at the distance from which echoes are returning at that time. Accordingly, the effective focal
length, Fe, is progressively increased as time proceeds so that Fe is equal to c te/2, where c is
P.21
the assumed speed of propagation, and te is time after excitation.
Figure 1.27. Ultrasonic wave front emanating from three array elements at three sequential instants of time.
Excitation pulses have been timed so that beam is centrally focused at time t2 .
These time-shift operations provide focused examination along a single scan line. Linear arrays may contain 256 elements and use a subset of, for example, 16
active elements for examinations along this line. Linear beam scanning is achieved by progressively shifting the groups of active elements that are used. Thus, the
total lateral extent of the scanned area is equal to the length of the transducer. Linear arrays provide electronic focusing in the azimuthal plane, perpendicular to the
long axis of the elements. A cylindrical lens is used to provide fixed focusing in the orthogonal elevation plane.
Phased arrays use similar operations to obtain sector scans, rather than linear scans, within a fan-shaped region of space. These transducers contain fewer
elements, and they focus along single lines, using the same time-shift procedures as linear arrays. However, the scan line orientation is varied by linearly
time-shifting the excitation of each element to produce a wavefront that is tilted along the desired direction. For example, to angulate the beam toward the left, the
element at the right edge of the array is excited first, and the left-edge element is excited last.
Annular arrays use concentric circular elements to control the effective focal length along the transducer's central axis. A focused beam is launched by exciting the
outer ring first and the inner element last. Again, the time shifts are computed so that pulses from all elements arrive simultaneously at the desired focal point. To
achieve dynamic focusing, echo components from each element are time-delayed and summed, as described for linear arrays. Annular arrays do not support
electronic scanning, but they do permit focal-point control and increased depths of focus. They are often used to vary the focal point around a default value set by
an acoustic lens placed in front of the array.
BIOLOGIC EFFECTS OF HIGH-INTENSITY ULTRASOUND

Intense ultrasound can modify tissue structures by a number of mechanisms that depend on the intensity and pressure amplitude of the incident ultrasonic beam
(34). Ultrasonic intensity is defined as the amount of ultrasonic energy passing through a unit area in a unit time. In a plane ultrasonic wave, the intensity, I, is
related to the amplitude, p(t), of the ultrasonic pressure variations and the characteristic acoustic impedance of the transmission medium:
where the superscript bar denotes an average over time. I is usually specified in watts/cm2. In practice, a set of subscripts denotes the spatial and temporal
averaging intervals used in intensity specifications, as subsequently described for the U.S. Food and Drug Administration (FDA) exposure indices.
At high exposure levels, ultrasound can alter tissues by thermal effects, mechanical phenomena, and cavitation (34). Thermal effects arise from absorbed
ultrasonic energy, which is converted to heat. When the incident energy is sufficiently high, the corresponding temperature rise can damage or denature tissue
constituents. Mechanical effects can occur when the incident beam is absorbed or reflected by a tissue structure. These phenomena redirect the beam's
momentum and generate radiation forces that can produce tissue motion or fluid streaming at high intensities. Cavitation, which occurs most readily at low
frequencies, can occur with large ultrasonic pressure oscillations; in these cases, the negative pressure may promote the formation of gas-filled microbubbles that
can grow until a positive pressure cycle causes their sudden collapse. The collapse can be accompanied by large, mechanical forces locally disrupting tissues in
the vicinity of the bubble.
In view of the widespread application of diagnostic ultrasound, the low intensities and pressures used in current systems pose no known threat of tissue damage.
Animal studies, the absence of reports of clinical damage, and extrapolation of laboratory data all point to the safety afforded by present diagnostic systems. FDA
guidelines, described in the following section, have been formulated to continue this record of safety. At high exposure
P.22
levels, however, animal studies reveal that physical alterations can be produced in ocular tissues.
Effects of high-intensity ultrasound on ocular tissues have been studied since 1938, when Zeiss (31) described in-vitro cataract production. A series of
investigations has shown that thermal phenomena can produce cataracts (35, 36, 37, 38) and lesions of the cornea (39), choroid, retina, and sclera (40, 41, 42).
Mechanical vascular compression can potentiate chorioretinal lesion production by suppressing blood-flow cooling (42). The exposure levels needed to produce
ocular damage were orders of magnitude larger than those used in diagnostic systems. This finding is consistent with a report that found no ocular damage after
diagnostic exposures as long as 4 hours (43). Animal studies have also found no damage following prolonged, elevated exposures using VHF (40 MHz) diagnostic
systems(44).
High-intensity focused ultrasound (HIFU) has been studied as a means of exploiting these physical interactions to treat diseases in various organs (45,46).
Typically, HIFU treatments use a series of focal lesions produced, using exposures of several hundred watts/cm2, with durations of several seconds; these
exposures are designed to produce desired effects before blood-flow cooling becomes significant. In the eye, glaucoma has been treated in humans, using thermal
ciliary body lesions to decrease aqueous humor production and potentiate alternative outflow pathways (47). In animals, chorioretinal lesions produced retinal
adhesion similar to that found using lasers; HIFU exposures prevented the spread of retinal tears and facilitated reattachment of detached retinas (48). In rabbits,
vitreous hemorrhages (49) and membranes (50) were successfully disrupted by pulsed HIFU beams, which cause mechanical agitation to promote intravitreal
dispersion. Tumor therapy has been investigated, using HIFU to treat human melanoma explants in nude athymic mice (51,52). Ultrasonic hyperthermia has also
been applied to tumors, using broad beams with lower intensities (several watts/cm2) to achieve sustained heating (near 45°C) for 30 minutes, for example (53,54).
FDA EXPOSURE INDICES

The FDA has devised safety guidelines to assure that all diagnostic ultrasound devices produce exposure levels that are below specific exposure thresholds
(55,56). These indices are defined as follows:
Spatial-peak pulse-average intensity: ISPPA.3
Spatial-peak temporal-average intensity: ISPTA.3
Mechanical index: MI = Pr.3/fc 0.5
Thermal index: TI = Wfc/210
Intensities are specified in W/cm2, using subscripts to denote spatial and temporal factors. The SP (spatial peak) subscript indicates the maximum intensity level in
the beam. The PA (pulse average) subscript indicates a temporal average over the duration of a single pulse. The TA (time average) subscript indicates a temporal
average over the time interval from one pulse to the next; this interval is set by the pulse repetition frequency (A-mode) or scan rate (B-mode). The subindex 0.3
indicates that the value is “derated” for the effect of attenuation (assumed to have a value of 0.3 dB cm-1 MHz -1) between the transducer and the measurement
point.
The mechanical index MI is defined as the derated peak rarefaction pressure, Pr (in megapascals), divided by the square root of center frequency, fc, (in
megahertz). MI is a unitless number related to the risk of cavitation. The thermal index TI is defined as the output power W (in milliwatts) times the center frequency
f (in megahertz) divided by 210 mW MHz. The denominator is considered to be the power level required to raise tissue temperature 1°C. Thus, the TI is a unitless
number, which at a value of unity indicates that a 1°C temperature increase in the insonified tissue would be expected.
For an ultrasound unit to be sold in the United States, it must meet FDA standards. The FDA provides two tracks under which a diagnostic ultrasound device can
meet these regulatory standards. Under Track 1, the instrument manufacturer demonstrates conformity with 510(k) standards, levels that are deemed to be safe in
diagnostic instruments, based on historic experience. For ophthalmology, these levels are as follows: ISPTA.3 = 17 mW/cm2, I SPPA.3 = 28 W/cm 2, MI = 0.23 (55).
These values are well below those in any other specialty. For instance, the ISPTA.3 thresholds for peripheral vessels, cardiac, and fetal imaging are 720,430, and 94
mW/cm2, respectively, as compared to 17 mW/cm 2 in ophthalmology. This conservative ophthalmic threshold is a consequence of the concern for cataract
formation, which is relatively high because of the high attenuation coefficient of collagen in combination with the lack of vascular cooling within the lens. Virtually all
ophthalmic diagnostic ultrasound units follow Track 1.
Track 3 devices follow the Output Display Standard (56). Under this approach, the instrument must display the MI and/or TI, if conditions exist, under which either
might exceed a value of 1.0. For Track 3 ophthalmic systems, the TI must not exceed 1.0, the MI must not exceed 0.23, and the ISPTA.3 must be under 50 mW/cm2.
Track 3 is used most commonly in general purpose instruments that might include a small-parts probe suitable for ophthalmic examinations.
Some debate continues as to the validity of the ophthalmic standards (57), especially for very high frequency (VHF) ultrasound (44). The FDA periodically reviews
existing standards in light of current research.
P.23
Figure 1.28. Diagram of schlieren system.
MEASUREMENT OF ULTRASONIC BEAM PARAMETERS

Ultrasonic exposure levels must be determined for some research applications, and manufacturers must specify them to assure compliance with FDA guidelines
(58). Small ultrasonic probes (hydrophones) are most often used for accurate measurements of ultrasonic pressure pulses. Needle hydrophones use small
transducers as receivers to measure local values of pressure as a function of time. Other hydrophones use large PVDF membranes whose electrodes define small
active areas for these measurements; such probes rely on PVDF because it is well matched to water so that their presence does not significantly affect the incident
beam. Both types of hydrophone can be calibrated and scanned through the incident beam in a water-filled tank. Their output voltages can be directly related to
pressure (in pascals) as a function of time.
Schlieren techniques use an optical system to produce a visible image of an ultrasonic beam. The light intensity at each point in a schlieren image is related to the
average pressure amplitude within the imaged beam and provides a semiquantitative measure of beam strength.
In a schlieren system (Figure 1.28), a point source of light and a collimating lens combine to produce a plane wave of light that passes through a fluid-filled optical
cell. The light exiting from the cell is focused by an integrating lens upon a small, opaque optical stop. If there are no ultrasonic waves propagating through the fluid
in the optical cell, all light is blocked by the stop.
The transducer to be studied is placed in the cell and excited with a continuous-wave sinusoidal voltage. Ultrasonic waves perturb the optical index of refraction
within the cell fluid so that the light emerging from the cell is nonplanar and, therefore, is no longer completely focused on the stop. That portion of light bypassing
the stop contains spatial and amplitude information relating to beam structure. A reimaging lens converts this information into image form. (If the optical stop were
not used, unaffected portions of the incident light would obscure the schlieren image.) Sensitive schlieren systems have been used in a more quantitative fashion to
examine pulsed beams.
Radiation pressure techniques have been used to measure the ultrasonic power emanating from a diagnostic transducer. In some implementations, a beam is
reflected at 45 degrees by a highly reflective plate. An analytic balance measures the small force on the plate that results from the redirection of wave momentum.
This force is directly related to the incident ultrasonic power; a power level of one mW produces a force of 0.067 mg. Carefully designed measuring systems with
sensitive balances are capable of measuring the low power levels encountered in diagnostic systems.
In the next chapter, the use of the physical principles in designing and constructing clinical instruments for ocular examination will be discussed.
REFERENCES
1. Mundt GH, Hughes WF. Ultrasonics in ocular diagnosis. Am J Ophthalmol. 1956;41:488-498.
2. Kinsler L, Frey A. Fundamentals of Acoustics. 4th ed. New York: John Wiley and Sons; 2000.
3. Kino GS. Acoustic Waves: Devices, Imaging, and Analog Signal Processing. Englewood Cliffs, NJ: Prentice-Hall; 1987.
4. Kremkau FW. Diagnostic Ultrasound: Principles and Instruments. Philadelphia: WB Saunders; 2002.
5. Pavlin CJ, Harasiewicz K, Foster FS. Ultrasound biomicroscopy of anterior segment structures in normal and glaucomatous eyes. Am J Ophthalmol.
1992;113:381-389.
6. Silverman RH, Lizzi FL, Ursea BG, et al. High resolution ultrasonic imaging and characterization of the ciliary body. Invest Ophthalmol Vis Sci.
2001;42:885-894.
7. Fujimoto JG, Brezinski ME, Tearney GJ, et al. Optical biopsy and imaging using optical coherence tomography. Nature Medicine. 1995;1:970-972.
P.24
8. Sherar MD, Foster FS. The design and fabrication of high frequency poly(vinylidene fluoride) transducers. Ultrason Imaging. 1989;11:75-94.
9. Chivers RC, Round WH, Zieniuk JK. Investigation of ultrasound axially traversing the human eye. Ultra Med Biol. 1984;10:173-188.
10. Oksala A, Lehtinen A. Measurement of the velocity of sound in some parts of the eye. Acta Ophthalmol. 1958;36:633-639.
11. Thijssen MJ, Mol MJ, Timer MR. Acoustic parameters of ocular tissues. Ultra Med Biol. 1983;11:157-161.
12. Jansson F, Kock E. Determination of the velocity of ultrasound in the human lens and vitreous. Acta Ophthalmol. 1962;40:420-433.
13. De Korte CL, van der Steen AF, Thijssen JM. Acoustic velocity and attenuation of eye tissues at 20 MHz. Ultra Med Biol. 1994;20:471-480.
14. Ye SG, Harasiewicz KA, Pavlin CJ, et al. Ultrasound characterization of normal ocular tissue in the frequency range from 50 MHz to 100 MHz. IEEE
Trans Ultrason Ferroelectr Freq Control. 1995;42:8-14.
15. Coleman DJ, Lizzi FL, Franzen LA, et al. A determination of the velocity of ultrasound in cataractous lenses. In: Francois J, Goes F, eds. Ultrasonography
in Ophthalmology. Basel, Switzerland: Karger; 1975:246-251.
16. Willard G. Temperature coefficients of ultrasonic velocity in solutions. J Acoust Soc Am. 1947;19:235-241.
17. Frucht AH. Die schallgeschwindigkeit in menoschlichen and tierischen geweben. Z Ges Exp Med. 1953;120:526-557.
18. Bamber JC, Hill CR. Ultrasonic attenuation and propagation speed in mammalian tissues as a function of temperature. Ultrasound Med Biol.
1979;5:149-157.
19. Muir TG, Carstensen EL. Prediction of nonlinear acoustic effects at biomedical frequencies and intensities. Ultrasound Med Biol. 1980;6:345-357.
20. Baum G, Greenwood I. The application of ultrasonic locating techniques to ophthalmology: part I. Am J Ophthalmol. 1958;46:319-329.
21. Feleppa EJ, Lizzi FL, Coleman DJ. Diagnostic spectrum analysis in ophthalmology: a physical perspective. Ultrasound Med Biol. 1986;12:623-631.
22. Lizzi FL, Feleppa EJ, Alam SK, et al. Ultrasonic spectrum analysis for tissue evaluation. Special Issue on Ultrasonic Image Processing & Analysis in
Pattern Recognition Letters. 2003;24:637-658.
23. Deng CX, Lizzi FL. A review of physical phenomena associated with ultrasonic contrast agents and illustrative clinic applications. Ultrasound Med Biol.
2002;28:277-286.
24. Begui ZE. Acoustic properties of the refractive media of the eye. J Acoust Soc Am. 1954;26:365-368.
25. Filipczynski L. Visualizing internal structures of the eye by means of ultrasonics. Proc Vibr Probl. 1967;4:357-368.
26. de Korte CL, van der Steen AF, Thijssen JM. Acoustic velocity and attenuation of eye tissues at 20 MHz. Ultrasound Med Biol. 1994;20:471-480.
27. Lizzi F, Burt W, Coleman DJ. Effects of ocular structures on the propagation of ultrasound in the eye. Arch Ophthalmol. 1970;84:635-640.
28. Redwood M. Transient performance of a piezoelectric transducer. J Acoust Soc Am. 1961;33:527-535.
29. Cannata JM, Ritter TA, Chen W, et al. Design of efficient, broadband single-element (20-80 MHz) ultrasonic transducers for medical imaging applications.
IEEE Trans Ultrason Ferroelectr Freq Control. 2003;50:1548-1557.
30. O'Neil HT. Theory of focusing radiators. J Acoust Soc Am. 1949;21:516-526.
31. Arditi M, Foster FS, Hunt JW. Transient fields of concave annular arrays. Ultrason Imaging. 1981;3:37-61.
32. Coleman DJ, Silverman RH, Chabi A, et al. High resolution ultrasonic imaging of the posterior eye. Ophthalmology. In press.
33. Shung KK, Zipparo M. Ultrasonic transducers and arrays. IEEE Eng Med Biol Mag. 1996;15:20-30.
34. Nyborg WL, Ziskin MC, eds. Biological Effects of Ultrasound. New York: Churchill Livingstone; 1985.
35. Zeiss E. Effects of ultrasound on excised bovine lenses. Graefe Arch Clin Exp Ophthalmol. 1938;139:301-322.
36. Lavine O, Langenstrass KH, Bowyer CM, et al. Effects of ultrasonic waves on the refractive media of the eye. Arch Ophthalmol. 1952;47:204-219.
37. Torchia RT, Purnell EW, Sokollu A. Cataract production by ultrasound. Am J Ophthalmol. 1967;64:305-309.
38. Lizzi FL, Packer AJ, Coleman DJ. Experimental cataract production by high frequency ultrasound. Ann Ophthalmol. 1978;10:934-942.
39. Rutzen AR, Roberts CW, Driller J, et al. Production of corneal lesions using high-intensity focused ultrasound. Cornea. 1990;9:324-330.
40. Purnell EW, Sokollu A, Torchia R, et al. Focal chorioretinitis produced by ultrasound. Invest Ophthalmol. 1964;3:657-664.
41. Lizzi FL, Coleman DJ, Driller J, et al. Experimental ultrasonically induced lesions in the retina, choroid, and sclera. Invest Ophthalmol Vis Sci.
1978;17(4):350-360.
42. Lizzi FL, Coleman DJ, Driller J, et al. Effects of pulsed ultrasound on ocular tissue. Ultrasound Med Biol. 1981;7(3):245-252.
43. Ziskin M, Romayandanda N, Harris K. Ophthalmologic effect of ultrasound at diagnostic intensities. J Clin Ultrasound. 1974;2:119-122.
44. Silverman RH, Lizzi FL, Ursea BG, et al. Safety levels for exposure of cornea and lens to very high-frequency ultrasound. J Ultrasound Med.
2001;20:979-986.
45. Hill CR, ter Haar GR. Review article: high intensity focused ultrasound—potential for cancer treatment. Br J Radiol. 1992;68:1296-1303.
46. Hynynen K. Review of ultrasound therapy. In: Proceedings of IEEE Ultrasonics Symposium. 1997:1305-1313.
47. Silverman RH, Vogelsang B, Rondeau MJ, et al. Therapeutic ultrasound for the treatment of glaucoma: results of a multicenter clinical trial. Am J
Ophthalmol. 1991;111:327-337.
48. Rosecran LR, Iwamoto T, Rosado A, et al. Therapeutic ultrasound in the treatment of retinal detachment: clinical observations and light and electron
microscopy. Retina. 1985;5:115-122.
49. Lucas BC, Driller J, Iwamoto T, et al. Ultrasonically induced disruption and hemolysis of vitreous hemorrhages. Ultrasound Med Biol. 1989;15:29-37.
50. Coleman DJ, Lizzi FL, El-Mofty AA, et al. Ultrasonically accelerated resorption of vitreous membranes. Am J Ophthalmol. 1980;89(4):490-499.
51. Lizzi FL. High-precision thermotherapy for small lesions. Eur Urol. 1993;23:23-28.
52. Lizzi FL, Astor M, Deng CX, et al. Control of lesion geometry using asymmetric beams for ultrasonic tumor therapy. Proceedings of SPIE Conference on
Ultrasonic Transducer Engineering, 1998;3341:99-106.
53. Coleman DJ, Silverman RH, Iwamoto T, et al. Histopathologic effects of ultrasonically induced hyperthermia in intraocular malignant melanoma.
Ophthalmology. 1988;95:970-981.
54. Coleman DJ, Silverman RH, Ursea R, et al. Ultrasonically induced hyperthermia for adjunctive treatment of intraocular malignant melanoma. Retina.
1997;17:109-117.
P.25
55. Information for Manufacturers Seeking Marketing Clearance of Diagnostic Ultrasound Systems and Transducers. Rockville, MD: Food and Drug
Administration, Center for Devices and Radiological Health; 1997.
56. Standard for Real-Time Display of Thermal and Mechanical Acoustic Output Indices on Diagnostic Ultrasound Equipment. Laurel, MD: American Institute
of Ultrasound in Medicine; Arlington, VA: National Electrical Manufacturers Association; 1998.
57. Herman BA, Harris GR. Theoretical study of steady-state temperature rise within the eye due to ultrasound insonation. IEEE Trans Ultrason Ferroelectr
Freq Control. 1999;46:1566-1575.
58. Ziskin MC, Lewin PA, eds. Ultrasonic Exposimetry. Boca Raton, FL: CRC Press; 1993.
> Table of Contents > 2 - Ultrasonic Systems
2
Ultrasonic Systems
The two most commonly used ultrasonic imaging modalities in ophthalmology are termed A-mode and B-mode. Each presents anatomic information in a distinctive
display format. A-mode refers to a graphic display of echo amplitude as a function of distance along one line of sight, or vector (Figure 2.1). A-mode was the first
display mode to be used in ophthalmology (1). It is used in characterization of tissues such as intraocular tumors and vitreous hemorrhage. It is also widely used in
biometric applications, such as axial length measurement and corneal pachymetry. B-mode, introduced in the late 1950s (2), refers to a display of two-dimensional
cross-sectional images (Figure 2.2). These images provide representations of the anatomy of the eye and orbit that have proven useful in diagnosis of a broad
spectrum of disease states. A- and B-mode systems may be found in instruments dedicated to one function only or may be combined in a single instrument.
A-mode displays may be generated using a special purpose A-mode transducer or may be generated from individual vectors comprising a B-mode display.
Although the basic physical principles discussed in Chapter 1 underlie the operation of all ultrasonic systems, it is electronic and computer technology that
translates these principles into practical clinical instruments. The electronic and computer components of modern ultrasound scanners are used to generate
ultrasonic pulses, process echoes, and display images and information. This chapter discusses the electronic components used in each of these stages for
generation of A- and B-mode images and how the characteristics of the individual components influence the quality of the resulting images. It also discusses the
means of recognizing and eliminating misleading results stemming from improper system adjustment. Throughout this chapter, emphasis is given to the overall
quality of an ultrasonogram in terms of three parameters: resolution, sensitivity, and dynamic range. Spatial resolution, defined in Chapter 1, refers to the ability to
distinguish two nearby reflectors. Temporal resolution, also to be considered here, refers to the ability to visualize tissue changes occurring over time. Sensitivity
refers to the weakest reflector that can be detected in a displayed ultrasonogram. Dynamic range describes the spread of echo amplitudes that can accurately be
portrayed in an ultrasonogram.
In addition to A- and B-modes, this chapter will describe modalities less commonly used or more recently introduced in ophthalmology, including Doppler, M-mode,
and swept-mode.
SYSTEM COMPONENTS
The ultrasound system, schematically represented in Figure 2.3, consists of the following components:
Transducer/probe
Servo (for B-mode systems)
Pulser
Receiver
Scan converter and display
The trend for system design in modern instrumentation is toward integrated digital components. By placing the previously mentioned functions on a single computer
board, the system becomes less expensive, more reliable, and easier to repair. Because of this integration, not all of the components are represented as
stand-alone devices, but we shall consider them as functionally separate entities.
PROBE
In B-mode systems, a mechanism is needed to sweep the ultrasound beam across a scan plane. In mechanical sector
P.28
scan systems, which dominate in ophthalmology, the transducer is enclosed within a sealed, fluid-filled housing with an acoustically transparent cap at one end. In
operation, the transducer is pivoted over an angle of 45 to 60 degrees at a rate of several times per second. This refresh rate, called the frame rate, is typically
about 10 Hz, but in some instruments rates of 30 Hz or more have been attained.
Figure 2.1. A-scans consist of a plot of echo amplitude as a function of range. This figure shows a typical
A-scan along the axis of a normal eye performed in contact mode through the eyelid. Peaks correspond to
eyelid (L), cornea (C), anterior (AL), and posterior lens (PL) surfaces and retina (R).
SERVO
The servo is a device that controls the motion of the transducer within the probe and registers the orientation of the transducer at each moment of time. The servo
controls a motor incorporated within the probe, and, as the transducer moves, the servo continually monitors its position. Each scan frame consists of a fixed
number of vectors (typically 256) that are evenly spaced within each scan frame. As the motor sweeps the transducer, the servo monitors its position and issues
signals to the pulser and other components such that pulse/echo vectors are acquired at appropriate positions.
Figure 2.2. Figure 2.1 illustrates the difficulty in establishing context for an A-scan seen in isolation. This
B-mode image shows the vector from which the A-scan in Figure 2.1 was derived but in the context of the ocular
anatomy revealed in B-mode.
Figure 2.3. Schematic representing electronic components comprising a B-mode system. These include the
probe (containing the transducer and mechanical scan mechanism), the pulser/receiver (which excites the
transducer and amplifies echoes), the servo (which controls and monitors transducer orientation), the scan
converter (which digitally formats echo data for display), and the display device.
PULSER
The ultrasonic pulser repeatedly “shock excites” the transducer with short voltage pulses applied across the electrodes of its piezoelectric element. Each excitation
results in the generation of an ultrasonic pulse. The pulse repetition frequency (PRF) must be low enough to allow all returning echoes to be received by the
transducer before the next pulse is generated. If we consider the distance to the optic chiasm (the longest distance that we would need to consider) to be
approximately 6 cm, then we can determine the two-way travel time for an acoustic pulse to be 2 × 0.06 m ÷ 1,540 m/sec = 0.078 msec. Typically, a 1 KHz pulse
repetition frequency is used so that pulses are generated at 1-msec intervals, much longer than the time for pulse and echo travel from the deepest portion of the
orbit.
The nature of the ultrasonic pulse is an important factor in determining the characteristics of the emitted acoustic pulse, and hence, of the attainable resolution of
the ultrasound system. The characteristics of the pulse that are significant include its form, duration, and amplitude. The most common pulse form used in A- or
B-mode imaging is a negative spike, although other modes, such as monocycle (a single sine-wave) are sometimes used. Long excitation pulses result in
correspondingly long ultrasonic pulses and concomitantly poor axial resolution. Thus, negative spike impulses are normally designed to be of short duration. Unlike
negative spike impulses, monocycles can be tuned to a specific frequency that may or may not correspond to the natural harmonic frequency of the transducer.
Monocycles can thus be used to alter
P.29
the output of a given transducer to frequencies that are somewhat higher or lower than its natural frequency. The amplitude of the excitation pulse is important in
achieving adequate sensitivity. Typically, an amplitude of 100 to 400 volts will result in ultrasonic pulses large enough to produce detectable echoes from weakly
reflecting interfaces, such as the vitreoretinal interface. Although higher voltages provide higher amplitude pulses, and, consequently, increased sensitivity,
extremely high voltages can result in distortion of the acoustic pulse (e.g., ringing) that results in decreased axial resolution. In some systems, access to pulser
characteristics, such as damping and energy, allows the excitation pulse to be varied between extremes of a short, low-amplitude pulse (providing high resolution
and low sensitivity) and a long, high-amplitude pulse (providing low resolution and high sensitivity). In instruments where pulse characteristics are accessible to the
user, the clinician can determine the combination of resolution and sensitivity most advantageous in a given situation.
RECEIVER
Following transducer excitation, echoes from intraocular and orbital structures impinge upon the transducer, which generates small voltages that are proportional to
echo amplitude. These voltages must be processed by the amplifier and associated electronics before they can be displayed in a useful format. The required
operations are carried out by the electronic receiver, whose most important function is amplification of the minute voltages generated by the transducer. Gains of
100 (40 dB) or more are needed to raise the amplitudes of these signals from their initially low levels (e.g., 1 millivolt) to levels that are compatible with display
requirements. The receiver's functions include a limiter to prevent damage from the high voltage excitation spike produced by the pulser, sometimes a low-noise
pre-amplifier (which, in some systems, is incorporated in the probe) to boost gain, and a variety of other functions, possibly including time-gain control (TGC),
compression, noise reject, and envelope detection.
A factor of prime importance when regarding receiver gain is saturation. This occurs whenever the amplified pulse reaches the maximum level (e.g., 5 volts) that
the amplifier can supply. If an input signal is too large (or if the gain is too high), saturation occurs and can eliminate clinically significant information regarding echo
strength (Figure 2.4). Saturation is always encountered at the beginning of an A- or B-mode display, when a part of the large excitation pulse is picked up by the
amplifier. This so-called “main bang” causes a dead space directly in front of the transducer and obliterates echoes from close objects. For this reason, the region
of the main bang is not displayed in most scan systems and is the reason that all transducers require some sort of a “stand off” between the transducer and the
tissue of interest.
Figure 2.4. All amplifiers have a saturation level above which signals are truncated or otherwise distorted.
Reject levels, which may be user-controllable, allow low-level signals (such as noise) to be suppressed. The
figure demonstrates the effects of saturation and reject on high- and low-level signals.
Electronic noise limits the detection of small echoes and can present severe problems with large bandwidth amplifiers. Noise consists of small, random voltage
variations that arise because of statistical fluctuations of electrons in system components. At high gains the resultant noise signals appear as “grass” in an
ultrasonogram and obscures low amplitude echoes. Some ultrasound units provide reject controls that establish a threshold level that can be set to prevent noise
from being displayed. Reject controls, if improperly adjusted, can suppress the display of small or modest echoes that may be diagnostically significant (Figure 2.4).
Amplifiers often include filters that pass signals within only a certain frequency range (usually surrounding the transducer center frequency). This frequency
bandpass operation reduces noise from extraneous sources.
Saturation and noise combine to determine the dynamic range of an amplifier. Dynamic range is defined as the spread of input signal amplitudes that result in
meaningful output signals. The largest useful input signal is one that just causes saturation; the smallest useful input signal is one that yields an output just above
the amplifier noise level (or reject level). Typical amplifiers have dynamic ranges of 60 dB, that is, the maximum input signal is one million times larger than the
minimum input signal. This range is sufficient for accurate displays of tissue echoes.
TGC is a form of amplification in which gain increases as a function of range. This feature is designed to compensate for the reduction in echo amplitude that
occurs with depth as a result of attenuation of acoustic energy with depth. As noted in Chapter 1, absorption causes a progressive loss in the strength of a
propagating ultrasonic pulse. Absorptive attenuation of echo strength increases exponentially with distance. Compensation for this effect involves increasing
amplifier gain with time, t, in a manner which is the inverse of absorptive
P.30
decay. The appropriate gain is thus of the form of e?ct, where ? is the absorption coefficient and c is the velocity of ultrasonic propagation. Although TGC is useful, it
cannot compensate exactly for absorption in the eye and orbit because absorption coefficients differ greatly in specific ocular and orbital tissues. Most instruments,
therefore, do not use the function shown previously, but rather let the user adjust gain arbitrarily over a series of “gates,” that is, successive nonoverlapping range
windows. An example of a TGC-modified image is provided in Figure 2.5.
Envelope detection, or demodulation, is used to convert the positive and negative voltages (and nulls) that are present in the raw radiofrequency (RF) echo data to
positive signals, sometimes referred to as the video signal or envelope. This operation, shown schematically in Figure 2.6, includes a rectification stage, in which
all negative voltages are converted to positive values, followed by a suitable low-pass filter operation, which is designed to smooth out the nulls in the rectified RF
waveform, without excessive blurring. Demodulation is a crucial step for formation of images, because A- and B-mode images can represent positive values only. It
is the signal envelope that is generally used for formation of A-mode and B-mode images.
Figure 2.5. Time gain control (TGC) consists of manipulation of gain as a function of range, thus allowing
enhancement of signals that would otherwise be lost as a result of attenuation. Top: B-scan of an eye and orbit
without use of TGC, as indicated by the flat TGC curve below the image. Bottom: Same image with TGC
applied, demonstrating enhancement of the orbit.
Figure 2.6. This schematic illustrates stages in processing of echo data. Radiofrequency (RF) data
represent the positive and negative voltages associated with the positive and negative pressures of the
acoustic waves but are inconvenient for display purposes. The first stage of processing is rectification,
in which the negative voltages are rendered positive. Center: The rectified data, however, are choppy
due to the nulls in the RF waveform. The rectified data are therefore smoothed using a low-pass filter,
as shown in the lowermost waveform. This waveform is termed the “envelope” and is used for A- and
B-mode displays.
Compression is an operation that reduces the contrast between high and low amplitude echoes. This operation is important because display devices have a
dynamic range that is far less than the amplitude range of linearly amplified echoes. Logarithmic amplification is one such compression scheme. In a standard
(linear) amplifier, the gain, G, is independent of the input signal level, and an output voltage is equal to G times the corresponding input pulse. In a logarithmic
amplifier, the
P.31
output voltage is proportional to the logarithm of the input voltage. As shown in Figure 2.7, this type of amplification yields a large gain for small signals and a small
gain for large signals. Thus, a logarithmic characteristic reduces differences in echo levels and permits a wider range of input signals to fall between the noise and
saturation levels of an amplifier or digitizer. Logarithmic amplification can be used effectively in B-mode systems, where it compresses the large range of input echo
signals to the smaller range of brightness levels that can be presented on digital displays. Other nonlinear functions have also been used to advantage in
ultrasonography. An S-shaped function, for example, provides the greatest dynamic range at intermediate signal levels and has proven useful in A-mode
examinations, as used in “standardized” A-scan ultrasonography (3). However, such nonlinear amplification makes it difficult to recognize relative echo amplitudes
and to compare clinical echo patterns with those obtained in standard systems. Characteristic A-mode patterns can be easily distorted, unless great care is
exercised in adjusting all system parameters.
Figure 2.7. Logarithmic amplification may be used to extend dynamic range by reducing contrast between bright
and faint echoes. This allows display of faint echoes without causing saturation of bright echoes.
Demodulation converts RF data into a convenient format for display but does this at the sacrifice of some of the information content inherent in the original echo
data. If we recall that the reflection coefficient is defined as (Z2-Z1)/(Z1 + Z2), where (as in Chapter 1) Z refers to acoustic impedance, then we can see that the
interface reflection, though of the same magnitude, will change in sign when going from a medium of high to low impedance (Z1 > Z2), versus low to high
impedance (Z1 < Z2). This effect is illustrated in Figure 2.8. Thus, RF data can provide information about the direction and magnitude of impedance change across
an interface, whereas the demodulated signal provides information regarding the magnitude only. RF data processing is now the preferred approach for extracting
subtle characteristics regarding tissue microstructure, and many manufacturers of sophisticated ultrasound systems used in specialties outside of ophthalmology
now incorporate RF signal capture.
Figure 2.8. Radiofrequency data derived from indicated (horizontal line) vector segment of 50-MHz scan of
cornea. Note that the phase of the posterior corneal surface echo (right arrow) is inverted in relation to that of
the anterior surface echo (left arrow). This effect arises from the opposite signs of the reflection coefficients of
the surfaces, which is positive anteriorly [(1640-1540)/(1640 + 1540) = 0.031] and negative posteriorly
[(1540-1640)/(1540 + 1640) = -0.031], where the speed of sound in cornea is 1,640 msec-1, and the saline
coupling medium and aqueous are 1,540 msec-1.
SCAN CONVERTER AND DISPLAY

Early ultrasound instruments used analog devices, such as oscilloscopes or television monitors, as display devices. This has been virtually superseded by digital
displays. An analog-to-digital converter (ADC) is needed to use digital display devices.
Analog-to-digital conversion involves transformation of a continuous signal into a discrete binary representation of the kind used in computers. The ADC samples
the output of the amplifier at a specific rate, which is normally at least twice the highest frequency component present in the signal. Although the original analog
signal is unconstrained in its amplitude values (within the dynamic range of the amplifier), ADC limits the range and number of levels of data that can be acquired.
An 8-bit ADC, for instance, can store up to 28 = 256 voltage levels. If a voltage level exceeds the range of the ADC, then it is “clipped.” At present, 8-bit digitizers
are most common in ophthalmic ultrasound systems, but deeper bit-depth components are available. A 12-bit digitizer, for instance, can represent 4,096 levels. As
deep bit-depth digitizers become incorporated in instruments, the need for compression, which always involves some information loss, will decrease.
For digital display of ultrasound data, a special type of ADC, the scan converter, is used. The scan converter
P.32
holds a series of memory locations that map directly to pixels of the digital display. The primary function of the scan converter is transformation of B-mode data into
an image representation in digital memory. This operation includes a remapping of every displayed pixel to specific scan vectors and ranges, such that proper
image geometry is obtained. With each successive scan frame, scan converter memory is overwritten and the new information displayed on the monitor. When the
freeze frame button of the scanner is pushed, the last update of the scan converter memory is continually shown, allowing prolonged examination of individual
frames and storage as computer files. Some systems also include a cine-loop feature in which not one but several successive frames, or even short segments of
real-time display, are stored in digital memory. This feature allows segments, usually a few seconds in length, of the real-time exam to be reviewed and individual
frames to be chosen and stored.
Most systems also include a digital-to-analog converter, whose purpose is to convert the digital image information into standard video formats, such as NTSC or
S-Video that allow the exam to be recorded on analog devices, such as printers and video recorders.
Most modern instruments use digital displays similar to those used in general purpose computers. Much work by computer designers has gone into high-speed
display of images for the very lucrative computer game market, and the ultrasound manufacturers have taken advantage of this technology to develop digital
systems that can display images at high refresh rates, with high resolution and with many colors. Digital ultrasound images are typically 512 X 512 pixels in
dimension and represent 256 shades of gray.
BANDWIDTH
The electronic components through which the ultrasound signal is processed before being displayed have a property known as bandwidth, which refers to the
range of frequencies that a component can process without distortion (usually defined as within a 3 dB range of a perfectly linear response). The echo data
undergoing amplification contain a broad spectrum, or bandwidth, of frequency components that must be accommodated by the amplifier and other components.
The bandwidth of a pulse is approximately equal to the reciprocal of its duration. If this bandwidth is reduced by a hardware component having a smaller bandwidth,
the amplified echo data will be stretched in time, and axial resolution will suffer. Amplification of 0.15-µsec pulses, therefore, necessitates a bandwidth of at least 7
MHz. Failure of one or more components to achieve adequate bandwidth is analogous to listening to a fine classical music recording on a system with a cheap
amplifier and/or speakers. In general, an instrument manufacturer will choose components with just sufficient bandwidth for the application, because this is
cost-effective and because as bandwidth is increased, noise levels tend to increase. As a consequence, one cannot simply plug a 20 MHz probe into a system
designed for a 10-MHz transducer, because (among other things) the bandwidths of one or more components may be inadequate.
A-MODE SYSTEMS
A-mode is the most fundamental ultrasonic modality and forms the basis for more complex modes of operation. The A-mode format consists of a plot of signal
amplitude versus range in a single line of sight. As such, A-scan can be performed with a dedicated A-mode instrument (Figure 2.9) or a combined A/B-mode
instrument. A-mode displays can also be produced using a B-scan instrument by displaying the envelope along individual vectors comprising the B-mode display.
A-mode systems exist in ophthalmology because of their special role in determining biometric properties of the eye, such as axial length, which play a crucial role in
surgical planning (i.e., correct lens implant powering). In design, the A-mode system does not require a motorized probe and, hence, a servo system or scan
converter. A-scan units require a high-resolution display to present the A-scan plot, but they do not need to provide many gray levels; even a simple bilevel display
will suffice.
For A-mode, the probe typically consists of an unfocused or weakly focused transducer. This is then placed in contact with the eye either directly (usually after
application of a drop of a topical anesthesia) or indirectly, using a fluid standoff such as can be established using an eye cup. Most systems for axial measurements
with
P.33
dedicated A-scan probes provide a fixation light in the center of the transducer that allows the transducer to be aligned with the visual axis.
Figure 2.9. A-scan instruments are designed to determine axial length and perform lens-power calculations.
Their simplicity allows compact design, with the most modern instruments being virtually handheld.
Figure 2.10. Typical A-scan printout for axial length measurement. Left to Right: Echoes represent cornea,
anterior and posterior lens, and retina. Note automatic determinations of anterior chamber and lens thickness, in
addition to axial length.
The most common use of A-scan is for axial length measurement (Figure 2.10). These units generally operate at a 10-MHz center frequency with weakly focused
transducers: a 5-mm aperture and 20-mm focal length are typical. A-scan probes may be either handheld or attached to a slit-lamp type mount. During operation,
excitation pulses are emitted at a high PRF, such that many waveforms are captured per second. A-scan units typically incorporate pattern recognition software
that detects the shape of A-scans and automatically acquires those that meet criteria expected when the probe is properly aligned. These data can then be
analyzed for measurement of axial length, anterior chamber depth, and lens thickness. Numerous formulas (4, 5, 6, 7) have been developed for computation of
lens implant power, based on these measurements and keratometry readings. These computations must take into account speed of sound in special instances,
such as a cataractous lens, aphakia, and so forth. A-scan instruments also allow measurements to be made manually from the A-scan trace, which may be
necessary in circumstances of unusual ocular anatomy.
Corneal pachymetry (derived from the Greek pachy, meaning “thick,” and metron, meaning “measure”) (8) is a form of A-scan, whose aim is to measure corneal
thickness, which is, in principle, no different from axial length measurement. In the case of corneal pachymeters, a transducer frequency of 20 MHz and aperture
less than 2 mm are typical, although units operating as high in frequency as 50 MHz are available. Data over a range of about 1.5 mm are plotted. Pachymetry
probes, like axial length probes, can be either handheld or mounted. Pachymetry can also be performed intraoperatively in LASIK to measure the thickness of the
residual stroma (9). A system whose bandwidth is sufficient for processing 20-MHz data can readily process 10-MHz data as well. In fact, some A-scan instruments
incorporate both types of probes. In such a case, the manufacturer will provide excitation pulses and filters suitable for each transducer configuration, along with
software suitable for performing the appropriate biometric determinations.
A special purpose A-scan probe is used in an ophthalmic ultrasound technique called standardized echography (10). Standardized echography units incorporate an
unfocused 8-MHz A-scan probe and a focused 10-MHz B-scan probe. In practice, the A-scan probe is placed in direct contact with the globe and used for
identification of intraocular tumors and differentiation of retinal detachment from vitreous membrane, among other applications. Standardized echography units use
an amplification curve that is S-shaped (a form of compression) to place the dynamic range of the system in the mid-range of echo amplitudes. Calibration is
performed by measurement of the reflections produced by a standard tissue model, thus linking the level of the S-shaped amplification curve to known decibel
levels.
A-scans can also be derived from B-mode images. If we recall that a B-mode image is simply a composite of many consecutive spatially offset vectors, then it is
apparent that we can readily obtain a plot of the amplitude envelope as a function of range along any of the vectors comprising the B-mode image. For this to be
quantitative, however, it is necessary to compensate for the effects of gain, TGC, and any other transformations used in generating pixels from echo data. An
A-scan generated in this manner would have the advantage of providing quantitative information in the context of the two-dimensional B-scan. For axial length
measurements, use of A-scans generated from B-mode data is generally not ideal, because scans are obtained (in contact mode) through the closed lid. Even if an
immersion technique is used, the B-mode probe does not provide a fixation light to ensure that the scan is aligned with the visual axis. However, in special
situations, for example, in eyes with irregular contours (staphyloma), A-scans derived from the B-mode image are the most useful, because they allow selection of
the anatomically appropriate vector (11).
B-MODE SYSTEMS
B-mode systems (Figure 2.11) combine transducer scanning and signal processing to produce cross-sectional images of the eye and orbit. The quality of these
images depends upon the factors already discussed for A-mode systems. However, there are additional electronic,
P.34
mechanical, and acoustic considerations that should be understood for optimal clinical use and proper diagnostic interpretation. Furthermore, B-mode images are
susceptible to several types of artifacts that are readily recognized and that can often be eliminated. In addition to treating these topics, the following sections
discuss useful image enhancement techniques and describe color-coding of pixels and three-dimensional imaging.
Figure 2.11. B-scan instruments incorporate a monitor for display of real-time gray-scale B-mode images. Some
B-mode instruments incorporate A-scan probes and scan modes suitable for axial length measurement as well.
(Photo courtesy of Innovative Imaging, Inc.)
B-MODE IMAGE GENERATION

Conceptually, the simplest B-mode system uses a linear transducer scan motion, as shown in Figure 2.12. As the transducer is moved perpendicular to its beam
axis, it emits acoustic pulses and receives echoes at a series of equally spaced positions. Each of these positions is referred to as a vector. Ultrasound systems
are generally designed so that the distance between adjacent vectors is smaller than the transducer beam width in the focal plane (focal ratio × wavelength). If a
larger distance is used, then the vectors are too sparse to guarantee that all anatomic structures will be captured in a single scan or sweep.
Figure 2.12. B-mode images are formed by scanning the eye by physical movement of the transducer. In the
diagram, a simple linear transducer motion is used. The image is formed by slaving the display to the transducer
position and range to each echo.
Although the scan motion is continuous, it is convenient to consider what happens at each vector location. Echo signals are processed, as in A-mode operation.
However, the resulting video signals are not used to generate a plot of echo amplitude versus range, as in A-scan. Instead, these signals are used to regulate the
intensity of the display. As described previously, this is accomplished by the scan converter, which is a rectangular array of digital memory locations having a
one-toone correspondence with each pixel on the display device. In the case of a linear scan motion, there is a very simple relationship between these locations
and the signal envelope. If we treat the scan converter memory as a rectangular array, then one axis corresponds to range, and the other to transducer position,
with appropriate scaling factors in both cases. Because the scan converter may have more lines than vectors, interpolation is performed to fill in these otherwise
empty memory locations. Thus, the display presents tissue reflectivity in terms of brightness as a function of time or, equivalently, distance from the transducer.
Thus, pixels indicate the two-dimensional position of reflective surfaces in the scanned tissue. In these images, sharply demarcated boundaries (e.g., the anterior
lens surface) that generate well-defined A-mode echoes are displayed as distinct surfaces; acoustically homogeneous regions (e.g., the normal vitreous) are
displayed as dark areas; acoustically heterogeneous areas (e.g., the orbital fat) that generate many closely spaced A-mode echoes are displayed as
correspondingly speckled brightness patterns.
An inherent assumption in B-mode image generation is that of a constant speed of sound value throughout the entire image. This assumption facilitates generation
of the image by allowing use of a single scaling factor to produce a geometrically correct image. However, the speed of sound varies throughout the eye, being
significantly higher, for instance, in the cornea, sclera, and certain tumors than in the vitreous. This difference results in anamorphic distortion, which, although
subtle, can introduce error into biometric determinations. Anamorphically corrected biometry can be performed by taking into account the depth and velocity of each
tissue traversed, as well as vector location.
SCAN PATTERNS
B-mode systems can use various scan patterns, as shown in Figure 2.13. The most useful patterns are those in which the ultrasonic beam is perpendicularly
aligned with reflective tissue surfaces. With perpendicular alignment, echoes travel directly back to the transducer rather
P.35
than being redirected along a path that bypasses the transducer. Linear scan patterns can achieve perpendicularity only over small segments of curved ocular
surfaces, such as those of the retina; therefore, they provide images of limited portions of the eye. Sector scan patterns are more compatible with these curved
surfaces and allow echoes from large segments of posterior surfaces to be captured for B-mode presentations. Arc scan patterns permit perpendicular alignment
over the anterior surfaces of the eye and lens. Of these patterns, the sector scan is the most widely implemented in ophthalmology. The sector scan is readily
implemented in a sealed, compact probe, with a small footprint suitable for contact examination of the eye. Additionally, the sector scan involves the least physical
motion of the transducer, and thus most easily achieves high scan repetition rates. The arc scan, however, is most suited for visualization of the anterior segment
with very high frequency ultrasound (12,13). With a transducer frequency of about 40 MHz, examinations must be performed using an immersion technique with a
lid speculum because of the effect of attenuation by the eyelid. In a noncontact exam, the arc scan is readily implemented and is ideal for visualization of the
cornea and other anterior segment structures, as illustrated in Figure 2.14.
Figure 2.13. Several mechanical scan modalities have been used. The sector scan is most popular in that it is
compact, most amenable to high scan rates, and provides approximate normality to the posterior retinal surface.
The linear scan is the most simple conceptually and has the advantage that vectors do not diverge with range.
The arc scan is complex in implementation but offers near normality to both the anterior and posterior surfaces
of the eye.
B-MODE IMAGE QUALITY

Under ideal conditions, the pixel intensities in B-mode images correspond precisely to the acoustic reflectivity at each tissue point. In practice, these
representations are constrained by the limited intensity ranges of display devices, including computer monitors, thermal printers, and video printers. Current digital
display devices devote a maximum of 1 byte (8 bits) to each of three colors: red, green, and blue. This allows millions of colors to be displayed at once but only 256
levels of each color individually. Because shades of gray are composed of pixels with equal intensity values of red, green, and blue, only 256 shades of gray are
available. This corresponds to a 24-dB dynamic range. Effective dynamic range can be increased by prior logarithmic amplification or other compression modes;
however, it is often most expedient to use B-mode images for assessments of general anatomy and to obtain A-mode results along specifically chosen directions
for quantitative reflectivity information.
Figure 2.14. High-frequency (50 MHz) scan of anterior segment produced using an arc-scan geometry. This
scan geometry maintains near-normality relative to the anterior surfaces of the globe, allowing display of the full
corneal contour.
The spatial resolution achievable in a B-mode system is limited both by acoustic constraints (frequency, focal length, aperture, and so forth) and by the pixel
resolution of the display device. Let us consider a system with a 200-micron pulse length in which an image of the eye and orbit 5 cm in depth is displayed over 256
pixels in the axial direction. This means that each pixel represents 195 microns, just sufficient to represent the axial resolution to which we are entitled. However, if
fewer pixels are used to represent the image, or a greater scan depth is displayed in the same number of pixels, the display resolution will be degraded. This effect
is also important in implementation zoom functions in B-scanners. The simplest way to implement a zoom is to double the pixel size. This method, however,
provides no actual increase in image detail. If, however, smaller pixels are derived from the stored data, then finer image detail can be achieved.
The resolution inherent in B-mode images can be limited by large pixel sizes, but these are usually not the limiting factors, and resolution is most often governed by
the same considerations that determine A-mode resolution. Axial resolution is determined by the duration of the ultrasonic pulse; thus, excessively long pulses will
cause apparent thickening of interfaces in the image and prevent detection of closely spaced surfaces.
P.36
Figure 2.15. A wire phantom allows characterization of beam width as a function of range. Note that wire
reflections in the region of the focus (dashed line) are less smeared than those in the near or far fields. Also
note that even in the focal zone, the wire appears somewhat elongated in the scan direction because the beam
width of focused beam is finite even at the focal point.
Lateral resolution is determined by the ultrasonic beam width. Wide beams exaggerate the apparent width of reflective structures in a manner that depends upon
the scan pattern being used. To examine this effect, consider a sector scan across an array of long, thin wires aligned normally to the scan axis, as illustrated in
Figure 2.15. Wires in the focal zone appear almost as points, because here the beam width reaches its minimum. Anterior or posterior to the focal plane, the
apparent width of the wires is exaggerated, and they take on an arc-shaped appearance as a consequence of the sector scan geometry and the broad beam width,
which causes detection of the wires across several adjacent vectors.
Figure 2.16. Tissue-mimicking phantoms are widely used to evaluate an ultrasound system's capacity to detect
wire targets, reflective and cystic structures embedded in a scattering background. Evaluation of a 10-MHz
transducer using a small-parts phantom (Radiation Measurements, Inc., Middleton, Wisconsin) is shown.
Wire targets are one type of “phantom” that can be used to characterize B-mode image quality. Several manufacturers offer ultrasound tissue phantoms suitable for
transducers of specific frequency ranges. Although no eye phantoms are offered commercially, small parts phantoms (Figure 2.16) can be useful in determining a
system's capacity to visualize cystic and echogenic targets of various sizes and contrast in relation to background.
Just as in A-mode operation, absorption limits the resolution attainable with B-mode systems. High-resolution images of the posterior segment can be obtained at
20 MHz (14), but only a thin layer of the retro-ocular orbit can be penetrated at this high frequency. Deeper orbital penetration requires lower frequencies (i.e., 5 to
10 MHz).
REAL-TIME IMAGING
During scanning, B-mode images are generated at a rate equal to the number of scans per second performed by the probe. Early mechanical sector scanners
provided perhaps four scans per second, but modern scanners can offer scans at 30 Hz or higher. This, essentially, offers real-time evaluation of ocular tissues.
Real-time imaging has particular value in evaluation of vitreous membranes, retinal detachment, and vitreous hemorrhage. It can also be useful in evaluation of
tumors by allowing visualization of vascular pulsatility. Real-time examinations can be captured using cine-loop (where available) or by attaching a video recording
device to the analog output of the B-scanner.
B-MODE ARTIFACTS
B-mode images are susceptible to artifacts resulting from ultrasonic and electronic sources. The most commonly
P.37
encountered artifacts are listed in Table 2.1 and are described later.
TABLE 2.1. Types of B-mode Artifacts

Source Effects
Acoustic Artifacts
Velocity differences Displacement artifact Contour distortion
Absorption Shadowing
Multiple reflections Surface duplication
Electronic Artifacts
Noise “Snow”
Saturation Obliteration of texture
Saturation (occurring with texture enhancement) “Swiss cheese” artifact
Inadequate superposition Blurring and duplication
B-mode artifacts can arise because of differences in the propagation velocities of various tissues. For example, Figure 2.17 illustrates distortions stemming from
the relatively high velocity of the crystalline lens. Along a central path OA through the lens, the rear wall appears to be displaced anteriorly because the high
lenticular velocity decreases the transit time from the transducer to point A. (This shortening is also present on an A-scan.) In addition, scan paths passing
obliquely through the lens (for example, OB) subject the ultrasonic pulse to refraction so that the point actually being imaged does not lie along the transducer axis.
On the other hand, paths bypassing the lens (OC) result in undistorted imaging. The overall effect of these phenomena is to distort the contours of tissues located
behind the lens.
Another type of artifact, acoustic shadowing, decreases the image light intensity in tissue regions posterior to highly absorptive structures, such as the lens and
certain types of tumors. An example of shadowing by a dislocated hypermature cataractous lens is provided in Figure 2.18. Shadowing often facilitates differential
diagnosis by allowing the clinician to categorize tumors according to their absorptivity. Because of these effects, the most accurate results are obtained only when
the transducer scan paths do not traverse the lens. Carefully oriented scans through the sclera result in only minimal degradations from velocity and absorption
effects.
Figure 2.17. Diagrammatic illustration of distortion of the posterior contour of the eye when imaged through the
lens. Because the speed of sound of the lens is higher than that of vitreous, more distal structures appear closer
than they really are (A versus A'). In addition, because of the convex shape of the crystalline lens and its
relatively high speed of sound, refraction causes the beam to diverge (B versus B') when it passes obliquely
through the lens.
Figure 2.18. Dislocated hypermature cataractous lens in an eye with vitreous hemorrhage and total retinal
detachment. The lens material is highly acoustically absorptive, resulting in an acoustic shadow trailing from the
lens.
Multiple acoustic reflections constitute another source of artifacts, introducing duplication of tissue contours, as shown in Figure 2.19. In this immersion scan,
ultrasonic echoes from the cornea and lens implant return to the transducer, where they are partially reflected back toward
P.38
the eye. These echoes are then reflected by the cornea and arrive for a second time at the transducer after the transit time determined by the transducer-cornea
separation. The multiply-reflected echoes appear in both A-mode signals and B-mode images, where they usually appear as phantom surfaces within the vitreous
or in posterior regions. Recognition of multiple reflections is straightforward: changing the transducer-cornea standoff distance alters the location of the artifacts
with relation to the other structures of the eye. These artifacts can be eliminated by making this standoff distance equal to the maximum tissue depth to be
examined. Reduplication artifacts also occur in contact scans, although they are less common.
Figure 2.19. Artifactual duplication of cornea (solid arrow) and lens implant (dashed arrow) in central vitreous.
Electronic artifacts can assume several forms. “Snow” can appear on B-mode images, if amplifier gain is high and electronic noise is not rejected prior to display.
Saturation can cause heterogeneous structures, such as orbital fat, to appear as uniformly bright areas. Recognition of these artifacts is aided by careful monitoring
of A-mode signals.
DIGITAL IMAGE PROCESSING

Digital storage of B-mode images confers great advantages in postprocessing. This digital representation allows application of various digital image processing
methods to enhance images. Most instruments include a set of simple operations, such as brightness and contrast adjustment and, possibly, a zoom function. An
entire literature exists regarding digital image enhancement (15), and these techniques are readily applied to ultrasound B-mode images stored in a generic format,
such as TIFF or JPEG. Examples of relatively useful and straightforward operations include modification of the pixel brightness intensity curve, thresholding,
blurring, and median filtering, among others (Figure 2.20). Such operations can be performed with available software, such as Photoshop or NIH-Image.
In addition, a color scale can be substituted (pseudocolor) for the usual gray scale in representing pixel brightness. The use of color provides increased contrast in
comparison with gray-scale. Color display is the default display mode in optical coherence tomography, for example. Color display, however, has not achieved
widespread acceptance in B-mode ultrasound imaging because color scales are essentially arbitrary and can conceal as much as they reveal, if not used
judiciously.
THREE-DIMENSIONAL IMAGING
Three-dimensional (3-D) imaging is made possible by digital storage of images. To form a 3-D image representation,
P.39
an ordered series of B-mode images must be acquired and stored (16, 17, 18). Various 3-D scan geometries are illustrated in Figure 2.21. Conceptually, the
simplest way to do this is to move the transducer in a linear fashion to sweep out a rectilinear B-mode image, move the transducer incrementally at right angles to
the sweep direction, and repeat. This way, scans are stacked like a pack of cards. There are other 3-D scan modes. For instance, a mechanical sector probe can
be moved linearly at right angles to the scan plane, or a sector probe can itself be sectored to sweep out a fan-shaped region. One commercially available 3-D
ophthalmic scanner (Ophthalmic Techologies, Inc., Toronto, Ontario) rotates a mechanical sector probe along its axis to sweep out a cone-shaped region.
Three-dimensional image data are rendered (Figure 2.22) with special purpose software. In addition to this general methodology, 3-D data can be acquired
freehand using a probe in which sensors monitor the probe position and orientation (19). From this information, the position of each pixel in space can be computed
and, using appropriate interpolation methods, a 3-D image can be generated. The 3-D renderings can be rotated, translated, zoomed, and sectioned, allowing
additional information to be extracted from the data. In addition, 3-D allows quantitative information to be determined, such as surface areas and volumes, that may
be useful for following tumors and other volume occupying pathologies. It should be understood, however, that 3-D image reconstruction is still subject to the same
principles as 2-D B-mode imaging. For instance, if a structure fails to provide a high amplitude echo as a result of oblique presentation in a single B-mode image,
this will not be improved by taking a series of parallel B-mode slices at the same oblique incidence.
Figure 2.20. Image processing enhancement of digitized B-mode images can be performed using a
variety of proprietary and public domain software. This figure shows application of sequential image
processing operations to a B-mode image of an eye with orbital mass. Top to bottom: Operations
degradation by addition of noise, Gaussian smoothing, thresholding, median filtering.
Figure 2.21. Three-dimensional imaging can be performed using a variety of scanning geometries, including
serial rectilinear (left), sequential sector (center), and meridional rotational (right), among others.
Figure 2.22. Examples of rendered 3-D ultrasound images. Upper left: This surface-rendered image was
derived from a series of parallel scans of an eggshell fragment resting on the retina. Produced by Silverman and
Coleman in the early 1980s, it is, to our knowledge, the first opthalmic 3-D ultrasound image. Upper right: Early
wire mesh surface rendering with hidden surface removal of a small choroidal melanoma, with computed values
of tumor dimensions. Lower left: Shaded surface rendering of a large choroidal melanoma with secondary
retinal detachment. Lower right: Volume rendered image of total retinal detachment. (see color image)
P.40
VERY HIGH FREQUENCY ULTRASOUND/ULTRASOUND BIOMICROSCOPY

The terms ultrasound biomicroscopy (UBM) and very high frequency ultrasound (VHFU) have generally been taken to refer to use of frequencies of 25 MHz or
higher. B-mode VHFU images do not differ fundamentally from those generated using conventional 10-MHz transducers, but the application of these high
frequencies impacts upon what can be imaged and certain technical aspects of instrument design.
A number of technologic advances in the early 1990s made VHFU possible, including new transducer materials (both polymer and crystalline) and new, reasonably
priced, broadband electronics, including high-speed digitizers. Pavlin and Foster (20, 21, 22) described clinical findings with a 50-MHz ultrasound biomicroscope
that was later developed into a commercial instrument, the ultrasound biomicroscope, or UBM (Paradigm Instruments, Salt Lake City, Utah). Our laboratory
independently developed a series of scan platforms for ophthalmic imaging at 50 MHz that included features, such as 3-D acquisition of radiofrequency data and
wide-angle scanning incorporating the entire anterior segment (12,13,16).
Because of the effect of attenuation, VHFU cannot be used for imaging of the posterior segment but can generate superb images of anterior segment anatomy and
pathology, such as corneal scars (including effects of refractive surgery), tumors and cysts of the iris and ciliary body, ciliary body detachment, glaucoma
syndromes (e.g., pupillary block), and hypotony. At a frequency of 50 MHz, we can expect a fivefold improvement in resolution compared to that of 10-MHz images,
with axial and lateral resolutions of about 30 and 60 microns achievable (depending upon specifics of transducer pulse length and focal ratio).
To image the anterior segment, VHFU scans must be performed using an immersion technique, with the attenuating eyelid absent from the acoustic path. This can
be accomplished using an eye cup or by forming a water-bath with a disposable surgical drape, usually in combination with a lid speculum. The Artemis-2 system
(Ultralink, LLC, St. Petersburg, Florida) uses a disposable eyepiece consisting of a viscoelastic foam ring that forms a seal around the eye. Acting like a reverse
swimming goggle, normal saline is introduced into the eyepiece to establish acoustic coupling. This system has the additional advantage of allowing optical
visualization of the eye during scanning by use of a coaxial video camera.
In addition to the UBM and the Artemis, other manufacturers have introduced cost-effective VHFU instruments with handheld 35-MHz probes.
Figure 2.23. High-frequency M-mode images of an iris vessel in a rabbit's eye, taken over about four cardiac
cycles. Left: In grayscale image, stationary structures remain at constant range, whereas flowing blood particles
change in range with time. Right: Colorized image of same data illustrates pulsatile flow, with some
regurgitation during diastole. (see color image)
M-MODE
M-mode (Figure 2.23) represents a cross between A- and B-modes. As in A-mode, the transducer interrogates a single line of sight, but as in B-mode, a
two-dimensional image is formed. In M-mode, however, the vertical axis represents time rather than lateral position (as it does in B-mode). M-mode is useful for
demonstration of tissue motion. Stationary tissue structures will maintain a constant range from the transducer, so echoes will appear vertical on the screen. Where
tissue motion occurs (i.e., vessel wall motion) the range will vary with time, and this will be evident in the M-mode image. M-mode is not generally available on
commercial ophthalmic systems.
SWEPT-MODE
Swept-mode combines M-mode and B-mode (Figure 2.24) (23,24). In conventional B-mode imaging, vectors are usually placed a beam width apart or less. In
swept-mode, vectors are placed much less than a beam width apart. The advantage of this is that groups of adjacent vectors within a beam width of each other can
be treated as viewing the identical spatial position over time (time being related to the pulse repetition frequency), while vectors that are more than a beam width
apart allow
P.41
formation of a conventional B-mode image. Thus, a swept-mode image is equivalent to a B-mode image composed of overlapping M-scans. Swept-mode is thus
capable of showing tissue motions, including blood flow, in the context of a B-mode image. The advantage of swept-mode in comparison to Doppler is that flow
information is obtained at the same high resolution as the underlying B-mode data. Disadvantages include lower sensitivity and slower frame rate.
Figure 2.24 Swept-mode is, essentially, a B-mode image in which vectors are spaced much closer together
than the beam width. Because of this, groups of adjacent vectors within a beam width of each other are not
spatially independent. This allows treating these groups of overlapping vectors as local M-scans. The uppermost
50-MHz image of the angle region in a rabbit's eye was constructed from 128 vectors spaced 18 microns (about
4 vectors/beam width). This has the appearance of a conventional B-mode image. The center image is of the
same tissue but with 1,024 vectors spaced 2.2 microns apart (about 30 vectors/beam width). In this highly
oversampled case, we can see areas where the echo phase is decorrelated compared to surrounding tissues
(arrows). This effect results from blood flow, where blood cells change in range over time. Based on the PRF
and the change in range per vector, flow velocity can be computed. The bottommost figure is a color-flow image
generated from the digitized echo data. (see color image)
LINEAR ARRAY SYSTEMS

At the time of this writing, ophthalmic ultrasonography is almost universally performed using mechanical sector scan probes, a technology that has almost
disappeared outside of this specialty. General purpose ultrasound instruments rely on linear array transducers for B-mode imaging. This distinction is attributable to
a number of factors. Ophthalmic ultrasound is performed at frequencies that are generally higher than those used in other specialties, with a few exceptions.
Fabrication of arrays and control circuitry become more difficult and expensive as frequency increases. These factors, taken in the context of the relatively small
ophthalmic ultrasound market, have kept ophthalmic ultrasound out of the technologic mainstream.
Small-parts linear array probes with a center frequency of 10 MHz or more are available. One should consider the advantages of linear array technology. These
include:
High frame rate

Large effective aperture—improves lateral resolution
Dynamic focusing—movable/multiple synthetic focal zones
Special scan modes
Continuous wave Doppler
Color flow Doppler
Power Doppler
Tissue harmonic imaging
The trend toward decreasing costs, greater compactness, and increasingly higher available frequencies for linear array systems suggests that this technology is
likely to have an increasing impact on the performance of ophthalmic ultrasonography in the future.
OTHER SCAN MODES
DOPPLER MODES
The Doppler effect (25) is well known: If a sound source is moving toward the listener, the wavelength is compressed, and the pitch increased (Figure 2.25). The
opposite effect occurs when the sound source moves away from the listener. This effect has been used in ultrasound systems for measurement and visualization of
blood flow and in ophthalmology for visualization and quantification of flow in the orbital vessels and tumors (26). The Doppler frequency shift, fd, is, by definition,
the difference between the emitted (fe) and received (fr)
P.42
frequencies, that is, fd = fe-fr. It is determined by fd = fe[2v ÷ (c-v)], where c represents the speed of sound, and v represents the component of the velocity of the
scatterer (e.g., blood cells) along the transducer beam axis. For instance, if a vessel has a flow velocity of 10 cm sec-1, then the Doppler frequency shift is 10 ×
106[0.02 ÷ (1540-0.01)] = 130 Hz for a 10-MHz source. Notice that this frequency is in the audio range. Doppler systems provide an audio output that allows the
sonographer to “hear” flow.
Figure 2.25. The wavelength of reflections from a particle in motion toward the transducer is shortened.
In Doppler ultrasonography, our interest is detection of frequency shifts associated with tissue motion. As such, Doppler becomes more sensitive as transducer
bandwidth is reduced or, equivalently, as pulse duration is increased. This means that as Doppler resolution increases, spatial resolution decreases.
The most basic Doppler mode is “continuous wave” (CW) Doppler. In CW Doppler, the pulser is replaced by an oscillator that produces a continuous sine wave
voltage that excites the transducer. Because this transducer is exclusively generating a continuous emission, a separate transducer is used to receive echoes.
Alternatively, in linear array systems, one subset of elements can be used to emit, while another set acts as the receiver. Color-flow or duplex Doppler (Figure 2.26)
involves a simultaneous display of a B-mode image with superimposed color information indicating areas of flow. In color-flow, a subset of elements in a linear
array emits pulses several cycles in duration. This allows measurement of Doppler shift as well as range and direction simultaneously (but with lower spatial
resolution than the underlying B-mode image and with reduced Doppler sensitivity compared to CW). The color-flow image can be used to select a vessel for CW
interrogation of the Doppler waveform, as shown in Figure 2.27.
Doppler electronics (Figure 2.28) differ in some ways from that of A- and B-mode systems. After the echo data are amplified, they are processed by a component
called a mixer. The mixer multiplies the amplified echo waveform by the excitation sine waveform. This provides a signal in which the sum (fe + fr) and the
difference (fe-fr) of the frequency components of the two inputs are combined. This signal is then bandpass filtered to remove the summation component, leaving
only fe-fr, which is,
P.43
of course, equivalent to fd, the Doppler frequency shift. It may also be necessary to perform a high-pass filter operation of the Doppler signal to remove extraneous
signal components associated with tissue motion. The combined amplifier and filter components are referred to as a demodulator. Because fd is in the audio range,
this signal is then amplified with an audio frequency amplifier.
Figure 2.26. Color-flow Doppler image of a normal human eye. The scan was taken with the transducer in a
vertical orientation such that superior is toward the right. The image shows the central retinal vessels as well as
the ciliary artery just superior to the nerve. (see color image)
Figure 2.27. The CW Doppler waveform associated with a vessel is obtained by choosing the area of interest
on a static color-flow image and adjusting the measurement angle (to perform cosine correction) where
necessary. (see color image)
Figure 2.28. Schematic drawing representing the electronic components used to generate directional color-flow
information.
The previously shown arrangement, however, provides only the magnitude of the Doppler frequency shift. Because we are often interested in the direction of flow
as well, a somewhat more complicated electronic processing method called phase quadrature detection is used. In this method, the echo signal is passed to two
separate demodulators, the first of which (as before) multiplies the signal by the excitation waveform, and the second multiplies the echo by an excitation waveform
that has been rendered 90 degrees out of phase. These are referred to as the in-phase and the quadrature signals, respectively. The in-phase and quadrature
channels allow determination of whether fd is positive or negative; if fd is positive, the quadrature channel lags behind the in-phase channel by a phase shift of 90
degrees, whereas, if fd is negative, then the quadrature channel is 90 degrees advanced, in respect to the in-phase channel.
CW Doppler systems provide a graphic representation of flow by converting the Doppler signals into positive and negative velocity values. These are plotted over a
period of several cardiovascular cycles to provide a good representation of the systolic and diastolic blood flow pattern in a vessel.
In color-flow Doppler systems, colored pixels representing flow are superimposed onto the B-mode image. A color-scale, usually ranging from reds (representing
arterial flow, i.e., flow toward the transducer) to blues (venous flow) is presented on the display. The sonographer can adjust several parameters to optimize the
color-flow presentation, including the range of velocities to be displayed, the write priority of color flow versus gray-scale information, and filtering functions (Wall
filters) used to suppress Doppler shifts associated with motions of solid tissues, such as vessel walls or respiratory motions. In addition, the user can generally
modify the PRF used for acquiring the Doppler signal. This is significant in that the highest Doppler frequency that can be accurately characterized is one half of the
PRF. If this is exceeded, then a phenomenon called aliasing occurs.
Color flow is advantageous because vessels (in the scan plane) are seen in the context of the B-mode image, which facilitates identification. Also, color-flow
imaging allows estimation of the angle of the vessel in relationship to the acoustic beam axis. Because Doppler systems can provide only a measurement of the
velocity component of flow in the beam axis, a cosine adjustment term must be used to correct Doppler velocity values, vd = v cos(?), where ? is the angle between
the transducer axis and the flow direction and, vd is the uncorrected Doppler velocity value. Color-flow Doppler systems allow the user to indicate vessel orientation
so that the cosine correction term can be applied to the uncorrected Doppler velocity values.
Power Doppler (27) is also provided on most instruments to perform color-flow Doppler. In Power Doppler, the Doppler frequency shift signal is integrated. This has
the effect of removing directional and velocity information but provides a color-flow map of perfusion that has a significantly higher sensitivity than conventional
color-flow Doppler and less sensitivity to angular orientation. Power Doppler is particularly useful in a situation of slow-flow and tortuous vasculature, as in some
tumors.
TISSUE HARMONIC IMAGING

Tissue harmonic imaging (THI) was developed in the late 1990s (28). It was discovered serendipitously as a consequence of attempts to develop a means for
improved detection of flow, using ultrasound contrast agents. Such agents consist of microbubbles, lipid shells filled with air, or other substances with a high
acoustic impedance inhomogeneity compared to blood. It was anticipated that the microspheres would resonate at specific ultrasound frequencies, and that this
would result in emission of echo data from the microspheres at harmonics of the emitted ultrasound frequency. By filtering out the fundamental emitted frequency,
then, tissue echoes would be suppressed, while the harmonic vibration modes of the microspheres would be detectable. Although this process was found to be
valid, early users discovered that even in the absence of contrast agent, tissues were seen with better contrast and clearer borders in images made at the second
harmonic (i.e., double the emission frequency). This effect is a result of nonlinear interaction between the ultrasound pulse and the tissues through which it
propagates. The speed of sound is affected by the density of the material through which it propagates. A sound wave, however, is by definition a pressure wave,
with compressive and decompressive components. Thus, as a sound wave travels through a medium, the compressed component of the
P.44
pulse tends to move slightly slower than the decompressed part of the pulse. This distortion is a function of both the tissues through which the pulse travels and the
overall distance—the effect is cumulative with range. The distortion of the pulse waveform results in generation of harmonics. The reason that this is interesting is
that the harmonic has comparatively small side lobes compared to those of the fundamental. Because side lobes effectively reduce lateral resolution, generating an
image at the harmonic produces more well-defined boundaries.
Tissue harmonic imaging requires a transducer with sufficient bandwidth to capture at least the second harmonic. This is generally accomplished by exciting the
transducer at two thirds of its center frequency and receiving at four thirds of center frequency. The simplest technology for producing tissue harmonic images is by
using bandpass filters centered at the second harmonic that filter out the fundamental. The disadvantage of this approach is that both bands must have fairly
narrow bandwidths, and hence poor axial resolution, to achieve this separation. An alternative approach is pulse-inversion. In this method, two pulses are emitted
in quick succession, one of normal phase and the other inverted. When echoes from the two pulses are acquired and added, the fundamental and all odd
harmonics are eliminated, leaving only the even harmonics. This provides an effective means of capturing and generating images at the second harmonic.
THI is now widely incorporated into linear array imaging systems. In fact, it is so effective that it is often used as the default imaging mode.
SPECTRAL PARAMETER IMAGING

The interaction between an ultrasound pulse and the tissue through which it propagates causes the reflected or backscattered signal to differ from the emitted
signal. An extensive literature exists regarding the effect of tissue microarchitecture on backscatter (29, 30, 31, 32, 33). It is understood, for instance, that as tissue
inhomogeneities become smaller compared to a wavelength, they more effectively scatter the higher frequencies present in the ultrasound pulse (Figure 2.29).
(This is the same physical principle that causes the sky to be blue.) Inhomogeneities that are much smaller than a wavelength become Raleigh scatterers, with
backscatter increasing with the fourth power of frequency. Also known, and intuitively obvious, is that as the number of scatterers per unit volume increases,
backscatter increases as well. However, if the scatterer concentration rises high enough, the scatterers effectively become background instead of foreground, and
the spaces between them behave as the scatterers. The geometric form of scatterers (spherical, filamentous, lamellar) also affects the backscattered signal. In the
case of nonisotropic scatterers (filaments or lamellas), their orientation relative to the ultrasound beam is an important consideration. Clearly, this is a complex
process. Nevertheless, if certain simplifying assumptions are made (e.g., weak scatterers and negligible attenuation), quantitative estimates of scatterer size and
concentration can be derived from measurement of the difference between the power spectrum of the emitted pulse and the received echo, referred to as the
calibrated power spectrum (CPS). Because the CPS is typically quasilinear in appearance, the linear best fit to the CPS is used as a means for characterizing
tissues.
P.45
Mathematical modeling of acoustic backscatter shows that the slope (dB/MHz) of the linear best fit to the CPS is directly related to scatterer size, whereas the
intercept (dB) relates to scatterer concentration and relative impedance. By measuring spectra at successive spatial positions within a B-mode image, estimates of
these quantities can be made and the gray-scale pixel values replaced by colors representing calculated scatterer size and concentration. This approach is used in
ophthalmology to characterize tissues, such as tumors, hemorrhage, and corneal scars. Spectral parameter images (Figure 2.30) provide a visual representation of
the physical properties of tissues and also quantitative values related to the mean value and variation of each parameter that allow comparison with cases of
known pathology or determination of changes occurring over successive examinations.
Figure 2.29. Mathematical modeling of acoustic backscatter has shown that as scatterer size increases from
much less than a wavelength to a half-wavelength, spectral slope (reflected amplitude versus frequency) goes
progressively from positive to negative values.
Figure 2.30. Spectral parameter images generated by determining calibrated spectra along each vector and
replacing the grayscale pixel values representing the envelope of the echo data with color values representing a
spectral parameter. In this case, three kinds of spectral parameter images were generated from a
high-frequency arc-scan of the anterior segment of the eye with hyphema. Upper left: A midband fit image is
presented in grayscale in the correct geometric format. The other three images show images in stretched
rectilinear format. Bottom left: Midband fit. Upper right: Slope. Lower right: Slope. Lower right: Intercept.
Midband fit allows reduction in speckle and other source of noise.
REFERENCES
1. Mundt GH, Hughes WF. Ultrasonics in ocular diagnosis. Am J Ophthalmol 1956;42:488-498.
2. Baum G, Greenwood I. The application of ultrasonic locating techniques to ophthalmology: part 2. Ultrasonic visualization of soft tissues. Arch Ophthalmol
1958;60:263-279.
3. Ossoinig KC. Quantitative echography: the basis of tissue differentiation. J Clin Ultrasound 1974;2:33-46.
4. Sanders DR, Kraff MC. Improvement of intraocular lens power calculation using empirical data. J Am Intraocul Implant Soc 1980;6:263-267.
5. Binkhorst RD. Intraocular lens power calculation. Int Ophthalmol Clin 1979;19:237-252.
6. Holladay JT, Prager TC, Ruiz RS, et al. Improving the predictability of intraocular lens power calculations. Arch Ophthalmol 1986;104:539-541.
7. Hoffer KJ. Pre-operative cataract evaluation: intraocular lens power calculation. Int Ophthalmol Clin 1982;22:37-75.
8. Aslanides IM, Aslanides MN, Reinstein DZ, et al. Have you ever seen a pachyderm [Letter]? J Refract Surg 1995; 11:162-164.
9. Flanagan G, Binder PS. Estimating residual stromal thickness before and after laser in situ keratomileusis. J Cataract Refract Surg 2003;29:1674-1683.
10. Ossoinig KC. Standardized echography: basic principles, clinical applications, and results. Int Ophthalmol Clin 1979;19:127-210.
11. Zaldivar R, Shultz MC, Davidorf JM, et al. Intraocular lens power calculations in patients with extreme myopia. J Cataract Refract Surg 2000;26:668-674.
12. Silverman RH, Reinstein DZ, Raevsky T, et al. Improved system for ultrasonic imaging and biometry. J Ultrasound Med 1997;16:117-124.
13. Reinstein DZ, Silverman RH, Raevsky T, et al. Arc-scanning very high-frequency ultrasound for 3-D pachymetric mapping of the corneal epithelium and
stroma in laser in situ keratomileusis. J Refract Surg 2000;16:414-430.
14. Coleman DJ, Silverman RH, Chabi A, et al. High resolution ultrasonic imaging of the posterior segment. Ophthalmology, 2004;111:1344-1357.
15. Rosenfeld A, Kak AC. Digital Picture Processing. New York: Academic Press; 1982.
16. Cusumano A, Coleman DJ, Silverman RH, et al. Three dimensional ultrasound imaging: clinical applications. Ophthalmology 1998;105:300-306.
17. Silverman RH, Coleman DJ, Rondeau MJ, et al. Measurements of ocular tumor volumes from serial, cross-sectional ultrasound scans. Retina
1993;13:69-74.
18. Romero JM, Finger PT, Rosen RB, et al. Three-dimensional ultrasound for the measurement of choroidal melanomas. Arch Ophthalmol
2001;119:1275-1282.
19. Delcker A, Martin T, Tegeler C. Magnetic sensor data acquisition for three-dimensional ultrasound of the orbit. Eye 1998;12:725-728.
20. Pavlin CJ, Sherar MD, Foster FS. Subsurface ultrasound microscopic imaging of the intact eye. Ophthalmology 1990;97:244-250.
21. Pavlin CJ, Harasiewicz K, Sherar MD, et al. Clinical use of ultrasound biomicroscopy. Ophthalmology 1991;98: 287-295.
22. Pavlin CJ, Harasiewicz K, Foster FS. Ultrasound biomicroscopy of anterior segment structures in normal and glaucomatous eyes. Am J Ophthalmol
1992;113:381-389.
23. Kruse D, Fornaris J, Silverman R, et al. A swept-scanning mode for estimation of blood velocity in the microvasculature [Letter]. IEEE Trans Ultrason
Ferroelectr Freq Control 1998;45:1437-1440.
24. Silverman RH, Kruse D, Coleman DJ, et al. High-resolution ultrasonic imaging of blood-flow in the anterior segment of the eye. Invest Ophthalmol Vis Sci
1999;40: 1373-1381.
25. Wells PN. Ultrasonic colour flow imaging. Phys Med Biol 1994;39:2113-2145.
26. Tanquart F, Berges O, Koskas P, et al. Color Doppler imaging of orbital vessels: personal experience and literature review. J Clin Ultrasound
2003;31:258-273.
27. Macsweeney JE, Cosgrove DO, Arenson J. Colour Doppler energy (power) mode ultrasound. Clin Radiol 1996;51:387-390.
28. Duck FA. Nonlinear acoustics in diagnostic ultrasound. Ultrasound Med Biol 2002;28:1-18.
29. Lizzi FL, Greenebaum M, Feleppa EJ, et al. Theoretical framework for spectrum analysis in ultrasonic tissue characterization. J Acoust Soc Am
1983;73:1366-1373.
30. Insana MF. Ultrasonic imaging of microscopic structures in living organs. Int Rev Exp Pathol 1996;36:73-92.
31. Hosokawa T, Sigel B, Machi J, et al. Experimental assessment of spectrum analysis of ultrasonic echoes as a method for estimating scatterer properties.
Ultrasound Med Biol 1994;20:463-470.
32. Hunt JW, Worthington AE, Kerr AT. The subtleties of ultrasound images of an ensemble of cells: simulation from regular and more random distributions
of scatterers. Ultrasound Med Biol 1995;21:329-341.
33. Lizzi FL, Astor M, Feleppa EJ, et al. Statistical framework for ultrasonic spectral parameter imaging. Ultrasound Med Biol 1997;23:1371-1382.
> Table of Contents > 3 - Ocular Diagnosis
3
Ocular Diagnosis
HISTORICAL BACKGROUND
The first use of ultrasound for ophthalmic diagnosis was reported in 1956 by Mundt and Hughes (1), who used industrial ultrasound equipment to examine
enucleated normal eyes and eyes with intraocular tumors. The first clinical use of A-scan in ocular diagnostic problems was described in 1957 by Oksala (2), in the
first of many pioneering papers. Later, Jansson (3) of Sweden described the use of ultrasound for ocular measurement and made critical in-vitro measurements of
the sound velocity constants for ocular tissues (4). Oksala (5) also made early measurements of sound velocities of tissues. Sorsby (6), using A-scan, compared
axial lengths in a large patient population and described sex and age differences. Coleman and Carlin (7) described the first axial length measurements of the eye,
using an electronic interval counter to make precise axial measurements and to document lens movement in accommodation. Giglio (8) also described axial
measurements, using a similar system.
Oksala's initial work was followed by many subsequent papers on A-scan for clinical diagnosis (9). Later, Bronson (10) developed ultrasonically directed intraocular
forceps for foreign bodies. Ossoinig (11) popularized a specific form of A-scan equipment and described clinical results with many original observations on the
A-scan properties of specific tissues. He developed a sophisticated diagnostic technique that emphasized the quantification of echo amplitudes, using a tissue
standard and “s” shaped amplification of the A-scan, which he termed standardized echography. He also described kinetic A-scans in which movement of both the
transducer and vascular structures is used to characterize tissues (12). This technique remains in widespread use.
Other investigators who have contributed early, original observations to A-scan diagnosis include Buschmann (13) and Gernet (14) of West Germany, Massin and
Poujol (15) of France, Francois and Goes (16) of Belgium, Vanysek and Preisova (17) of Czechoslovakia, Bertenyi (18) of Hungary, and Gallenga (19) of Italy. In
the United States, in the hospital-based laboratory at the Wills Eye Institute, Sarin et al. (20), under the direction of Keeney (21), made early contributions regarding
A-scan evaluation. At the Walter Reed Army Hospital, Penner and Passmore (22) and Cowden and Runyon (23) described the uses of A-scan in the diagnosis of
foreign bodies. Coleman (7) demonstrated high frequency (25-MHz A-scan) evaluation of the choroid to measure the in-vivo thickness of this highly vascular
erectile tissue.
B-scan diagnosis was first developed by Baum and Greenwood (24, 25, 26) in 1958. They made numerous original observations on B-scan evaluation of the eye
and orbit. Their work, featured on the cover of the Journal of the Acoustical Society of America (27) of a B-scan of the eye, was the adumbration of developments
to come. Baum's efforts were devoted primarily toward the development of equipment with increased accuracy and better resolution (Figure 3.1).
Purnell and Sokollu (28) used a similar prototype B-scan (developed by General Precision Instruments) and made many seminal observations that had a major
influence on B-scan diagnosis (Figure 3.2). His laboratory described orbital B-scan evaluation and provided the first systematic classification of orbital disease with
B-scan ultrasonography (29). He and others were the first to use the magnetic properties of a foreign body in ultrasonic diagnosis (28). Purnell and Sokollu (30,31)
also described special techniques for using continuous wave ultrasound (generally related to therapeutic applications) in early experiments for the treatment of
retinal detachment with ultrasound. They developed the first handheld contact B-scanner for use in ophthalmic ultrasound diagnosis (which preceded the Bronson
contact
P.48
ultrasound instrument), but it was never marketed (Figure 3.3) (32).

Figure 3.1. The immersion B-scan tank used by Baum and Greenwood provided sector scans of both eyes.
About 5 gallons of water and a special fitted face mask were required. These B-scans began the evolution of
B-scan ultrasonography of the eye.
Figure 3.2. Purnell used the same General Precision scanner as Baum but used a simple goggle and latex
cover to provide the immersion standoff necessary to conduct the ultrasound wave. Much of the early clinical
data provided by B-scan was achieved by Purnell et al.
Coleman et al. (33), at the Riverside Research Institute, developed the first commercially available B-scan with simultaneous A-scan and a simplified
hand-operated linear B-scan, using an immersion bath around the eye, created using a plastic surgical drape (Figures 3.4 and 3.5). Bronson and Turner (34)
developed a handheld B-scanner, which was the first of many easily used contact B-scanners commercially available (Figure 3.6). These instruments led to the
widespread use of ocular ultrasound. Fisher et al. (35) has made many clinically significant observations with this early instrument. Figure 3.7 shows the Sonomed
instrument, which was the first contact B-scan using an oscilloscope for more accurate morphic outlining. Coleman (36,37) presented an evaluation of the reliability
of ocular and orbital diagnosis with A-, B-, and M-scan ultrasound and a systematic description of ocular and orbital diagnosis. Coleman, Lizzi, and Jack (38)
published the first book on ultrasonic diagnosis of the eye and orbit. Coleman and Lizzi (39,40) and the Riverside Research Institute made many innovations,
including the use of color monitoring and encoding, and isometric viewing. Also, Coleman, Silverman, and Rondeau worked with power spectrum analysis for tissue
characterization, three-dimensional (3-D) ultrasound, and digital signal processing (I-scan) principles (56).
M-scan diagnosis, first described by Coleman and Weininger (41, 42, 43), has been used to study physiologic changes during accommodation and the magnetic
properties of foreign bodies. It has also been used for examining the vascular and respiratory pulsations in ocular and orbital tumors. Silverman, Kruse, and
Coleman (44) pioneered the use of swept-scan analysis for use in evaluating vascular flow in various ocular conditions. Color-flow Doppler (CFD) imaging of the
orbital vessels was first described by Erickson et al. (45) in 1989.
Figure 3.3. The first contact B-scan transducer system as devised by Purnell, Sokollu, and Holasek. The
instrument was never commercialized.
P.49
Figure 3.4. The equipment console used by Coleman for clinical ultrasonic evaluation. Two separate
interchangeable A-and B-scans were used with different frequencies and simultaneous oscilloscope displays.
An electronic interval counter and color and isometric displays were used. The laboratory instrument developed
at the Harkness Eye Institute by Coleman and Lizzi provided variable frequency examinations up to 35 MHz
along with precise electronic interval measurements of axial optical dimensions. This equipment was more
complex than required for routine clinical use. (see color image)
Figure 3.5. The first commercially available A- and B-scanner, developed by Coleman and Katz and marketed
by Sonometrics Systems, Inc. Both A- and B-scan modes were observed simultaneously by the examiner, with
a separate oscilloscope display available for photography of implementation of the M-mode. Bottom: Scanning
in “immersion,” performed while the patient lies supine on an examination table. This reduces the patient's head
movement and permits the examiner to observe the relationship of the transducer to the eye while also
observing the scan display on the oscilloscope.
Figure 3.6. The Bronson Turner contact B-scan provided an inexpensive B-scan sector scanner that led to
widespread use of the contact B-scan technique. The television raster lines altered the shape of the scans but
provided a very inexpensive way to add gray scale.
P.50
Figure 3.7. The contact B-scan unit with A-scan using an oscilloscope, developed by Katz and Coleman. The
oscilloscope provided more accurate morphology and amplitude quantification.
The use of Doppler ultrasound in the eye and orbit had its start as a method to evaluate the hemodynamics of patients with cerebrovascular disease and its
ophthalmic sequelae. The ophthalmic community was slow in adopting Doppler ultrasound as a diagnostic tool before the clinical availability of color Doppler
imaging (CDI) in the late 1980s. Pioneers who used continuous wave Doppler evaluation of the orbital and ocular vasculature, such as Yamamoto and Ardouin (46,
47, 48, 49) in the 1970s, used stand-alone Doppler equipment (without B-scan control) that required extraordinarily careful interrogation of the eye and orbit. With
the advent of duplex scanners, where a vector and Doppler gate could be positioned on a B-scan for localization, Doppler ultrasound became an important
diagnostic tool in cardiovascular, peripheral vascular, and obstetrical ultrasound. The cost, availability, limited frequency range, and, possibly, higher power levels
of such equipment ultimately limited Doppler use in the routine ophthalmic exam. The introduction of CDI changed this. Even with duplex Doppler, the identification
of the relatively small and tortuous orbital and ocular vasculature had been problematic. With the superimposition of color flow information, rapid and correct
identification of vessels became possible, and the diagnostic advantage of the technique began to outweigh the cost and access factors. Lieb et al. (50,51) were
among the first to popularize the use of CDI studies for a range of ocular and orbital conditions.
Ophthalmic ultrasound can be divided into two phases during its half-century history. Initially, techniques and applications were described and perfected. In the
second phase, advances in instrumentation and computer technology led to improved resolution and image quality as well as diagnostic and measurement
accuracy.
Recent years have seen many improvements in the quality of images, as a result of higher frequencies of examinations, improved electronics and transducers,
and, most important, computer power and software to allow 3-D scans (52), tissue characterization (53, 54, 55, 56, 57, 58), and other improved imaging techniques
(59, 60, 61, 62). Measurement accuracy has significantly improved owing to higher frequencies and computer enhancement techniques, such as digital signal
processing techniques, including deconvolution and analytic signal magnitude rectification of radiofrequency (RF) signals (63,64).
The most notable improvements in current ophthalmic ultrasound diagnosis have been with the high frequency B-scan. The ultrasound biomicroscope (UBM),
introduced by Pavlin and Foster (59,60,65,66), was the first commercial instrument to take advantage of polyvinylidene fluoride (PVDF) film technology for high
frequency scans. This instrument permitted numerous advances in anterior segment diagnosis, particularly in diagnosis of glaucoma, tumors, and trauma of the
anterior segment. A different high frequency scanner (Figure 3.8), developed in our laboratory at Weill Medical College of Cornell University by Coleman et al.
(64,67), uses an arc scan to display the entire anterior segment and to align the transducer orthogonally with the anterior segment, to maximize accuracy of
measurement for various applications,
P.51
such as corneal mapping, pre- and post-LASIK surgery, physiologic studies of lens changes in accommodation, ciliary body and lens movement in presbyopia, and
angle-to-angle and sulcus-to-sulcus measurements for intraocular lens powering and surgery.
Figure 3.8. An Artemis II high frequency scanner, developed by Ultralink LLC from technology devised at
Cornell, for 50-MHz anterior segment imaging. This scanner provides orthogonal transducer alignment for
viewing of the entire anterior segment by means of an arc scan. The definition is the current state-of-the-art for
measurement of corneal thickness and anterior chamber dimension for both LASIK and intraocular lens surgery.
In addition, it provides superb definition of intraocular pathology, such as intraocular tumors and ciliary body
cysts.
Three-dimensional imaging of the eye was first described by Coleman et al. in 1987 but was not widely available until less expensive computer power and software
became available. Three-dimensional imaging, as will be noted later, offers a significant advantage in measuring tumor volume for growth or regression
posttreatment. It also provides a perspective for scanning that can aid interpretation of 2-D images, as well as interactive analysis of ultrasound images. Fisher et
al. (61), working with Ophthalmic Technologies, Inc. (OTI), have developed a commercially available 3-D ultrasound system (Figure 3.9). Finger et al. (52) have
also used this OTI scanner to examine intraocular tumors and to demonstrate the advantages of 3-D perspectives.
Figure 3.9. The OTI is a sector B- and A-scan system operating at 12 MHz or higher frequencies that provides
integrated 3-D scans with a contact system.
There are many new instruments available for not only 10-MHz scanning, but also 20- to 30-MHz B-scanning. Many of our figures were taken with a Sonovision
scanner (Figure 3.10), which is no longer commercially available,
P.52
but provides the radiofrequency data necessary for our tissue characterization and parameter image analyses. We also use the Quantel Cinescan 20-MHz scanner
(Figure 3.11) for evaluation of the posterior pole of the eye, as well as the vitreous. Figures throughout this chapter will have been produced using one of these four
systems. A-scans, when shown, are always quantitative derivations from the radiofrequency, except for A-scans used in axial biometry.
Figure 3.10. The Sonovision ultrasound scanner, which was the first commercial B-scanner able to provide
digital capture of the radiofrequency data. This permitted mathematical analyses of the data and the
development of power spectrum analysis.
OCULAR DIAGNOSTIC TECHNIQUES

Synopsis
TWO DIAGNOSTIC APPROACHES
1. B-scan with A-scan ultrasonography. A-scan generated along vector of B-scan with same focused transducer.
Class A: A-scan is calibrated and has radiofrequency display for “quantitative” digital display.
Class B: A-scan vector uses same amplifier that generated the B-scan. It is a simple rectified trace and has not been calibrated for amplitude variation.
2. Standardized echography, using a separate B-scan and A-scan with a nonfocused transducer. The A-scan is calibrated against a tissue standard.
As has been discussed in Chapter 2, there are two primary techniques for ocular ultrasonic examination: A-scan and B-scan, with supplementary display
techniques of M-scan, I-scan, 3-D scans, and kinetic scanning. With these techniques, different systems for ultrasonically evaluating patients have evolved. A-scan
and B-scan are not mutually exclusive methods of diagnosis. A thorough knowledge of A-scan or B-scan ocular ultrasonography can provide reliable diagnostic
information, but a combination of both A- and B-scan diagnostic methods is optimal and almost universally used. In our laboratory, for diagnostic purposes we rely
primarily on B-scan, and we use a constant A-scan monitor to obtain maximum quantitative echo amplitude information from the calibrated RF digitized data.
B-scan provides the two-dimensional display that provides the cross-section basis for comparison of characteristic echo amplitude variation (or third dimension)
provided by the A-scan. The “third dimension” of the levels is displayed on the B-scan as gray scale but varies with amplifier characteristics and gain settings.
Amplitude comparisons can be very useful on B-scan, but amplitude character of a tissue is more accurately obtained by observing the A-scan monitor
simultaneously with the B-scan to identify the orientation of the A-scan vector.
Figure 3.11. Quantel Cinescan 20-MHz sector scanner providing excellent resolution of the posterior pole on
B-scan.
Ideally, B-scan ultrasound with good dynamic range, or gray scale, can present a tomogram, or a thin cross section of the eye, with highly accurate resolution of
tissue surfaces, such as the cornea, the anterior chamber, or tumor characteristics. It also displays reflectivity patterns as a cross section within the tissue being
observed. Amplification of echoes or dynamic range is best shown using logarithmic amplification or the “s” shaped amplification of Ossoinig (12). M-scan is a
technique that has been beneficial in demonstrating consistent or reproducible pulsations, such as the respiratory or vascular pulsations of certain tissues, or the
magnetic properties of foreign bodies, but is used only occasionally in ocular diagnosis.
Ossoinig (11) has emphasized the value of A-scan ultrasonography in providing quantitative echo information and developed an “s” shaped amplifier to compress
the dynamic range of echoes and emphasize echo amplitude variations with a separate and independent A-scan evaluation, using a nonfocused A-scan
transducer. However, most ophthalmologists can more readily interpret two-dimensional B-scan patterns and rely on the A-scan from the same instrument and
transducer to interpret the pathology. This has given rise to two separate “schools” of ophthalmic ultrasonography. Most practitioners use the B-scan and obtain a
vector A-scan, using the same transducer that produces the B-scan.
Our approach has been to use an amplifier that generates an RF signal. This signal can be rectified to produce a characteristic A-scan along a vector on the
B-scan. The amplitude is calibrated for the transducer (the same focused transducer that produces the B-scan when the trace
P.53
is modulated). With this technique of calibration from the RF trace, which is termed quantified A-scan, a very accurate amplitude is determined. We find this method
preferable because it provides the most accurate amplitude quantification while allowing it to be compared to the precise tissue area as localized on the B-scan.
However, in most commercial systems, the A-scan is generated along the vector, from the displayed B-scan pixel intensities. These instruments are less expensive
and may be adequate for routine examination, but the A-scan amplitude comparisons are obviously less desirable. The “standardized echography” school uses a
separate A-scan with nonfocused transducer standardized against a tissue standard using the “s” shaped amplifier developed by Ossoinig (70, 71, 72). This
method is more time-consuming because the examiner may switch from B- to A-scan and cannot identify the tissue of reference as easily as when it is selected
from the simultaneous B-scan. In addition, angle of incidence of the beam and the tissue that affects amplitude are less confidently identified.
Excellent courses and books on standardized echography are available that detail this technique, notably Frazier-Byrne and Green (73) and DiBernardo and
Schachat (74). Although ideological differences thus exist, both methods essentially use the B-scan for orientation and rely on an A-scan for tissue quantification
and identification.
The combined A-scan and B-scan technique, because of its clinically demonstrated value and availability, has come to play a critical diagnostic role in
ophthalmology, particularly when the media is opaque or a lesion is occult.
This book will outline the method of diagnosis, which uses the B-scan to provide the broad, topographic information about tissue geometry and morphology and the
A-scan and other digital analytic techniques to provide specific comparative information regarding the reflectivity and backscatter from tissue structures, as well as
accurate measurement of their dimensions.
Figure 3.12. A typical 10-MHz B-scan with a vector selected A-scan. The power and other settings for
modification of the scan are shown, as is the time variable gain graphic on the lower right.
DIAGNOSTIC PARAMETERS
The typical ophthalmic examination proceeds in two stages to identify the anatomic ocular features listed in Table 3.1. First, general tissue features, such as size
and position are established. Second, specific features, particularly anomalous structures, are identified by examining echo characteristics that are indicative of
finer, more discriminative morphologic features.
Although the general architecture of the eye and orbit is readily discerned in B-scans (Figure 3.12), finer features can be interpreted only with an understanding of
how
P.54
tissue structure influences ultrasonic reflectivity. As noted in Chapters 1 and 2, different tissues transmit, absorb, and reflect in various manners, depending upon
factors such as density, elasticity, and internal structural features. At boundaries between tissues, ultrasound is reflected to a degree determined by the acoustic
impedance mismatch between the tissues and by the size, orientation, and roughness of the boundary. To use an analogy, a mirror that is very smooth and that
lies perpendicular to a flashlight beam will reflect most of the energy back toward the flashlight (a specular reflector) but very little reflectance back to the flashlight,
if the light is off-axis. If the mirror is roughened (a diffuse reflector) or smaller than the total beam or angled away from the beam, it will reflect proportionately less
light back to the flashlight. Conversely, a diffuse reflector will reflect some energy even off-axis, allowing curved surfaces, such as the lens, to be better outlined if
blood or fibrin converts the smooth to a diffuse reflective surface (Figure 3.13).
TABLE 3.1 Diagnostic Parameters
Gross Morphologic Features
Location
Size
Outline/Contour/Shape
Associated Ocular Changes
Changes with Time
Fine Morphologic Features
Boundary Layer Properties
Acoustic Impedance
Roughness of Surface
Internal Tissue Properties
Internal Texture (Homogeneous or Heterogeneous)
Type of Internal Structural Elements
Spatial Distribution of Internal Structural Elements
Acoustic Absorption
Internal tissue characteristics also influence ultrasonic transmission and reflection. If a tissue has a homogeneous structure (e.g., lens, optic nerve, or a solid tumor,
such as a malignant melanoma), there are few internal reflective surfaces, giving a “cystic” or sonolucent hollow or hypoechoic appearance on B-scans. This
appearance contrasts sharply with the dense, speckled hyperechoic appearance generated by reflections from internal features of heterogeneous structures, such
as in hemangioma, angioma, or vitreous hemorrhage. In these heterogeneous structures, echo amplitude and spatial distributions depend on the type and
distribution of the internal structural elements (e.g., blood vessels, calcific deposits, or necrotic regions). In addition, the falloff of echo amplitude with increasing
depth is indicative of attenuation of the ultrasound beam through absorption and scattering. (In homogeneous structures, attenuation can be manifested by a
“shadowing,” or blocking of detail, in more posterior tissues.) The attenuation of ultrasound frequency-related scattering in tissue can be used to identify specific
ocular tissues and pathologic conditions. New developments in instrumentation promise to provide even more gray scale and computer-enhanced information than
is presently available for ultrasound-based tissue diagnosis.
Figure 3.13. Immersion ultrasound examinations of two traumatized eyes, one with an older 10-MHz instrument
(left) showing how blood can outline the entire lens. On the right, a newer contact B-scan used in immersion
demonstrates how blood can help outline the lens by converting it to a diffuse reflector.
TYPES OF DIAGNOSTIC INFORMATION

Synopsis
INDICATIONS FOR DIAGNOSTIC ULTRASOUND
Measurement of distances or volumes (corneal thickness, ocular biometry)
Opaque media—no view (cataract, blood, etc.)
Occult areas of globe (retroiridal)
Trauma—foreign bodies
Retinal/choroidal detachment
Orbit and optic nerve
We have found it convenient to consider ultrasonic diagnosis in terms of the (a) unique, (b) supplemental, and (c) documentary information available from
ultrasound. Unique information indicates that which is obtained with opaque ocular media or, for example, in occult ciliary body tumors. Supplemental information is
exemplified by tumor diagnosis. Although a mass may be visualized ophthalmoscopically, differentiation by means of clinical appearance is often inaccurate or
misleading. The ultrasonic characteristics of a mass (e.g., shape, height, and acoustic transmission properties) may be added to information that is obtained
visually. Documentary information refers to the ability of ultrasound
P.55
to measure accurately the thickness of the cornea or lens, the length of the globe, the dimensions of a tumor, the motion of the lens during accommodation, or
ocular properties that vary under the influence of drugs. All of these are measurements that are not easily, or not at all, obtainable by other means. Although the
unique information is the most dramatic and most often described utilization of ultrasound, (e.g., detection of tumors in opaque eyes), the supplemental and
documentary uses, such as in characterizing tumors or measuring eye length for an intraocular lens, are of equal or, in many instances, greater importance.
INDICATIONS FOR OCULAR ULTRASOUND

A summary of the indications for the ophthalmic use of ultrasound is presented in Table 3.2. In addition, there are specific uses of ultrasound, such as detecting loci
of choroidal effusion in patients with flat chamber, preoperatively studying the character and form of vitreous hemorrhages prior to vitrectomy, measuring the size
and volume of tumors prior to radiation, plaque, proton beam or other treatment, and measuring axial dimensions to determine keratoprosthetic or intraocular lens
dioptric powers.
In general, where visual techniques fail to provide sufficient information as to the structural configuration of the eye, ultrasonic imaging is indicated. It is safe,
economical, and rapid. Even where it fails to provide optimal imaging, as, for example, when MR or CT is superior for orbital evaluation, or CT is superior for
identifying foreign bodies, follow-up evaluation may be useful and most economically provided by ultrasound.
THE NORMAL EYE
A-SCAN ULTRASONOGRAPHY
Synopsis
A-SCAN MEASUREMENTS FOR INTRAOCULAR LENS
Known velocities for each tissue traversed by the round trip “time of flight” echo-detected, convert the time measurement to distance (mm) using the formula:
Distance = Velocity × (time/2)
Maximal amplitude of lens and posterior pole surfaces indicate alignment with the optical axis.
TABLE 3.2 Indications for Ocular Ultrasonography
Opaque Media
(Corneal Leukoma, Hyphema, Hypopyon, Cataract, Vitreous Hemorrhage)
Occluded or Markedly Miotic Pupil
Ophthalmoscopically Visible Mass Lesion
Suspicion of Tumor Underlying Retinal Detachment
Ocular Trauma
Ocular Foreign Body
Axial A-scan Ultrasonography

The axial ultrasonogram is obtained by using the visual or optical axis as the path of the examining ultrasound beam, so that echoes are obtained from structures
along the path of the central cornea and on posterior through the lens to the retina. Echoes arise from the ocular tissue interfaces that produce acoustic impedance
mismatches. These echoes are displayed as vertical deflections on a display device. In the optical axial echogram of the normal eye (Figure 3.14), high amplitude
echoes are produced by the corneal surfaces, by the lens surfaces, and by the vitreoretinal interface. The vitreoretinal interface echo is followed by a complex of
echoes representing retina, choroid, sclera, and retrobulbar fat. The echoes in the retrobulbar fat diminish gradually to baseline as the sound is absorbed. Certain
parts of the eye are normally acoustically homogeneous at typical ophthalmic ultrasound frequencies. These include the cornea, anterior chamber, the lens, the
vitreous, and, to some extent, the optic nerve. These areas appear as baseline (zero echo or anechoic) segments between echo groups from their surfaces.
Optical axis measurements are obtained when the echo amplitude is maximized as a result of the orthogonal or perpendicular relationship of the transducer beam
and the tissue, that is, cornea, lens surfaces, and retina. This feature is used to insure alignment of the optic axis when taking axial measurements for computation
of lens power (Figure 3.15).
The visual axis may be more important in some situations but can be obtained only by having the patient visually align his or her eye with a target or light in the
center of the transducer beam. Because this is subjective and required only in special situations, there are only a few transducers so specially designed, and
optical axes are the norm for A-scan measurement.
The velocity of sound constants for all of the anatomic structures in the ultrasound beam path, that is, cornea, anterior chamber, lens, and vitreous, are required to
convert the round trip “time of flight” measurements to distance in millimeters (time/2 × velocity = distance) (Table 3.3). Several formulas based on optical models
of the eye, regression, or some combination of the two, can then be used to indicate proper lens power for intraocular lens (IOL) implantation.
P.56
Figure 3.14. An A-scan along the optical axis at 10 MHz showing the amplitude variation of the optical surfaces.
Note that the posterior lens surface is of lower amplitude than the anterior surface because of lens absorption
and the concave surface reducing the area incident to the beam. This feature is useful in maximizing true optical
axis measurements because on-axis measurement is assured when the posterior lens echo is maximal.
B-SCAN ULTRASONOGRAPHY
Synopsis
B-scan alignment can be “on axis” (through the lens) or diascleral. The lens and lids attenuate sound, more so with higher frequencies.
High frequency has better resolution but less tissue depth penetration or sensitivity, and lower frequencies give less resolution but deeper tissue sensitivity.
Artifacts are produced by anatomic and electronic causes and must be accounted for in examinations.
B-mode ultrasonographic systems have been described in Chapter 2. B-scan ultrasonography provides a two-dimensional “acoustic section” of the globe along any
desired scan plane. As in A-scan ultrasonography, the appearance of the normal eye varies according to the scan plane selected.
B-scans are anamorphic displays. Depth or distance is related to sound transmission, whereas cross section or lateral position is related only to the orientation of
the transducer.
Axial B-scan Ultrasonography (10-MHz Sector Scan)

A typical 10-MHz B-scan ultrasonogram along an axial scan plane (Figure 3.16) shows both the anterior and posterior surfaces of the cornea, separated by a
sonolucent interval representing the corneal stroma. With sector scanners the cornea will have a “reverse curve” because of the sector movement and the width of
the transducer beam in the near field. The anterior chamber appears as a uniformly, acoustically clear (hypoechoic) area. The anterior surface of the iris is usually
demonstrable. The echoes from the posterior iris surface usually merge with those from the anterior lens surface. However, with a dilated pupil, the anterior lens
curvature is more prominently seen. The interior of the normal lens also appears as an acoustically homogeneous (hypoechoic) space. The posterior curvature of
the lens is usually well demonstrated, at least centrally with a sector scanner, but the equator is not seen because of its oblique orientation to the beam. The
vitreous compartment normally appears as an anechoic or sonolucent cavity with no internal sound reflections. The vitreoretinal interface forms a smooth, concave
curvature. Echoes from the retina merge with echoes from the choroid and the sclera, and in the normal eye these contiguous echoes cannot be well separated at
normal examining frequencies of 10 and 15 MHz. These boundaries between the retina, choroid, and sclera can be better identified using digital signal
P.57
processing techniques. The scleral fat boundary (Tenon's capsule) is, however, well seen acoustically. In B-scan ultrasonography, the area between the ora serrata
and the equator of the globe is poorly demonstrated with axial orientation of the transducer, because, as with the equator of the lens, these areas are more parallel
to the sound beam. Thus, for a complete ultrasonic examination, multiple scans with the scanner placed meridionally at each clock hour and using diascleral scans
that bypass the lens is necessary. In an axial B-scan, the retrobulbar fat forms a W-shaped pattern, with a black notch formed by the relatively homogeneous optic
nerve. The orbital fat appears as a highly reflective mass with extraocular muscle bellies forming the outline for the fat. The normal orbital B-scan appearance is
discussed in Chapter 5 on orbital diagnosis.
Figure 3.15. Left: An A-scan demonstrating maximized posterior lens echoes along the optical axis. It is shown
as a vector on the B-scan to help demonstrate the positioning of the ultrasound beam. The anterior lens echo is
clipped or saturated, explaining the difference in appearance to Figure 3.14.
TABLE 3.3 Reported Mean Velocities of Ultrasound in Ocular Tissues
Tissue (Accepted Velocity) Velocity (M/Sec) Temperature (°C) Frequency (MHz) Investigator
Cornea 1,639 M/sec 1,632 22 4 Chivers
1,550 22 4 Oksala
1,553 22 10 Thijssen
1,572 20 20 De Korte
1,575 37 60 Ye
Sclera 1,744 22 4 Chivers
1,630 22 4 Oksala
1,597 20 20 De Korte
1,622 37 60 Ye
Vitreous 1,532 M/sec 1,508 22 4 Chivers
1,495 22 4 Oksala
1,532 37 4 Jansson
1,514 20 20 De Korte
Lens 1,641 M/sec 1,548 22 4 Chivers
1,650 22 4 Oksala
1,641 37 4 Jansson
1,629 37 15 Coleman
1,590 20 20 De Korte
Silicone Oil
1,000 CS 972.0 37 7.5 Silverman
5,000 CS 978.5 37 7.5 Silverman
Figure 3.16. A typical immersion B-scan at 10 MHz demonstrating the excellent imaging of the vitreous and
retina, as well as moderate imaging of the lens. The cornea has only a small area visualized because of the
mismatch of sector scanning and the corneal curvature. The echoes in the vitreous are a “multiple” of the
anterior segment. The reverse arc of the cornea is explained in Figure 3.19.
Diascleral (Off-axis) A-scan Ultrasonography

To obtain the diascleral ultrasonogram, the examining ultrasound beam must pass peripheral to the cornea and the lens to avoid the absorption of the ultrasound
beam by the lens. The transducer is placed on or anterior to the sclera, and aligned toward the posterior pole. The normal diascleral ultrasonogram (Figure 3.17)
consists of a high-amplitude echo complex, representing sclera, followed by a long acoustically empty (anechoic) interval, representing the normal vitreous cavity.
The final echo complex produced by retina, choroid, sclera, and retrobulbar fat is similar to that seen in the axial ultrasonogram. Oksala (75) first pointed out that
the echoes from the posterior ocular wall in the diascleral ultrasonogram are
P.58
higher and broader than those obtained in the axial ultrasonogram, because there is no sound absorption from the lens. Tumor measurement and such amplitude
features that aid diagnosis of retinal detachment often benefit from this variation of beam orientation.
Figure 3.17. A diascleral B-scan demonstrating improved sensitivity of the posterior pole when the ultrasound
beam is not partially absorbed by the lens.
Transducer Frequency Variation

As we have discussed in Chapters 1 and 2, there is a balance, or “trade-off,” between resolution and penetration. Figure 3.18 shows B-scans of the same normal
eye, taken at 10 and 20 MHz, to illustrate the higher resolution obtainable with the 20-MHz transducer. The penetration of the ultrasound beam from a 20-MHz
transducer, however, is much less than that obtained with a 10-MHz transducer, which depicts more of the orbital fat and optic nerve. In general, the 10-MHz
examining frequency is the best compromise for initial examination, with higher (or lower) frequency transducers then substituted for the study of specific tissues or
areas.
Figure 3.18. B-scans of a normal eye at 10 MHz (left) and 20 MHz (right), which demonstrate improved
resolution at higher frequency with concomitant reduced sensitivity.
Effect of Scanning Mode

Figure 3.19 shows the difference between a sector scan and an arc scan of the eye, using a single transducer (see also DVD). An arc scan is better at outlining the
contours of the anterior segment and the equator of the globe than is the sector scan. The arc scanner and sector scanner are equivalent at the posterior pole,
because at the lower 10- to 20-MHz frequencies needed for the posterior pole, the sector and reverse arc are the same. (The pivot point of the arc scanner is
placed in midvitreous so that the reverse arc follows the contour of the posterior pole.)
Effect of Lid Attenuation

We prefer to use a water bath standoff with a lid speculum in most diagnostic work; however, both immersion and contact B-scan ultrasonography can be
performed through the closed eyelid. Figure 3.20 demonstrates the marked attenuation of ultrasound energy caused by passage through the lids. In addition to the
marked absorption of the sound beam, ocular structures immediately posterior to the lid are obscured. For optimum B-scan ultrasonography, a water bath standoff
of some type with lids open, usually with a speculum, is recommended. The contact B-scan is easier to use and for routine evaluation, such as evaluating possible
retinal detachment or choroidal elevations, may be the preferable technique because of ease of use.
P.59
Figure 3.19. Schematic demonstration of the difference between a sector and an arc scan for imaging a
convex surface. The sector scanner will give a small reverse curve corneal echo, caused by the edge of
the transducer beam striking the reverse slope of the cornea but imaged as if it were the center of the
beam. This oblique angle also reduces the area of the cornea that can be displayed with a sector scan.
The arc scan remains, generally, perpendicular over the area of travel of the transducer, thus producing
a more accurate image. (See also DVD.)
ARTIFACTS ENCOUNTERED IN OCULAR ULTRASONOGRAPHY

Occasionally, artifacts arise in the course of ultrasonic evaluation of the eye, and familiarity with their appearance will avoid erroneous interpretation. Artifacts may
be classified into four groups: (a) electronic artifacts, (b) reduplication echoes, (c) refraction artifacts, and (d) absorption effects. The sources of these artifacts are
treated in Chapter 2. Examples of the types of clinically relevant artifacts are presented in the next sections.
Figure 3.20. 10-MHz B-scan with A-scan demonstrating reduced attenuation by bypassing the lid. C, cornea;
AL, anterior lens; PL, posterior lens; R, vitreoretinal interface.
ELECTRONIC ARTIFACTS
In certain scan situations, artifacts may arise from unsatisfactory electronic processing of the ultrasonic echoes. A typical artifact is referred to as snow, which is
produced by background noise (“grass” on the A-scan trace) and resembles interference on a television screen. Background noise can usually be eliminated
electronically by requiring incoming echo-generated energy to exceed a certain threshold level before triggering the B-scan presentation, thus rejecting
low-amplitude background noise. This problem is rare with modern B-scan ultrasound systems.
REDUPLICATION ECHOES
These echoes (also known as multiple echoes) occur commonly and have been extensively analyzed by Kossoff (76). They usually appear along the axis of the
cornea and lens. They occur when the transducer is aligned perpendicular to a tissue surface and high-amplitude echoes are reflected back to it. These echoes
can then be reflected from the transducer back to the tissue and then rereflected, producing what is called a reduplication echo at a multiple of the distance
between the transducer and the reflecting surface. An echo of this type is often seen in midvitreous when, using a water bath standoff, the transducer is positioned
a short distance (e.g., 1 cm) from the eye. The artifact would then appear in midvitreous, that is, 1 cm posterior to the cornea, although this artifactual echo can be
displaced farther back, even into the orbit fat by positioning the transducer farther away from the eye. Echoes bouncing back and forth between the transducer and
the cornea may mimic abnormal tissue or foreign bodies. These echoes may be distinguished from real echoes by moving the transducer either toward or away
from the eye. This causes a displacement of the reduplication echo relative to tissue, allowing it to be identified (Figure 3.21; see also DVD).
REFRACTION ARTIFACTS
Other artifacts relating to the transducer position are produced by refraction of ultrasound within ocular tissues. On B-scan, the relatively high lenticular propagation
velocity can produce apparent abnormalities of the posterior pole that resemble tumor formations or thickening of the choroid (Figure 3.22). Purnell (29) has
referred to these refraction abnormalities of the posterior pole as “Baum's bumps,” because they were originally described by Baum (77). In general, if a mass is
seen at the posterior
P.60
pole, scans should be made through different planes of preceding tissue to ascertain that the abnormality is not a reduplication echo or caused by refraction
through the lens. Scans of the posterior pole should be made through the limbus or a more peripheral position, if practical, so that only normal sclera and vitreous
precede the area of interest.
Figure 3.21. The arrows show reduplication artifacts from the anterior segment as a result of high gain used in
the electronic display. C, cornea; IP, iris plane.
ABSORPTION EFFECTS (SHADOWING)

Absorption of sound energy by anteriorly located structures may cause abnormal ocular ultrasonic patterns. The absence or attenuation of echoes in the posterior
segment, giving an appearance of a defect in the ocular wall, is a typical example. Commonly encountered causes of such defects are dense cataract (Figure 3.23)
and organized hemorrhage or a foreign body, for example, metal or silicone oil.
Figure 3.22. Baum's bumps on a contact B-scan. (See also DVD.)

Figure 3.23. Artifacts posterior to a calcific cataract are seen, producing distortion and hypoechoic areas
caused by deflection and absorption of the ultrasound beam.
Whenever an abnormal ocular B-scan pattern is encountered, the previously mentioned artifacts should be considered and ruled out. Recognition of these artifacts
is aided by (a) careful monitoring of the A-scan, which permits recognition of many electronic artifacts; (b) repositioning of the transducer, if a reduplication echo or
shadowing is suspected; and (c) analysis of any ocular abnormality that may cause absorption defects in the acoustic pattern and, perhaps, most important, a good
history.
ABNORMALITIES OF OCULAR SIZE AND SHAPE

B-scan ultrasonography graphically portrays anomalies of ocular size and contours. B-scan ultrasonography, with its capability for two-dimensional, cross-sectional
display, allows this information to be derived from a single scan plane. Figure 3.24 is a B-scan ultrasonogram of a patient with a posterior staphyloma as a result of
high myopia. This aberration from the normal posterior contour of the globe appears acoustically as a concave dip in the globe wall. Figure 3.25 is the
ultrasonogram of a patient with a coloboma of the choroid, in addition to a posterior staphyloma. The coloboma gives a pronounced aneurysm-like defect in the
ocular wall, with a sharply defined rim.
B-scan portrayal of an enlarged globe and A-scan axial measurement documenting increased axial length allow differentiation of pseudoproptosis from true
proptosis. This feature is discussed further in Chapter 5.
P.61
Figure 3.24. B-scan demonstrating the posterior outpouching of the globe in the macular area, as seen in a
typical staphyloma.
ANAMORPHIC DISPLAY
The two-dimensional B-scan does not use the same scale for depth and cross section, which is essential to remember. Depth or range is dependent on sound
transmission, whereas the lateral or cross section is dependent entirely on transducer sweep (see Chapter 1) and the electronic tracking of the sweep (Figure
3.26).
VERY HIGH FREQUENCY ULTRASOUND AND ULTRASOUND BIOMICROSCOPY

The very high frequency ultrasound, first introduced by Pavlin and Foster (59) as ultrasound biomicroscopy, or UBM, optimizes anterior segment imaging.
Frequencies of 50 MHz and higher provide superb imaging of the cornea and anterior segment. Resolution of 30 microns or less can be achieved, and
reproducibility with I-scan (digital signal processing) can approach 5 microns for the cornea thickness. Chapter 4 describes more fully corneal and anterior segment
measurement in relation to very high frequency ultrasound and refractive surgery considerations.
Figure 3.25. B-scan demonstrating a coloboma at the posterior pole of an infant showing the relatively sharp
edges or clivus of the defect.
Figure 3.26. Schematic of the anamorphism of ultrasonic imaging produced by different scales for the axes
depth and cross section. Depth relates to the speed of sound and density of tissue, whereas cross section
relates solely to transducer displacement. This feature must be recognized and accounted for in any
calculations that are taken off-axis to the transducer beam.
ANTERIOR SEGMENT ABNORMALITIES

Synopsis
ANTERIOR SEGMENT
Very high frequency ultrasound (VHFU), that is, 50 MHz and greater, is the preferable way to evaluate the cornea, iris, and ciliary body. Lower frequencies are
required to outline the lens. Unless blood or fibrin converts the surfaces to a diffuse reflector, the entire outline is not seen.
VHF ultrasound, or UBM, provides excellent definition of the cornea and anterior segment, including iris and ciliary body tumor detection.
CORNEA
Size and Shape Abnormalities

As discussed earlier, anomalies of corneal size, such as megalocornea or microcornea, may be demonstrated with
P.62
B-scan ultrasonography. Corneal curvatures can be directly measured from the B-scan display, and abnormalities of corneal shape (such as keratoconus) are
demonstrable but generally require a high index of suspicion to merit the effort to map the posterior cornea and to measure corneal density.
Corneal Thickening
Thickening of the cornea can be determined by A-scan ultrasonic biometry (Chapter 4), and a form of A-scan is commonly used for “pachymetric” measurements.
Pachymetry is becoming increasingly important to the glaucoma specialist in estimation of the intraocular pressure. Immersion B-scan ultrasonographic techniques,
however, may also be used and can provide imaging and measurements over the entire structure, thus providing additional diagnostic information to the glaucoma
or cornea surgeon (Figure 3.27).
Keratoprosthesis
In addition to the examination of the globe prior to keratoplasty, ultrasound is valuable in the evaluation of eyes prior to prosthokeratoplasty. As previously
mentioned, clinical evaluation of eyes with opaque corneas is, at best, difficult, whereas ultrasound permits accurate determination of the status of the posterior
segment of the globe. The axial length of the eye can be obtained ultrasonically, allowing the placement of accurate dioptric correction in the optic cylinder of the
prosthesis. Although most eyes proposed for keratoprosthesis insertion are aphakic, in some cases a lens or lens remnant may be present. Ultrasonography
determines the presence or absence of a lens and prepares the surgeon for a lens extraction at the time of keratoprosthesis placement, if the eye is phakic. If a
cyclitic membrane is found ultrasonically prior to prosthokeratoplasty, it may be planned for surgical excision during prosthesis placement.
Figure 3.27. Top left: A 10-MHz immersion scan and Top right: a 50-MHz scan of the cornea, demonstrating
the resolution of this higher frequency in a patient with corneal opacification. The scarring was caused by
corneal abrasion. Lower left: A cross section of a thickened cornea showing the irregular outline of Descemet's.
Lower right: The relationship of the angle and the cornea in the same patient.
Visual evaluation of a globe is difficult because of the very limited field of view (2 disc diameters) through the keratoprosthesis. The two-dimensional acoustic
section of the globe provided by B-scan ultrasonography facilitates recognition of possible pathologic conditions.
Abnormal Eye with Keratoprosthesis

Coleman et al. (78) described results in 22 patients who were referred for ultrasonography because of unexplained visual loss after months or years with a
satisfactory result from keratoprosthesis insertion. Nineteen of these patients were found to have ultrasonically demonstrable posterior segment abnormalities
accounting for their visual loss. These abnormalities were classifiable into four groups: (a) cyclitic membrane, (b) choroidal detachment, (c) vitreous hemorrhage,
and (d) retinal detachment. In some of the eyes, two of these conditions coexisted. Surgery can be more cogently planned, using gas and vitreosurgical methods,
with foreknowledge of these anatomic conditions (79).
P.63
Figure 3.28. A 50-MHz arc scan of a normal anterior segment, demonstrating in one frame the cornea, iris,
ciliary body, and ciliary processes, as well as the anterior lens surface. This format allows measurement of all of
the anterior segment dimensions more accurately than possible with collages, such as with the UBM.
ANTERIOR SEGMENT ULTRASOUND
ANTERIOR CHAMBER DEPTH

Accurate measurement of anterior chamber depth is obtained by A-scan ultrasonic biometry or by high frequency B-scan (50 MHz or VHFU). The A-scan beam
should be positioned along the visual or optical axis to obtain a central representative and repeatable measurement. The optical axis is generally used because the
use of maximal anterior and posterior lens echoes is easiest to align. The visual axis requires a target alignment transducer system, as described in Chapter 2.
B-scan ultrasonography can provide accurate two-dimensional information, as well as anterior chamber depth, and can allow evaluation of the iris and sulcus plane
depths. A normal anterior chamber is shown in Figure 3.28, and Figure 3.29 is a B-scan ultrasonogram of a phakic patient with a flat anterior chamber. The lens is
displaced anteriorly, and the iris is seen to lie against the corneal surface. The echoes from the posterior cornea and anterior iris merge, and the interface between
these two structures can be outlined. Figure 3.30, conversely, demonstrates a deep anterior chamber in a phakic eye following trauma with a cyclitic membrane
producing hypotony. To provide the optimal accuracy in measuring chamber dimensions, careful three-dimensional alignment is required to avoid off-axis errors.
Figure 3.29. A 50-MHz arc scan of a phakic patient with a flat anterior chamber.
Figure 3.30. An abnormally deep anterior chamber in a patient following trauma, with a hypotony and
iridodialysis.
Hyphema
Hyphema (blood in the anterior chamber) appears as an echoic structure of variable echogenicity, depending on duration and clot lysis. Fresh hemorrhage will
generally have low echogenicity, increasing with organization into a clot. Blood will sometimes be traceable to the site of bleeding and will often tend to accumulate
in the angle inferiorly. Blood on the surface of the lens enhances imaging by producing a diffuse reflector, as seen in Figures 3.13 and 3.31.
Iris
Normal Iris
The normal iris is highly reflective and can be well imaged at 10 MHz, but it is far better outlined at 50 MHz. The melanin laden surface is reflective and is of
interest
P.64
largely for congenital anomalies, for physiologic studies, in trauma, and for evaluation of tumors. Iris conformation is significant in glaucoma as well, as in pupillary
block.
Figure 3.31. A patient with a normal anterior segment and the presence of a hyphema, which helps outline the
zonular attachment to the anterior capsule in this 50-MHz scan.
Figure 3.32. An iris melanoma at 50 MHz showing thickening of the iris but no evidence of extension into the
angle or ciliary body.
Tumors
Iris tumors can be detected with ultrasonography, even if they are less than one millimeter in thickness. Generally, these lesions are small and only a “solid” versus
“cystic” differentiation can be made. However, ultrasonic evaluation is valuable in determining the possible extension of the tumor into the ciliary region (Figure
3.32) (see Tumors in later section). Generally, very high frequency examinations at 50 MHz are the best way to satisfactorily visualize the iris.
Iris Cysts
Iris cysts are seen ultrasonically (VHFU) as rounded hypoechoic areas and may be differentiated from iris tumors, which appear solid or acoustically opaque
(isoechoic or hyperechoic with the iris). The B-scan ultrasonogram of a patient with a ciliary body cyst is shown in Figure 3.33. Often the origin of the cyst may be
undifferentiable as to iris or ciliary body in nature. However, it is important to distinguish between solid and cystic masses (see Ciliary Body Tumors in later section).
Figure 3.33. Ciliary body retroiridal cyst can be demonstrated in this occult area as clear, usually rounded,
single, or multiple cyst spaces. They are nearly always clear acoustically and may, at times, contact the lens
and conceivably cause cataract formation.
Figure 3.34. A patient with iris bombé, demonstrating adhesions of the iris sphincter to a cataractous lens.
Iris Bombé/Plateau Iris/Post-glaucoma Surgery

A B-scan ultrasonogram of a patient with iris bombé is shown in Figure 3.34. This iris is pushed forward in a convex fashion. Other iris changes, such as plateau
iris (Figure 3.35), have been described by Pavlin et al. (65)
P.65
and by Pavlin and Foster (80). These anatomic changes are best seen with 50-MHz ultrasound. The cause of plateau iris has been established by ultrasound to be
the result of an anterior positioning of the ciliary processes that prevents the iris from falling away from the trabecular meshwork following iridotomy.
Figure 3.35. A patient with plateau iris. In plateau iris, the relation of iris to ciliary body and lens as well
as corneoscleral angle can be shown and the ciliary processes demonstrated as anteriorly placed. (Top
figure: Courtesy of Charles Pavlin, MD.)
The angle between the iris and the cornea can be accurately measured, providing a quantitative gonioscopy to complement visual gonioscopy (Figure 3.36). This
can be useful in cases where visualization of the angle is compromised for whatever reason.
Postsurgical evaluation of glaucoma patients includes examining bleb configuration postfiltration surgery and examination of the position of filtration devices, such
as an Ahmed valve (Figure 3.37). Iris adhesions may develop, as in Figure 3.38, of a patient with a displaced haptic. The iris remained adherent to the cornea even
after the lens had been repositioned (Figure 3.38). The patient's visual symptoms disappeared following lens repositioning.
Figure 3.36. 50-MHz scans demonstrate single planes through the anterior segment that allow the cornea-iridal
angle to be measured. Top: This scan shows a patient with closed angle glaucoma. Middle: This scan shows a
patient with narrow angle glaucoma. Bottom: This scan demonstrates how lines can be used to actually
measure the angle in different meridians. This is termed digital gonioscopy.
Figure 3.37. Left: 50-MHz scan of a filtering bleb showing the bleb space as well as possible anatomic
changes of underlying sclera, which may include hypotonus changes of separation of the ciliary body
from the sclera, as is shown in this figure (arrow). Right: Position of an Ahmed Glaucoma Valve, which
is used to control intraocular pressure, lowering the chance of hypotony.
Figure 3.38. Top: A patient with iris touch noted superiorly and a partially dislocated lens. Middle: This pair of
ultrasonograms show the lens and haptic positions. Bottom: These scans show the same eye following lens
repositioning. The iris remains adherent to the cornea. (see color image)
P.66
Figure 3.39. A cross section at 50 MHz of the eye demonstrating the variation of ciliary process morphology
and position as distinct from the ciliary muscle. The arrow points to the ciliary body, whereas the arrowhead
indicates the ciliary processes. Note an anterior vitreal strand (VS), often seen as part of the zonular
suspension.
Ciliary Body
The ciliary body can be well defined at higher frequencies. Although there are three muscle groups anatomically, only two are characteristically seen with
ultrasound, that is, the sphincter component (Mueller's muscle) and the longitudinal component. In scanning, it is important to distinguish the muscle from the ciliary
processes (Figure 3.39), which take various morphologic patterns. The ciliary processes are best distinguished with serial or 3-D scans that allow the base to be
identified most easily (Figures 3.40 and 3.41; see also DVD). The processes tend to vary with globe dimensions (i.e., myopic or hyperopic eye) and with age and/or
pressure of the crystalline lens or intraocular lens.
Figure 3.40. Serial scans made coronally through the ciliary body, demonstrating ciliary processes in cross
section. Note the cyst in the scan in the upper right. (See also DVD.)
Clinically, we have found several conditions or situations where it is important to distinguish the ciliary
P.67
muscle anatomically. These are tumors, hypotony, accommodation, and glaucoma management. Tumors will be discussed later in the section on ocular tumor,
where it is essential to distinguish an iris melanoma from a ciliary body melanoma or to recognize extension of an iris tumor into the ciliary body.
Figure 3.41. A color rendering and 3-D animation of the scan seen in Figure 3.40, further
demonstrating the difference of the anatomy of the ciliary body and the processes. (see color image)
Figure 3.42. Separation of the iris from the sclera (iridodialysis) in a patient with an iridotomy and narrow angle.
Hypotony will be discussed later in the section on ocular trauma. It bears pointing out that the separation of the ciliary body from the sclera is a clinically important
distinction, whether it be from iridodialysis (Figure 3.42) or an isolated separation (Figure 3.43). The movement of the ciliary body is an important measurement
observation as it relates to production of accommodative changes in the lens. Thus, good visualization and measurement of the anatomy are required for both lens
implantation, particularly with intraocular contact lenses (ICL) or accommodating lenses, and placement of surgical incisions for presbyopic surgery, whether with
implants (Figure 3.44) or by laser (Figure 3.45).
Figure 3.43. Separation of the ciliary body from the sclera with hemorrhage and hypotony.
Figure 3.44. 50-MHz anterior segment scan with implants for treatment of presbyopia. This patient was
scanned 2 years postsurgery.
P.68
Figure 3.45. 50-MHz anterior segment scan demonstrating lesions placed for treatment of presbyopia by laser
surgery.
Physiologic Measurements
Physiologic changes can be measured both statistically and in real time, using ultrasound. In addition to measurements of blood flow that can be imaged with
Doppler or swept-scan techniques, physiologic changes in the choroid, ciliary body, and lens position and shape are routinely measured with ultrasound. The effect
of accommodation, pressure, and light can be shown, and the effects of pharmacologic agents can be documented.
The development of an instrument to demonstrate a forward or translational movement of the lens in accommodation led to the first use of an electronic interval to
measure axial length (33), which is useful in determining lens power for surgery.
Studies of accommodation in our laboratory led us to the catenary diaphragm theory of accommodation that better explains the paraboloid anterior lens curvature in
accommodation and helps explain how accommodating intraocular lenses can work, as well as why presbyopic surgery techniques are possible (67,81).
The definition possible with early radiofrequency 20-MHz A-scan ultrasound is shown in Figure 3.46, which was used to demonstrate a mass or translational
forward lens movement. Using VHF ultrasound, we have demonstrated how the anterior lens curvature is similar to the paraboloid curvature proposed by Koretz et
al. (82). This aspheric lens surface and its depth-of-field advantages explain many of the inconsistencies noted with other theories, such as the capsular or
Helmholtz theory.
Pupil diameter changes with light stimulation are shown in Figure 3.47. The position of the iris during accommodation (Figure 3.48), relative to the cornea-scleral
P.69
angle and the lens, can be critical in planning possible intraocular lens placement (83,84). Measurements of anterior segment dimensions and possible placement
of explants or surgical or laser incision for presbyopia are shown in Figure 3.49.
Figure 3.46. 20-MHz A-scan demonstrating both the RF and video traces that can allow very accurate
measurement of ocular dimensions. Note on the RF that the first quarter cycle of sound can be positive as the
sound enters the cornea and negative as it leaves the cornea, denoting the change of media and speed of
sound.
Figure 3.47. Iris position during light illumination and darkness. These and other physiologic measurements of
the iris can easily be made with very high frequency ultrasound.
Figure 3.48. Position of the iris and its relation to the lens during accommodation and unaccommodation.
Anterior lens curvature can also be measured.
Figure 3.49. The position of the lens equator relative to external landmarks can be measured and predicted
from B-scan imaging. Here the circle demonstrates the expected position of the lens equator for possible
presbyopic surgery correction; similar prediction is possible postcataract extraction or IOL position.
Effect of Pharmacologic Agents

Dilation, and the degree of dilation or contraction of the pupil, can be demonstrated with B-scan ultrasound, and, ordinarily, the pupil size can be accurately
estimated (Figure 3.50). Positioning the transducer parallel to the iris in contact B-scan ultrasonography can show the actual movement of the iris sphincter in a
graphic manner. Studies of the effect of drugs, such as pilocarpine or other agents on both the pupil and ciliary body, provide unique measurements of
pharmacologic effects on anatomic and vascular structures (Figure 3.51) (44,85).
Figure 3.50. B-scan demonstrating the pupillary opening, which can be measured in both accommodation or
pharmacologic or other physiologic conditions. See also Figures 3.47, 3.48.
P.70
Figure 3.51. Ciliary body area with color enhancement of the scatterer dimensions in the ciliary body in studies
of pharmacologic effects on the ciliary body. (see color image)
Lens
The normal lens has been described previously and is normally a clear (anechoic) space as a result of homogeneity of the lens cells. The surfaces have a high
acoustic index but are mostly specular reflectors, which may not be seen unless the transducer is orthogonal. On B-scan, the arc scan easily shows the anterior
surface but may show only a “highly reflective center” of the posterior surface. The reverse is true for a sector scan (Figure 3.52). For structures deeper than 5 to 6
mm, lower 10- to 25-MHz scanning frequencies are required. Fibrin or blood can convert the surfaces to a diffuse reflector and permit a better outline (Figure 3.13).
Figure 3.52. A 10-MHz sector scan shows only a small segment of a reverse image of the cornea (as a result of
transducer beam width) and of the anterior lens surface. The posterior lens surface is better appreciated as a
result of its concave “fit” of the lens surface and the sector. The arc scan is much better at showing anterior lens
surface but does not show the posterior lens surface well.
Absence and Displacement

Variations of normal lens position may be depicted ultrasonically. The absence of the lens from its proper position in patients suffering from trauma should initiate a
thorough search of the vitreous compartment for a displaced lens, as seen in Figure 3.53.
Cataract
The ultrasonographic appearance of a cataractous lens differs from the normal lens in that optical opacities also produce acoustic inhomogeneities. The A-scan
trace through the lens changes from a picture of acoustic homogeneity and sonolucent areas (with echo return only from the anterior and posterior lens surfaces) to
an acoustic heterogeneity, where numerous echoes are seen within the nucleus and cortex of the lens. The position of these echoes indicates the area of acoustic
change, which usually corresponds to the optical changes, and the degree of visual loss. Figure 3.54 demonstrates the separation of nucleus and cortex and shows
a posterior cortical cataract. B-scan ultrasonograms taken of a cataract demonstrate multiple intralenticular echoes. Pathologic changes responsible for these
acoustic alterations include nonuniform lens fiber swelling and water cleft formation.
Pre-intraocular and Post-intraocular Lens Implant

Ultrasonography is useful in documentation of the status of the anterior segment of the globe and accurate determination of the axial length of the eye, which
permits selection of dioptric correction with an intraocular lens.
The diameter of the cornea-iris angle and of the ciliary body sulcus is not uniform. Meridional scans will allow the axis of maximum diameter to be evlauated. This
may be useful for lens haptic placement to prevent lens movement, called propellering.
Figure 3.53. In this traumatized eye, the crystalline lens was completely dislocated and can be seen as a
rounded mass in a suitable plane. Differentiation from a tumor is usually not a problem and is facilitated by
having the patient move his or her eye, thus inducing lens movement.
P.71
Figure 3.54. Cataract shown at 10 MHz in a patient with a flat anterior chamber (left), and 50 MHz image of
cataractous lens in another patient showing high internal reflectivity within the lens (right).
The angle-to-angle and sulcus-to-sulcus measurements can be critical for sizing of IOL placement. Rondeau et al. (86) have shown that the coronal section of the
eye at the angle and at the sulcus is not round but generally oval. In addition, the white-to-white measurement that has been advocated is not an accurate
substitute for actual measurement, with errors reported up to 2 mm. The oval shape of the coronal section is usually longer vertically, but the axis does not
correspond exactly with the axis of astigmatism (Figure 3.55).
Figure 3.55. Top: 1. pupillary alignment vector for the scan series. 2. angle-to-angle measurement plane. 3.
sulcus-to-sulcus measurement plane. Bottom: The scan geometry for the semimeridional scan series is
demonstrated with lines representing individual scans. We have demonstrated that this coronal measurement is
commonly ellipsoidal, and the long axis can be determined for optimal placement of lens haptics to avoid
“propellering.”
A B-scan ultrasonogram of an eye with an intraocular lens in correct place is shown in Figure 3.56. Similarly, eccentric placement or dislocation can be shown as
well as haptic malposition, as shown in Figures 3.57 and 3.58. This can be helpful in directing surgical intervention. Certainly, “sizing” is the critical element in
developing better lens designs, and VHF ultrasound is the preferred method for its accuracy in measuring all ocular anatomies.
POSTERIOR SEGMENT ULTRASOUND

Synopsis
Vitreous hemorrhage as a result of diabetes, trauma, or other causes can appear of variable density. Blood in the formed vitreous can be best seen with a narrow
band transducer, using kinetic scanning for movement.
Light hemorrhage, endophthalmitis, or the vitreous changes in uveitis or central nervous system (CNS) lymphoma may be easily detected but acoustically
indistinguishable.
Retinal detachment hallmarks are high amplitude surface, always attached at the optic nerve and (except for giant tears) at the ora serrata. Retina movement on
kinetic scanning can indicate recent (fluid movement) or old (fixed, rigid) detachments.
Retinal detachment over tumors is characteristically bullous followed by a smooth tumor surface, which may or may not be attached to retina.
Posterior pole abnormalities, like proliferating membranes, small melanomas and nevi, or age-related macular degeneration (AMD), are better visualized at 20 to
30 MHz.
Severe vitreous hemorrhage may produce visual loss by chemical changes (i.e., hemosiderosis of the retina, hemosiderogenic syneresis of the vitreous) and
mechanical
P.72
changes (i.e., formation of strands or membranes producing permanent opacification of the media and retinal detachment). Biologic changes occur relatively early,
and to intervene surgically is often desirable. The decision for surgical intervention requires a careful appraisal of the extent of the pathologic changes in the
vitreous, including ultrasonic evaluation of the retina, choroid, and hemorrhage character.
Figure 3.56. A 50-MHz scan of the anterior segment demonstrating the lens haptics position following
uneventful cataract removal and lens implantation in the capsular “bag.”
Cibis (87) was a pioneer in demonstrating the changes that take place within the vitreous body secondary to blood and its breakdown products. Cibis and
Yamashita (88) documented retinal degeneration secondary to hemosiderogenic changes. Regnault (89) supplemented these studies with a description of the
temporal relationships in the formation of vitreous membranes, as did Machemer and Williams (90). Fibroblasts, which spread along the path of hemorrhage into
the solid vitreous, organize the vitreous along the hemorrhagically disrupted plane into membranes (90, 91, 92). These membranes, as they contract, tend to
produce stress on the retina that may lead to retinal detachment. Even if detachment does not ensue, the eye may remain visually useless because of optically
dense membranes.
Knowledge of the pathology cascade is important to the ultrasonographer because vitreous changes as a result of diabetes or trauma are not static, and repeat
evaluations are often indicated to determine present or impending retinal or choroidal detachment (93). In the normal eye, the vitreous appears as an acoustically
clear (anechoic) cavity. On the A-scan, no echoes are seen above baseline between the posterior lens capsule and the retina. On B-scan, the vitreous appears as
a uniformly sonolucent area. The retina in the normal eye appears on B-scan ultrasonograms as a smooth, concave, acoustically opaque (20) surface formed by
echoes arising from the vitreoretinal interface (Figure 3.59). These echoes are contiguous with, and inseparable from, the choroid-sclera complex. At 10 MHz,
hemorrhagic vitreous, which is opaque to optical examination methods, remains acoustically clear on B-scan at low gain. Denser hemorrhages appear as irregular,
opaque areas (Figure 3.60; see also kinetic scan on DVD). The location, extent, and density of vitreous hemorrhage can be shown by ultrasonography.
Figure 3.57. A slightly horizontally displaced intraocular acrylic haptic lens as a result of folding of one of the
haptics. Despite the minimal displacement, patient symptoms were severe enough to warrant lens replacement.
Figure 3.58. Top: A 50-MHz ultrasonogram of a lens haptic adherent to the iris with resultant traction of the iris.
Bottom: A similar situation as above with more severe retraction of the iris plane as a result of adhesions.
EXTENT AND DENSITY OF VITREOUS HEMORRHAGE

Light, diffuse, unclotted blood produces little echo response, so that the vitreous may appear sonolucent (Figure 3.61; see also DVD). Clumps of cells will produce
echoes
P.73
higher than the normal baseline echo of the vitreous. Low-amplitude echoes are usually best seen on the A-scan display or with a narrow band transducer with
B-scan (see later text), because the amplitude of the echoes from small clumps of cells is low (94) (Figure 3.61; see also DVD). The density of hemorrhage is
estimated from the character of echoes and the area of vitreous involvement, as determined from the B-scan. Thus, comparison of A- and B-scans is critical.
Movement of the eye causes these low-amplitude echoes to move freely within the globe and helps to distinguish them from more fixed vitreous membranes. A
more damped, lesser movement is apparent when the clumps are restrained by the “solid” primary vitreous.
Figure 3.59. Left: The posterior segment of an eye at 10 MHz and Right: at 20 MHz showing a smooth contour
on sector scan. The retina, choroidal, and scleral reflections form a smooth transition with separation of sclera
and Tenon's, often emphasized by higher amplitude echoes.
Figure 3.60. A moderately dense retrohyaloid hemorrhage with accentuation of the posterior hyaloid. (See also
DVD.)
The extent of clotted blood is more easily appreciated on the B-scan (Figure 3.62), and serial sectioning can be used to delineate the hemorrhage. A coagulum
within the hemorrhage is indicated by moderately high-amplitude, closely spaced echoes, giving the appearance of a solid mass. Changes in appearance can be
useful, if surgery is contemplated, particularly if rebleeding can be detected.
LOCATION AND SOURCE OF VITREOUS HEMORRHAGE

The localization of hemorrhage to areas of solid or fluid vitreous is based on the position and movement of the
P.74
hemorrhage. Hemorrhage localized in the anterior vitreous compartment is usually an indication of incorporated solid vitreous. Often blood along the posterior
limiting membrane or “hyaloid face” of the solid vitreous will form a “veil” or membrane that separates and outlines the fluid and solid compartments (Figure 3.63).
This veil may be studied with kinetic B-scanning. Kinetic scans are obtained by asking the patient to move his or her eye while fast sector scanning is performed.
The “after” movements of hemorrhage and membranes are observed after the eye has come to rest in its new position. Motion of hemorrhage in the solid vitreous
is damped more quickly than that of hemorrhage in the fluid vitreous. The final or resting position of an area of hemorrhage varies with gravity in fluid vitreous but
remains constant in solid vitreous.
Figure 3.61. A light hemorrhage along the posterior hyaloid of the vitreous with variable amplitude and usually a
marked variability on kinetic scanning. (See also DVD.)
Figure 3.62. B-scan at 10 MHz showing an area of dense coagulum, with a border of solid primary and more
diffusely represented tertiary vitreous.
Figure 3.63. B-scan of blood in the tertiary vitreous outlining the hyaloid face of the primary vitreous.
Figure 3.64. Recent hemorrhage in a patient with a solid vitreous demonstrating the location of a probable
bleeding site.
The importance of localizing the hemorrhage in the solid or fluid vitreous is important in that patients with spontaneous hemorrhage of light density limited to the
posterior vitreous have a good chance of clearing within a short time. Patients with dense hemorrhage into the solid vitreous, whether anterior or posterior,
regardless of etiology, clear more slowly, if at all. Such organized hemorrhages in one study had only about a 33% chance of clearing (94).
The position of hemorrhage relative to the limbusiris plane, lens, and optic nerve can be determined with B-scan ultrasound. In younger patients with a solid
vitreous, the source of bleeding can be frequently recognized as the point where echoes extend to the globe wall on the B-scan display (Figure 3.64) (42). A kinetic
scan of the moving eye can aid in tracing the point of origin of a vitreous hemorrhage. The vitreous attachments to the optic nerve and/or to the macula can often
be seen (Figure 3.65). When only attachment to the nerve is noted, vitreous veils along the posterior limiting membrane of the vitreous may simulate and resemble
retinal detachment.
VITREOUS VEILS (MEMBRANES)

Blood cell collection along the vitreous surfaces or the hyaloid can resemble a veil or even a membrane. Vitreous veils are usually distinguishable from
hemorrhagic clots by their pattern (Figure 3.66) and echo height (which is moderately high but usually lower than that of the retina). Occasionally, vitreous veils and
membranes may be difficult to distinguish from a localized retinal
P.75
detachment, particularly when retinitis proliferans is present (Figure 3.67). Tracing the veils to their attachment on the globe wall may be helpful; if the attachment is
anterior to the ora serrata, a vitreous membrane is indicated (or possibly a choroidal detachment), whereas if attachment to the ora serrata and the optic nerve
head is demonstrable, a retinal detachment is usually present. The B-scan is essential for tracing membranes, because it provides the topographic pattern of
amplitude contours that is not readily apparent from the A-scan alone. Kinetic B-scanning may graphically define a membrane by showing its failure to attach at the
optic nerve. On A-scan, echoes from a retinal detachment have a higher amplitude than most vitreal membranes. The retina echo is equivalent in height to the
sclera, whereas membranes are usually about 50% or less of the scleral echo height. Membranes, however, can be of variable amplitude, and they will appear as
incomplete “lines” on the B-scan, with occasional high amplitude segments (Figure 3.68; see also DVD).
Figure 3.65. A recent dense vitreous hemorrhage with a retracted primary vitreous—a fluid zone and
hemorrhage anterior to the retina. A retinal adhesion of vitreous to the macula is shown here. (See also DVD.)
Figure 3.66. Vitreous hemorrhage with complex pattern as a result of retracted vitreous and blood along the
syneretic vitreous cavity walls.
Figure 3.67. B-scan at 10 MHz demonstrating vitreous attachment to a proliferative membrane (arrow) in a
patient with diabetes. The echoes produce a characteristic “x” shape at the nexus of vitreous and proliferative
membrane. It is best seen with a kinetic scan. (See also DVD.)
Figure 3.68. 10-MHz B-scan of vitreous membranes.
P.76
PROLIFERATIVE DIABETIC RETINOPATHY WITH VITREOUS HEMORRHAGE

Vitreous hemorrhage secondary to proliferative diabetic retinopathy is a common indication for pars plana vitrectomy. Presurgical evaluation of such patients with
opaque media is greatly enhanced by ultrasound. B-scan ultrasonography in an eye with diabetic retinopathy can demonstrate (a) vitreous hemorrhage, (b) retinitis
proliferans, (c) vitreous veils or membranes, and (d) retinal detachment.
As has been mentioned in the previous section, vitreous filled with diffuse hemorrhage often appears acoustically clear on B-scan ultrasonograms. A nonfocused
10-MHz transducer is best to demonstrate diffuse low-amplitude vitreous hemorrhage. The narrow band B-scan is easiest to use and to image low-amplitude
echoes. The A-scan trace through the vitreous may stay at baseline at low gain. Denser vitreous hemorrhages present an ultrasonic picture varying from scattered
dots throughout the vitreous to a dense sheet of white opacities filling the vitreous compartment. They have indistinct borders and a relatively amorphous
appearance. In simple vitreous hemorrhage, the lens is in its normal position and the retina is in place.
Retinitis Proliferans
An area of retinitis proliferans appears on the B-scan ultrasonogram as an echo configuration forming a stalk that arises from the retina (Figures 3.67 and 3.69; see
also DVD). The echoes often tend to diverge as the stalk extends forward in the vitreous. At the distal end of retinitis proliferans, areas of the nexus of retinitis
proliferans and vitreous membranes may be demonstrable ultrasonically. Often an “x-shaped” area is seen at this nexus, where vitreous membranes and
proliferans intersect.
Retinal Detachment
A-scans and kinetic B-scans are both useful in differentiating vitreous membranes from retinal detachment. The kinetic B-scan will often show the vitreous body
and membranes floating away from the disc, while the retina remains attached. It should be remembered that solid vitreous may be firmly attached to the disc
and/or the macula (usually in young patients) and, with hemorrhagic collection along the hyaloid, can mimic a detachment on both static and kinetic B-scans.
Figure 3.69. B-scan of a proliferative membrane attached to retracted vitreous; both wide (left) and narrow
(right) bands are shown, with the narrow band giving greater sensitivity to the area of vitreous traction. (See
also DVD.)
Retinal Schisis
Schisis of the retina is a splitting of the neural layers of the retina producing an elevated, convex (usually peripheral) echo pattern similar to a total retinal
detachment. The thickness of the “split” retina cannot usually be differentiated from full thickness retina (Figure 3.70).
Boldt (95) has pointed out that pressure of the B-scan probe directly over the schisis area will show that a schisis cavity does not collapse, unlike retinal
detachments, which do collapse, as a result of a higher intraschisis pressure than vitreous pressure.
Traumatic Vitreous Hemorrhage

Vitreous hemorrhage, as a result of trauma, is discussed again in a later section, Ocular Trauma.
MISCELLANEOUS VITREOUS ABNORMALITIES
Primary Vitreous Detachment (PVD) and Asteroid Hyalosis

Normal aging of the vitreous is marked by retractions of the solid vitreous (primary vitreous detachment, or PVD) and a clear fluid replacement posteriorly in the
“tertiary vitreous” (96). With B-scan, low-amplitude echoes in clear vitreous can be seen, and on kinetic scans, areas of residual attachment to the retina may be
seen. This feature can be useful for looking for potential retinal tears when a history of flashing lights and floaters with opaque media is given.
Deposition of calcium soap crystals (asteroid hyalosis) produces high-amplitude echoes scattered throughout the solid or fluid vitreous and is usually accompanied
by vitreous retraction so that a clear tertiary vitreous zone is seen on B-scan.
These calcium soap particles show widely scattered low-amplitude spikes on the A-scan. On kinetic A-scans, they can be seen to “dance” or move quickly. On the
B-scan, asteroid hyalosis appears as a plethora of echoes
P.77
in the vitreous cavity (Figure 3.71). As Jaffe (97) has noted, asteroid hyalosis tends to occur preferentially in the primary vitreous and thus is often best seen
posterior to the lens. The peripheral vitreous may be acoustically clear as noted previously.
Figure 3.70. Two views at 10 MHz of a schisis cavity. The cavity has higher intraschisis pressure than the
vitreous, thus should always remain smoothly convex. It is important to be perpendicular to the surface to
maximize the B-scan. The wall thickness may be indistinguishable from retina, and balloting with the transducer
may be valuable.
In the aging vitreous, cholesterol crystals (cholesterolosis) and dissolution of the vitreous can be seen on B-scan with real-time or kinetic scanning. These deposits
seem to float and settle like the “snowflakes” in a decorative globe, unlike the suspended, highly reflective echoes of asteroid hyalosis.
Amyloidosis of the Vitreous

Extensive experience with ultrasonography in patients with amyloidosis of the vitreous, a rare abnormality, has not been obtained. We have examined one patient
with pathologically confirmed amyloidosis of the vitreous in the fellow eye. There were irregular strandlike echoes in the vitreous, which were of low amplitude on
the A-scan. This pattern is similar to that seen in vitreous hemorrhage, endophthalmitis, and hyphema.
Figure 3.71. Typical asteroid hyalosis pattern of high amplitude reflector and a clear zone separating the
retracted vitreous from the retinal surface.
Endophthalmitis
The B-scan pattern seen in patients with endophthalmitis is characteristic, though not diagnostic. The vitreous pattern will have the synchesis clefts emphasized by
bacterial or detritus collections along their margins, producing a diffuse polyploid matrix (Figure 3.72; see also
P.78
kinetic scan on DVD). The pattern may resemble convex curves that may straddle the ora serrata and simulate a choroidal detachment. This particular acoustic
pattern has not been noted in any other ocular abnormality and seems related to bacterial spread different from that seen by pressure or diffusion of blood. In
many, if not most cases, however, endophthalmitis is similar to a vitreous hemorrhage and, at least in early stages, is best diagnosed with caution, a good history,
and a clinical indication. The extremely rare condition of CNS lymphoma should also be borne in mind, because it can also present as vitreous debris (Figure 3.73).
Figure 3.72. Two scans of a patient with endophthalmitis shown with both a wide band (top) and a narrow band
(bottom) transducer. The narrow band transducer provides better sensitivity of the vitreous changes. Note that
the vitreous appears to separate from the retina with a point of contact remaining at the optic nerve. This is best
seen on kinetic scan. (See also DVD.)
Figure 3.73. B-scan of a CNS lymphoma, which is relatively absorbent like a granuloma. There is also vitreous
debris present.
ULTRASONOGRAPHY AND VITREOUS SURGERY
RETINAL ABNORMALITIES
Retinal Detachment
Modern surgical techniques for retinal detachment have been considerably influenced by the widespread use of indirect ophthalmoscopy and ultrasonography,
which permits a thorough evaluation of the peripheral fundus. Since optical evaluation techniques, such as indirect ophthalmoscopy, are useful in examination of
the vitreous and retina only if the ocular media are clear. They are useful in examination of the subretinal space, only if the overlying retina is transparent and the
subretinal fluid is clear. Ultrasonic evaluation techniques are not subject to these limitations. B-scan ultrasonic evaluation of the vitreous, retina, and subretinal
space (especially in eyes with clouding or opacification of the ocular media) adds a significant dimension to the diagnosis of retinal detachment and to the
management of patients with vitreoretinal abnormalities.
The use of A-scan ultrasound in the diagnosis of retinal detachment was first described by Oksala and Lehtinen (98) in 1957. B-scan ultrasonography of retinal
detachments was subsequently reported by Baum (99), Baum and Greenwood (100), Purnell (28), Coleman and Jack (101), and numerous others.
Rhegmatogenous Retinal Detachment

The retina normally appears on B-scan ultrasonograms as a smooth, concave, acoustically opaque surface formed by echoes from the vitreoretinal interface.
These echoes are contiguous and, with the attached retina, inseparable from echoes from the choroid and the sclera. A detached retina also appears on B-scan
ultrasonograms as a thin, continuous, acoustically opaque line of echoes separate from, and anterior to, echoes from the wall of the globe (Figure 3.74). A relatively
flat detachment has a narrow, acoustically empty (sonolucent) space between detached retina and the globe wall. A highly elevated, totally detached retina
appears as convex bullae extending far into the vitreous from attachment points at the ora serrata and at the optic nerve. The space posterior to the elevated retina
is sonolucent (anechoic).
The extent of a detachment (whether partial or total) is ascertained by performing ultrasonic B-scans in serial, horizontal, or radial planes of the eye. Starting above
the superior limbus, scans are made at roughly 2-mm intervals along the vertical dimension of the globe. The serial sections thus obtained differentiate a true
retinal detachment from simulating structures, such as a
P.79
choroidal detachment, which extends anterior to the ora serrata, and from a vitreous veil (hemorrhage along the posterior hyaloid face), which cannot usually be
traced back to the optic disc. The amplitude characteristics of returned echoes, best observed on the A-scan, further differentiate retinal detachments from vitreous
membranes, because retinal echoes have uniformly higher amplitude than do vitreous membrane echoes. Retinal echoes are approximately as high in amplitude
as the normal posterior globe wall echoes, whereas membrane echoes are typically only about 50% of this height, as noted previously. If the anterior segment is
particularly dense acoustically, as in a calcified cataract, the amplitude of the retinal echo can obviously be greatly reduced.
Figure 3.74. A relatively flat detachment exhibiting uniformly high amplitude. This will usually move freely on
kinetic scan.
B-scan ultrasound indicates the thickness of the detached retina and the extent of retinal organization and shrinkage. A freshly detached retina appears as a thin
white line, equal in length to the scleral arc from ora to ora. On kinetic scans it is flexible and dances with eye movement (Figure 3.75; see also DVD). In
long-standing detachments, the retina is thickened and its overall length often shrinks to form a cord from the optic disc to the ora serrata, thus often forming a
funnel-shaped or “morning-glory” configuration. A contracted retinal detachment of this type is shown in Figure 3.76. Cystlike structures of the retina in
long-standing detachments are indicated ultrasonically by a thickened and convoluted echo pattern (Figure 3.77).
Rhegmatogenous retinal detachments are usually the result of traction of vitreous membranes or bands on the retina. Vitreous membranes and bands, as noted
previously, are demonstrable ultrasonically, as are sites of vitreous membrane attachment and their associated traction of the retina, indicating the “stress sites” at
which retinal holes may be expected to have occurred.
Figure 3.75. A freshly detached retina is highly reflective along its entire surface, as is seen on both A- and
B-scans. (See also DVD.)
Figure 3.76. A long-standing retinal detachment with blood in the preretinal and postretinal spaces, producing a
“morning glory” shape. The retina is rigid on kinetic scanning, and contraction membrane can be seen
connecting the retinal leaves.
Retinal detachment may be seen in conjunction with choroidal detachment. Choroidal detachments are limited posteriorly by the ampullae of the vortex veins and,
as noted in a later section on choroidal effusion, will extend anterior to the ora serrata. Both retinal detachment and choroidal separation will have high amplitude
walls and are best distinguished by the B-scan pattern (Figure 3.78). See Choroidal Effusion in later section.
Nonrhegmatogenous Retinal Detachment
Retinal Detachment Secondary to Choroidal Melanoma.

The use of binocular indirect ophthalmoscopy and scleral transillumination significantly aids in the clinical diagnosis of retinal detachment secondary to choroidal
malignant melanoma. Mistaken diagnoses do occur, however, and a significant incidence of operation of retinal detachment in eyes containing choroidal melanoma
occurs (2% in one series) (102). As Norton (103) has emphasized: “The most serious misdiagnosis of retinal separation is the failure to recognize an underlying
malignant melanoma of the choroid as the cause of the retinal elevation. Once the patient is subjected to retinal surgery and the integrity of the sclera disrupted, the
patient's life may be in jeopardy. On the other hand, removal of an eye with an idiopathic retinal detachment because of misinterpretation of the fundus changes is
a similar tragedy.”
P.80
Figure 3.77. 10-MHz scans of an eye with a long-standing detachment. Left: Immersion B-scan shows total
retinal detachment with membranes and with cystlike structures produced by coaptations of the walls and
resulting adhesions. Right: Coronal plane shows retina in cross section with dense vitreous debris.
The use of B-scan ultrasonography can reliably demonstrate the presence or absence of a tumor underlying a retinal detachment. Upon serial sectioning, the
profile of a choroidal mass can be detected and characterized. Tumors appear as acoustically solid (echoic) masses at low examining frequencies and can be
differentiated from hemorrhages or exudative elevations, which are acoustically clear. Because certain detachments overlie tumors, patients with suspected
secondary detachments benefit from B-scan ultrasonography to detect the obscured etiology of the detachment. All cases of retinal detachment do not necessarily
require ultrasonic evaluation. This diagnostic test is most valuable in selected cases of clinically atypical retinal detachment. The following features are clinical
suggestive, but not always diagnostic, of a nonrhegmatogenous retinal detachment:
Figure 3.78. A retinal detachment in the presence of choroidal detachments as well. Both have uniformly high
reflective surfaces, but the retina attaches to the nerve, and the choroidal elevations extend anterior to the ora.
1. Absence of breaks: Retinal breaks were not seen by the clinician in 22 of 26 eyes with choroidal melanomas enucleated after detachment surgery (102). Of
the four eyes in this series in which breaks were suspected clinically, they were not found on pathologic examination. Retinal breaks, of course, do not rule
out a choroidal melanoma.
2. Smooth bullae and shifting fluid.

3. Elevated intraocular pressure: There is a significant incidence of elevated intraocular pressure in eyes with choroidal melanoma.
4. Large iris nevi: The association of iris nevi and melanomas of the choroid has been discussed by Reese (106).
B-scan ultrasound can be an essential test in suspicious or atypical retinal detachment to rule out a solid tumor under the elevated retina. Serial sections should be
used to localize the tumor, which will often appear as an acoustically opaque, hyperechoic mass (see Tumors in later section). Associated ocular changes, such as
hemorrhage, can be demonstrated ultrasonically. Ocular tumors will be discussed later.
Other Secondary Retinal Detachments.

Retinal detachments may also be secondary to inflammation, exudation, and cicatricial conditions. The subretinal space in these conditions is acoustically clear,
except in the region of the inflammatory focus or cicatrix. That is, the subretinal fluid is acoustically quiet (unless hemorrhagic), but tissue
P.81
abnormalities may be seen. Evidence of inflammation may also be detectable in the overlying sclera and Tenon's space, or in the optic nerve sheath (Figure 3.79).
Figure 3.79. Scleritis of the posterior pole produces an accentuation of the Tenon's surface posterior to the
sclera and is commonly traceable to the meninges as they pass posteriorly along the optic nerve. This
accentuated reflective “space” is often called a “T-sign” (arrow).
POSTOPERATIVE SITUATIONS
After retinal detachment surgery, numerous situations may arise in which management may be facilitated by ultrasonography.
Occasionally, after an encircling procedure, especially if extensive cryotherapy has been used, choroidal effusion or hemorrhage may occur and cause angle
compromise with elevation of intraocular pressure. Usually, the choroidal effusion with hemorrhage is visible with the indirect ophthalmoscope. However, a cloudy
cornea, a miotic pupil, or a cataract may prevent visualization. B-scan ultrasonography in this condition shows the pathognomonic appearance of a choroidal
effusion, that is, a smooth, convex, circumferential elevation straddling the ora serrata. Choroidal effusion may be differentiable from choroidal hemorrhage with
A-scan quantification techniques. Demonstration of a localized choroidal effusion may indicate the site for posterior sclerotomy, if this should be clinically
warranted.
Figure 3.80. Left: B-scan showing the indentation of the globe wall by a sponge element. The retina remains in
place. Right: A sponge element with residual retinal detachment.
The extent of settling of the retina after detachment surgery may be difficult or impossible to evaluate optically because of opaque media. In this situation, B-scan
ultrasonography can provide accurate information as to the position of the retina (Figure 3.80). Posterior migration of an encircling element can be shown
acoustically.
B-scan ultrasonography can also demonstrate the presence of an encircling element or large, local implant in patients where the history is unclear and the media
are cloudy, as shown in Figure 3.81.
Figure 3.81. 10-MHz B-scan of eye after sponge implant for treatment of detached retina. Note the deformation
of posterior globe supertemporally as a result of implant and the presence of moderate vitreous debris. Right: A
residual area of detachment.
P.82
Figure 3.82. Residual perfluorocarbon bubbles in fluid vitreous after surgery. The retina is attached. (See also
DVD.)
POSTOPERATIVE EVALUATION OF THE VITREOUS SUBSTITUTE

The use of gas and tamponade liquids in retinal detachment and vitreous surgery produces characteristic echogenic and conductive artifacts. Gas bubbles are
reflective and thus distort or prevent posterior pole evaluation. Perfluorocarbon liquid is usually removed at the end of surgery, but small bubble remnants may be
seen (Figure 3.82). They are highly reflective, often looking like small foreign bodies, either in the ciliary body region or at the posterior pole (Figure 3.83).
Silicone filled eyes produce a characteristic lengthening of the eye because of the slower velocity of sound transmission than saline or vitreous. Our measurements
of velocities are 972.0 meters per second for 1,000 CS silicone at 37°C and 978.5 meters per second for 5,000 CS silicone at 37°C (standard deviation = 4.5).
Silicone is less dense than water and floats to the top of the vitreous compartment, giving a “split” type of B-scan when viewed at the vertical or 6 to 12 meridian
(Figure 3.84). Silicone can produce acoustic artifact changes that are difficult to interpret but, usually, positioning will help make the area of interest accessible
acoustically.
Retinitis Proliferans
The ultrasonographic appearance of retinitis proliferans has been discussed in earlier section on vitreous abnormalities.
Macular Edema
In severe macular edema, or in the “sunny-side-up” stage of Best disease, a cystic structure may be ultrasonically demonstrable. The B-scan shows a convex
anterior projection of the vitreoretinal interface echo line in the macular region, followed by a localized sonolucent (anechoic) area (Figure 3.85). Although not as
useful as optical coherence tomography (OCT) in demonstrating vitreoretinal traction and edema, in most cases the traction may be visualized with B-scan
ultrasound. Certainly with opaque media, the retinal changes can be well visualized with 20-MHz transducers (see Figure 3.153).
Figure 3.83. Small perfluorocarbon fragments are seen in these 50-MHz images. These small bubbles produce
ringing artifacts and are generally only seen at very high frequencies.
Age-related Macular Degeneration (AMD)

B-scan ultrasonography, in cases of elevated disciform macular degeneration, shows an elevated lesion at the posterior pole (Figure 3.86). The typical AMD lesion
is usually too small to differentiate. The acoustic appearance of the interior of the lesion may vary. In predominantly hemorrhagic lesions, a cystic appearance with
a bright line of front surface echoes separated from the posterior wall of the globe by a sonolucent (anechoic) zone is noted. An acoustically hypoechoic opaque
appearance is noted in fibrotic, scarred lesions. This range of appearances is also characteristic of a small choroidal melanoma. The phenomenon of choroidal
excavation, the appearance of replacement of the choroid with tumor (see discussion on choroidal tumors), often seen with choroidal malignant melanomas, is not
seen, in our experience, with disciform macular degeneration. Because disciform lesions are restrained by Bruch's membrane and do not involve the choroid, one
would not, on theoretic grounds, expect choroidal excavation to occur in this condition.
P.83
Figure 3.84. Upper left: Contact 10-MHz B-scan in horizontal plane of globe filled with silicone oil. Artifactual
increase in axial length results from the slower speed of sound in oil. Upper right: Vertical scan of the same
eye 10 months later shows artifactual distortion of the superior portion of the globe as a result of the presence of
the oil. Bottom: 10-MHz B-scan in horizontal plane of inferior (aqueous fluid filled) globe is not distorted. Note
suggestion of peripheral choroidal/retinal detachment. There is also vitreous debris of moderate amplitude.
Figure 3.85. 20-MHz image using the Quantel Cinescan demonstrating a vitreous traction membrane in a
patient with cystoid macular edema.
Figure 3.86. 10-MHz B-scan through the macula of a patient with disciform AMD (arrow) and vitreous
membranes.
P.84
Evaluation of the A-scan trace through the lesion is also important for differentiation. In disciform macular degeneration, irregular echoes, often of low amplitude,
follow the high-amplitude echo from the retinal surface, and in melanomas gradual echo decay follows the high-amplitude echo from the retinal surface, although a
small melanoma may be too small to show this well on A-scan (see Small Melanomas in later section).
Choroid in Age-related Macular Degeneration

A new technique for evaluating the macular area, especially the choroid behind the macula, involves the use of parameter image tissue characterization. Defining
the choroid can be useful in documenting progressive aging changes as well as effects of therapy.
Parameter imaging and tissue characterization were discussed in Chapter 2. Choroidal measurement is discussed in later section on small melanoma evaluation.
Normal choroid and globe wall thickness can best be seen at 20 MHz and higher frequencies (Figure 3.87) (107).
Figure 3.87. Left: In vivo 22 MHz B-scan through the posterior pole of the eye, with measurements of retinal,
choroidal, and scleral thickness. Right: Comparative histologic section. Note relative thinness of choroid in the
absence of perfusion. (see color image)
Coleman, Rondeau, et al. (105) have developed a special analytic mathematical modeling technique for imaging the “normal” choroid using wavelet analysis. The
definition of the choroidal thickness and appreciation of vascular channels is outlined by the scattering elements that surround them. Digital RF scans of the
macular region are acquired, and power spectrum and wavelet analysis were used to identify changes in the backscatter architecture, providing uniquely enhanced
images, as shown in Figure 3.88. This technique is being explored for evaluation of disease progression as well as effects of therapy.
UVEAL TRACT ULTRASOUND AND TUMORS
Retinoblastoma
Retinoblastoma is the most common primary intraocular malignancy in children. A highly malignant retinal tumor found in infants and young children, it usually
presents with focal areas of calcification within the tumor. B-scan ultrasonography is frequently used for the initial and follow-up evaluation of retinoblastoma.
Ultrasound can clearly reveal calcium, which is characterized by highly reflective foci within the tumor or vitreous. When small, the tumors are smooth, dome
shaped, and have low to medium reflectivity. As the tumors grow, they become more irregular in configuration and more highly reflective as the amount of calcium
accumulates (Figure 3.89). There may also be associated retinal detachment. Ultrasound has become a useful and cost-effective way to follow these tumors as
treatment is delivered. Baseline tumor size measurements and tumor locations are obtained, and these parameters are monitored closely during and after
treatment. Ultrasound may have its major application in following therapy of such tumors.
Figure 3.88. Wavelet analysis of a parameter image of scatterers that outline the interstitial elements of the
choroid (ch), thus also outlining the vascular elements. Top: This scan (a) is of a normal subject. Bottom: This
scan (b), also of macular degeneration, is of drusen (sd) in a patient with wet AMD but taken in an area adjacent
to the macula.
P.85
Figure 3.89. 10-MHz ultrasonogram of an eye containing a retinoblastoma. It is estimated that over 95% of
retinoblastomas have calcium deposition, which is highly reflective. In this case, the short arrow indicates a
concentrated area of calcium in a surrounding area of tumor. The long arrow indicates a smaller calcium
element. Note the shadowing behind the concentrated calcium.
Choroidal Effusions
Synopsis
Choroidal effusion and hemorrhage usually extend anterior to the ora serrata and thus are distinguishable from a localized retinal detachment, which is limited by
the ora. The posterior extent of a choroidal detachment is usually limited by the ampullae of the vortex veins.
Hypotony and ciliary body separation are best visualized using VHF or UBM ultrasound.
Choroidal effusions (detachments) occur occasionally following intraocular surgery and often after trauma. In some patients, these cannot be diagnosed clinically
because of a cloudy cornea, hazy aqueous or vitreous, or miotic pupil. A choroidal effusion appears ultrasonographically as a convex surface of echoes extending
into the vitreous compartment from the globe wall in any quadrant. In a fully developed choroidal effusion, the echo lines may extend into the central vitreous from
each side of the eye (baseball-stitch sign) and may even appear to touch (“kissing choroidals”) (Figure 3.90). The choroidal effusion is nearly always seen to
straddle the ora serrata anteriorly, and at its juncture with the posterior wall echo forms an acute angle. It is limited posteriorly by a vortex vein. The space internal
to the line of echoes is acoustically clear (anechoic). This characteristic differentiates a choroidal detachment from a choroidal hemorrhage, which shows
low-amplitude echoes in the choroidal space, similar in amplitude to echoes obtained in vitreous hemorrhage. Infrequently, however, large choroidal effusions will
show scattered low-amplitude echoes in the subchoroidal space, indicative of organized cells or exudate (Figure 3.91). Other ocular abnormalities, such as retinal
detachment, may coexist with choroidal effusion (Figure 3.78).
Figure 3.90. Typical “kissing choroidal” elevations. They are smoothly convex as a result of the pressure
producing them. They are limited posteriorly by the vortex vein ampulla and extend anterior past the ora serrata.
A flat or shallow anterior chamber is most commonly encountered as a postoperative condition. It may also appear in other situations, such as penetrating injury,
P.86
pupillary block, dislocated lens, swollen lens, choroidal hemorrhage, or intraocular tumors. Most flat chambers are associated with a serous choroidal detachment
(effusion). Examination of the posterior segment may therefore provide information useful for the diagnosis and management of each individual case.
Figure 3.91. Choroidal hemorrhages demonstrating the lower amplitude echoes from blood in the choroid. The
clear zone, noted in one bullae, indicates hemolysis, and this feature can be used to indicate areas and timing
for drainage.
Figure 3.92. Upper left: Immersion B-scan shows thickened choroid posteriorly and possible cyclodialysis cleft
(arrow) temporally. Upper right: 50-MHz ultrasonogram shows internal echoes in lens consistent with cataract.
Bottom left: Ciliary body is detached nasally (arrow). Bottom right: Complete cyclodialysis temporally with
complete separation of the iris root (arrow).
In evaluating an eye with a flat or shallow anterior chamber, it is useful to search for relevant abnormalities in the anterior segment, to search carefully for evidence
of external fistulization, to measure the intraocular pressure, and to determine the presence or absence of a choroidal detachment in the posterior segment.
Factors, such as corneal edema, hyphema, cataract, miosis, and vitreous hemorrhage, may intercede to prevent adequate visualization of the posterior segment.
Such problems are likely to be encountered in the more complex cases where information regarding the posterior segment is most needed.
Figure 3.93. An example of a patient's anterior segment with pupillary block glaucoma is shown, with evidence
of occluded angles and a flat iris.
An eye with a flat chamber following a filtering procedure for angle-closure glaucoma always raises the possibility of malignant glaucoma. In this condition, aqueous
humor is trapped within the vitreous compartment, despite a patent iridectomy, and the lens is anteriorly displaced (Figure 3.92). Malignant glaucoma is
distinguished from postoperative pupillary block glaucoma (Figure 3.93) and is characterized by lack of a patent iris coloboma and aqueous humor trapped in the
posterior chamber causing iris bombé (Figure 3.94), without anterior displacement of the lens. In malignant glaucoma, the intraocular pressure is usually high, but it
may fall within the normal range during the early phase. The tension in the flat chamber syndrome is generally soft but may transiently rise to a normal level in
some cases, despite the continued presence of a choroidal detachment. Hence, a differential diagnosis between malignant glaucoma and the flat chamber
syndrome cannot always be made by tonometry alone. In equivocal cases, the presence of a choroidal detachment would be diagnostic of the flat chamber
syndrome, whereas the absence of such a finding would strongly suggest incipient malignant glaucoma. Thus, when visualization of the posterior segment is not
possible, and the intraocular pressure is not elevated, ultrasonography may be helpful.
Choroidal Thickening and Choroidal Folds

Choroidal folds can be seen generally as thickened choroid at 10 MHz and may simulate a displaced implant or band or even a foreign body on B-scan. Higher
frequencies can help document the choroidal thickening, but 10 MHz is required to define the retrobulbar space to determine whether a retrobulbar mass is
present.
Hypotony and Ciliary Body Detachment

Very high frequency ultrasound or UBM is essential for evaluating any form of ciliary body elevation or separation from the sclera. A 10-MHz B-scan can detect
ciliary body separation in some cases, if a standoff is used, but the definition is clearly suboptimal compared to the higher frequency examinations.
P.87
Figure 3.94. Iris bombe is demonstrated with adhesions of the sphincter to the lens capsule and typical bulging
of the iris, and a narrowed iridocorneal angle.
The characteristic picture on B-scan of a ciliary body detachment is shown in Figure 3.95. An anechoic region is characteristic of the separation of detachment and
its extent can be measured in terms of meridians involved with serial scans. We have found that hypotony typically has two or more quadrants of ciliary body
detachment involved (108). In prolonged hypotony, the surgical management (either suturing or gas tamponade) can thus be directed by the determination of the
location and extent of the separation.
Detachment of the ciliary body can be seen in both hypotony and trauma of the eye. A case of detachment is shown in Figure 3.96, where the iris root was totally
separated.
OCULAR TUMORS
Synopsis
CHOROIDAL TUMORS
Melanoma, hemangioma, and metastatic carcinoma are the most common choroidal tumors.
Distinguishing Characteristics
Melanoma: solid; shapes vary, but convex to collar-button; most common; reflectivity decreases rapidly on A-scan. Hemangioma: solid; convex to flat shape;
A-scan reflectivity remains high amplitude as a result of vascular clefts.
Metastatic: solid; placoid to convex shape; A-scan reflectivity moderate and uniform, that is, less than hemangioma but posterior segment higher than melanoma,
which is more homogeneous.
Tissue characterization techniques can subclassify melanomas in terms of lethal extravascular matrix (EVM) patterns. Observation for growth of small melanomas
and the follow-up of treated tumors are important to both diagnosis and treatment. Accurate measurement of thickness, perhaps including 3-D ultrasound for
volume, is clinically important.
Accurate diagnosis of choroidal masses remains a challenge to ophthalmologists. Lesions demanding different therapy may have similar clinical appearances.
Even with the use of indirect ophthalmoscopy and scleral depression (109); scleral transillumination (110); fluorescein angiography (111, 112); visual field studies
(113, 114); and MR, CT, and OCT (115, 116, 117, 118, 119), inaccurate diagnosis of ophthalmoscopically visible masses still occurs. The recent Collaborative
Ocular Melanoma Study (COMS) claims an accuracy of 99% with all modalities used for melanoma diagnosis confirmed for enucleated eyes, but, significantly, any
eye with a questionable diagnosis was not included in the study. Diagnosis of tumors with concurrent trauma and hemorrhage or other opacity thus probably
remains in the accuracy range of approximatley 97% (120, 121, 122, 123, 124).
Figure 3.95. Ciliary body detachment at 10 MHz (top) and at 50 MHz (bottom). The 10-MHz scan shows
typical changes, such as a shorter eye and choroidal thickening in hypotony. The 50-MHz scans show the ciliary
body detachment with greater accuracy.
P.88
Figure 3.96. Separation of the ciliary body from the sclera in a typical case of hypotony. Meridional scans can
outline the full extent of the separation and can be very helpful, not only in diagnosing the condition, but in
selecting surgery location and monitoring resolution.
TABLE 3.4 Differential Diagnostic Criteria for Choroidal Tumors

Characteristic Malignant Melanoma Metastatic Carcinoma Angioma/Hemangioma
Morphologic
Size Varies from 0.5 mm to Variable Usual range 0.5 to 2

15 mm in height mm in height
Shape Convex commonest, Convex or placoid Convex most

polyploidy infrequent common
Location Anywhere in globe Usually posterior Usually posterior pole

near optic nerve
Associated Frequent Frequent large May be associated

ocular nonrhegmatogenous nonrhegmatogenous retinal detachments
changes retinal detachments retinal detachments
Evidence of Extraocular extension Multiple lesions None

spread can be shown at times common
Changes with Growth or regression Growth or regression Growth or regression

time demonstrable demonstrable demonstrable
Acoustic
Boundary Sharp, smooth leading Sharp, smooth leading Sharp, smooth

properties edge edge leading edge
Acoustic quiet Polypoid: solid, no quiet Usually solid at 10 and Quiet zone seen even
zone zone; convex: usually 20 MHz at 5 and 10 MHz in
has quiet zone at 15 and large lesions; small
20 MHz lesions may appear
solid
Choroidal Prominent Not seen Not seen

excavation
Absorption Often shows shadowing Occasionally shows Seldom seen

effects shadowing
A-scan Increasing attenuation Little attenuation Relatively constant,

amplitude through tumor, sharp through tumor high amplitude spikes
decay after initial echo
Texture of Relatively closely spaced Relatively closely Relatively wide

A-scan pattern echoes spaced echoes spaced echoes
Ultrasound is particularly useful in distinguishing solid choroidal tumors from simulating lesions filled with blood. Echo amplitude studies with A-scan can usually
differentiate an acoustically “solid” ocular tumor from conditions simulating an ocular tumor, such as choroidal detachment or retinal detachment with subretinal
hemorrhage or fluid. In addition, A- and B-scan acoustic criteria may help identify the various types of choroidal tumors (Table 3.4).
The A-scan ultrasonographic appearance of a choroidal melanoma was first reported by Oksala (125) in 1959. Further discussions of choroidal tumor patterns were
published by Oksala (126), Ossoinig (127), and Poujol (128). The B-scan appearance of choroidal tumors has been described in papers by Baum (129), Purnell
(28), Coleman et al. (130), and more recently by Greenwald (131), Ossoinig (132), Byrne (133), and DiBernardo (134).
The important role of ultrasonography in choroidal tumor diagnosis in eyes with opaque media has been well documented by the previously mentioned researchers.
P.89
Eyes with opaque media often harbor unsuspected malignant melanomas (10% of enucleated eyes in one study from the Armed Forces Institute of Pathology
contained malignant melanomas) (135) and thus may present a diagnostic problem. Ultrasound is the only test for preoperatively discerning a tumor in such eyes,
other than MR or CT scans, which are not as accurate or economically reasonable.
In a famous series of 529 eyes with clear media and ophthalmoscopically visible lesions clinically thought to be malignant melanomas that were enucleated (prior
to the use of ultrasound), the clinical diagnosis was incorrect in 100 (19%) (136). Combined A- and B-scan ultrasonography of eyes with suspected intraocular
tumors provides morphologic and acoustic characteristics, which can help significantly in making a correct diagnosis in these cases, just as in those with opaque
media.
Consistent and systematic techniques must be used to follow tumors over time. Although the information presently obtainable with ultrasound does not allow
absolute “tissue diagnosis,” tumor differentiation can be performed with a high degree of reliability by combining ultrasonic information with a knowledge of tumor
characteristics. Newer tissue characterization data are described later that not only diagnose melanomas but can provide high sensitivity in identifying “high-risk”
melanomas (56).
In general, there are three tumor types and one simulating pathology that must be differentiated, because their treatment differs: malignant melanoma, metastatic
carcinoma, hemangioma, and organized subretinal hemorrhage. Although metastatic carcinoma is the most common neoplastic choroidal tumor, malignant
melanoma is most commonly presented to the ophthalmologist as an unknown, and the subsequent discussions of tumor differentiation will be in terms of variation
from melanoma patterns.
We have found it useful to describe ultrasonic features of tumors in terms of their morphologic characteristics, primarily two-dimensional analysis, and in terms of
their acoustic characteristics, primarily one-dimensional amplitude or A-scan analysis. Obviously, the acoustic properties will affect, indeed produce, the patterns
seen in the two-dimensional portrayal, and the diagnostic A- and B-scans are performed together. The discussion of morphology and tissue characterization is
undertaken here to emphasize the different features.
The progression of ultrasonic evaluation from morphologic characterization, primarily from B-scan diagnosis, to acoustic characterization, primarily from A-scan
diagnosis, proceeds much as histologic evaluations proceed from gross inspection through increasingly higher powers of microscopic examination. Because the
evaluation is in vivo, dynamic tissue characteristics, such as vascularity, provide unique insights not comparable to any histologic technique.
Morphologic Characteristics
Size
Malignant melanoma and metastatic carcinoma occur in a continuum of sizes ranging from minimally elevated masses to lesions almost filling the globe, all of
which can be well demonstrated on the B-scan display. Hemangiomas, in our experience, have not typically shown an elevation higher than 5 to 6 mm; however,
they can have a very wide cross section. Subretinal hemorrhages sent for ultrasonic evaluation have characteristically low elevations, usually less than 4 mm.
20 MHz systems offer axial resolution of as little as 75 microns. However, highly accurate tumor measurements are made difficult because the base of the tumor
may not be clearly distinguishable from the scleral echoes, and subsequent measurements may not obtain the same fiducial points, particularly if different
examiners and equipment are used. The standard method is to measure maximal tumor thickness (or tumor height). With our equipment, using 10 MHz, we
consider measurement to be accurate to 0.1 mm for purposes of chronologic comparison.
Shape
Malignant melanomas occur primarily in two characteristic shapes, polypoid or convex, as seen on B-scan. The most common form of melanoma is an elevated
convex mass (Figures 3.97 and 3.98), which occurs when growth of the lesion is restrained by Bruch's membrane. A
P.90
polypoid, or “collar-button” shape, is assumed when the tumor breaks through Bruch's membrane into the subretinal space and protrudes into the vitreous.
Ultrasonically, these tumors have a mushroom appearance when the scan plane is passing through the tumor stalk (Figures 3.99 and 3.100). However, when the
scan plane does not intersect the stalk, this type of lesion could resemble an isolated intravitreal mass. In such an instance, serial ultrasound sections are essential
to trace the stalk of the tumor emanating from the choroid.
Figure 3.97. 10-MHz B-scan of a melanoma, showing the characteristic convex shape with a small area of
adjacent retinal detachment.
Figure 3.98. 10-MHz B-scan of a malignant melanoma filling a major portion of the eye. Acoustic attenuation is
evident in decreasing echogenicity within the tumor with depth, but this is more clearly demonstrable on A-scan.
Figure 3.99. Large collar-button melanomas on B-scan demonstrating difference in texture between the button
portion and the base of the original lesion next to the choroid. (Shown in 3-D in Figure 3.102.)
All metastatic tumors that we have examined have appeared as convex or placoid masses (Figure 3.101), and none has exhibited a collar-button shape. However,
collar-button metastatic adenocarcinoma has been described (137). In general, metastatic masses have a lower silhouette (i.e., a lower height-to-base ratio) than
do malignant melanomas. Choroidal hemangiomas and organized subretinal hemorrhages are also usually flattened or slightly convex. In relatively flat lesions, all
four tumor types can appear as a simple convex mound and require A-scan and often repeat examinations for diagnosis, unless they are thick enough that tissue
typing can be used (generally greater than 1.5 mm).
3-D Ultrasound and Volume

Although height and base or chord measurements have a long history of acceptability for tumor measurement, volume measurements are clearly superior as a
means of following tumor growth or regression posttreatment (138). Recent instruments make these measurements more feasible and practical. Because volume
relates to a cube function of measurements, accuracy of volume is increased, but the error percentage also increases. We use a 10% change from baseline
volume for a significant growth bench mark (Figure 3.102).
Location
Localization of the tumor may aid in differentiation because hemangiomas tend to occur in the posterior pole, particularly near the optic nerve; metastatic tumors at
the posterior pole; and malignant melanomas throughout the choroid.
Localization of tumors should be done with respect to normal ocular structures (Figure 3.103), because proximity to the optic nerve may have prognostic
significance,
P.91
and measurement of distance from the lens or ciliary body may be useful when considering placement of a radioactive plaque.
Figure 3.100. Immersion A- and B-scans of collar-button melanoma. Note high amplitude echoes in the
“button” (A) versus lower amplitudes in base (B).
Figure 3.101. 10-MHz A- and B-scans (left) and 20-MHz A- and B-scans (right) of a metastatic carcinoma
demonstrating a relatively placoid shape and moderate reflectivity throughout entire tumor.
Evidence of Spread
It is possible to preoperatively identify subclinical extension of a melanoma beyond the globe with ultrasound. In the case shown in Figure 3.104 (see also DVD),
ultrasonography was consistent with extension of the tumor into the optic nerve and was confirmed pathologically. In general, the diagnosis of subtle extrascleral
extension cannot be made with high confidence and must be confirmed with additional imaging and clinical data. Massive orbital extension, however, can be easily
shown, as demonstrated by Figure 3.105, with a flat intraocular choroidal melanoma.
Size Progression
Repeating ultrasonic evaluation at various time intervals has been valuable to document the progressive growth (Figure 3.106) or regression of a mass lesion.
Variation in scan plane is unavoidable in scans performed at different times, but scans through the area of maximum elevation with maximization of the A-scan
vitreoretinal echo height will reduce measurement error. Ultrasonography has also been useful in following the response of tumors to local radiotherapy (I-125
plaques) (Figure 3.107) as well as proton beam radiation and transpupillary thermotherapy (TTT) (139) (Figure 3.108). Hemangiomas can also be followed after
photocoagulation to document regression when hemorrhage is encountered, or with other causes of opaque media.
Metastatic carcinoma usually grows more rapidly than malignant melanoma, but we seldom have the opportunity to follow ultrasonically a patient with such a lesion
because of other medical priorities.
P.92
Figure 3.102. A typical 3-D portrayal of a melanoma, as shown in cross section in Figure 3.99.
Figure 3.103. Tumor adjacent to the optic nerve, causing a widening of the typical optic nerve shadow. In this
case, a vitreous hemorrhage obscured the tumor. Differentiation of melanocytoma from a perioptic melanoma
may not be possible with conventional 10-MHz ultrasound.
Figure 3.104. A small melanoma is demonstrated at 20 MHz. An area of subretinal fluid is noted anterior to the
tumor, and an area of sonolucence is noted posterior to the tumor, indicating possible orbit extension. (See also
DVD.)
Figure 3.105. An area of orbital sonolucence is seen posterior to a relatively flat ocular melanoma, consistent
with a large orbital extension of the tumor.
P.93
Figure 3.106. Ocular tumor growth of a relatively small ocular melanoma is demonstrated here over a period of
years.
Acoustic Profile
The acoustic profile of an ocular tumor includes the height of the echo from the leading boundary of the tumor (reflection coefficient), the attenuation of the sound
beam as it is transmitted through the tumor (absorption coefficient or decay slope), the presence, spacing and height of reflecting surfaces within the tumor (internal
tissue texture), and the variation of absorption and texture, with changes of frequency of the transducer. These acoustic parameters will be discussed individually
and are interrelated, as are the morphologic characteristics.
Figure 3.107. The effects of radiation treatment on a ciliary body tumor, by use of parameter image staining, is
shown. Tumors may not regress significantly in size but may show a progression of scattering elements. (see
color image)
Boundary Properties
Echo amplitude is directly related to the change in impedance between the different tissue layers traversed by the examining ultrasound beam. High-amplitude
echoes are produced at boundaries where there is great discontinuity in tissue sound velocities or tissue densities, producing an acoustic impedance mismatch.
Examples of this occur at fluid-tissue boundaries or at boundaries between highly disparate tissues, such as lens and vitreous.
The leading edges of choroidal tumors produce high-amplitude echoes when the sound beam is perpendicular to the mass. For A-scan evaluation it is essential
that
P.94
this boundary echo be maximized to obtain proper relative values for the internal tissue echoes, as discussed in a later section, Attenuation Coefficient. The height
of this leading edge is the high-amplitude echo from the vitreoretinal interface (Figure 3.109). The boundary properties of malignant melanomas, metastatic
carcinomas, hemangiomas, and subretinal hemorrhages are thus similar and have not aided materially in differentiation. Identification of this feature, however,
separates tumor masses from intravitreal hemorrhages, which, lacking regular boundaries, reflect low-amplitude echoes from their anterior surfaces.
Figure 3.108. Top: This figure demonstrates the use of ultrasound to follow the shape of a melanoma pre- and
post-proton beam therapy. Bottom: Comparative broadband (left) and narrow band (right) images of the same
post-treatment tumor. Narrowband image demonstrates the presence of a vitreous hemorrhage.
Attenuation Coefficient
The attenuation coefficient is a measure of the rate of energy loss in the ultrasound beam as it passes through tissue. The tissue absorbs ultrasound, and thus the
height or strength of the echoes returned to the transducer diminishes. The amount of ultrasound energy absorbed is dependent on viscoelastic characteristics,
and their absorption makes up over 90% of the loss of power. Ultrasound is also attenuated by reflection and scattering from tissue elements. In a heterogeneous
tissue both absorption and scattering losses combine to create a gradual falloff (decay slope) in echo height following the initial boundary spike. This decay slope is
specific to each type of tissue, and it is approximated for each tissue by a line connecting the peaks of echoes from within a tumor on A-scan. The decay slope of a
heterogeneous tumor that has internal reflecting surfaces produced by blood vessels or variations in tissue type is less easily approximated than the slope of a
tumor with only one type of internal scattering element or tissue.
Ossoinig (10) has termed the decay slope as an angle Kappa, referring to the angle between the echo peaks (i.e., attenuation) and a horizontal line parallel to the
baseline. It should be noted that the decay slope in reality follows an exponential curve and appears as a line only when a logarithmic amplifier is used. Because
the amplifier that Ossoinig uses (the S-shaped curve amplifier) approaches a logarithmic amplifier in character, the decay
P.95
slope will be somewhat linear with this instrument. With linear amplifiers, which we use, the decay slope is exponential (see Chapter 2). This curve is best
discerned with the radiofrequency display, which is one reason we use radiofrequency as well as video A-scan signals. The other major reasons for observing the
radiofrequency signal are (a) the radiofrequency has minimal electronic processing and thus is less susceptible to amplifier overload or reject and is a more realistic
and faithful display of tissue reflections than is the “envelope” or video trace; (b) tissue texture is better visualized with the radiofrequency than with the video
display; and (c) digital image processing is more informative with the unadulterated RF signal.
Figure 3.109. A-scan through a melanoma demonstrating the internal tissue character once the A-scan has
been oriented properly so that a maximal boundary echo is obtained. The figure shows the characteristic
attenuation of a melanoma compared to subretinal hemorrhage.
Nevertheless, as Ossoinig et al. (127) and Coleman (36) have emphasized, the decay slope is essential in differentiating tumor types. Malignant melanomas show
a high-amplitude leading portion with a steep decay (or “angle Kappa”), often reaching baseline as the tumor adjoins the sclera (Figures 3.110 and 3.111; see also
DVD). Hemangiomas generally exhibit a relatively mild, uniform decay slope, lacking the final low-amplitude section seen with malignant melanoma (Figures 3.112
and 3.113). The average amplitude of the hemangioma is usually about 70% of the scleral echo height with our system but is normally about 95% to 100% of
scleral echo height, with the system used by Ossoinig et al. (127).
Metastatic carcinomas, like hemangiomas, have relatively flat decay slopes but have lower internal amplitudes, usually about 50% of the scleral height (Figure
3.114; see also DVD). Subretinal hemorrhages show low-amplitude internal reflections, which are only 10% to 20% of the scleral height (Figure 3.115; see also
DVD). The sclera is used as a reference for the height of choroidal tumor reflectance because it is posterior to the tumor and has already been subjected to tumor
absorption. Thus, attenuation within the tumor by structures anterior to the tumor is subtracted from the scleral echo, and relative absorption differences between
tumor and sclera are thereby maintained.
P.96
Figure 3.110. Left: B-scan of small dome-shaped malignant melanoma with overlying serous detachment.
Right: High-resolution A-scan reveals typical low-amplitude internal echo pattern.
Figure 3.111. Melanoma on B-scan with the A-scan. Typical attenuation, taken on a vector through the tumor.
(See also DVD.)
Figure 3.112. Left: B-scan of choroidal hemangioma with typical dome shape and high internal echo levels.
Right: High-resolution A-scan sustained high-amplitude echoes comparable in amplitude to orbital fat. The
horizontal arrow indicates tumor position.
P.97
Figure 3.113. Hemangioma on B-scan with the A-scan at 10 MHz (left), 20 MHz (middle), and 22 MHz (right).
Typical attenuation, taken on a vector through the hemangioma.
Internal Tissue Texture

Echoes occurring within a tumor following the boundary spike delineate tissue “texture” by their spacing and height. Internal echoes are the result of
inhomogeneities in the tissue, such as blood vessels or poolings of fluid. These relatively small interfaces, known as scatterers, generally, do not contribute
significantly to absorption, but their presence and position are useful in differentiation between tumors.
Melanomas often have high-amplitude discontinuities, usually large blood vessels, that produce echoes rising above the decay slope (Figure 3.116; see kinetic
scan on DVD). These echoes may show time variations in amplitude and position within the tumor. Ossoinig (11) has described these variations as “spontaneous
movements.” The rest of the radiofrequency echo complex in malignant melanoma appears as clustered, relatively coarse, widely spaced echoes mixed with
closely spaced echoes. Subretinal hemorrhage, generally, appears as fine-textured, closely spaced echoes, and metastatic carcinoma appears as coarse-textured
echoes.
Figure 3.114. Left: B-scan of metastatic carcinoma (lung primary) shows placoid shape with overlying retinal
detachment. Note accentuation of Tenon's space and orbital involvement. Right: High resolution A-scan of
marked vector segment shows that internal echo pattern is moderate in amplitude with negligible attenuation.
(See also DVD.)
Frequency Variation
In discussing the previously mentioned internal tissue characteristics, we have assumed that only a single transducer with a given frequency is used. When the
ultrasound frequency is changed (by using a different transducer), each of the internal tissue texture properties (i.e., echo amplitude, echo spacing, and acoustic
absorption) may vary. We have found this variation in tumor
P.98
“acoustic profile” between examining frequencies of 10 and 20 MHz (Figure 3.117) to be valuable in distinguishing melanomas from metastatic carcinomas,
hemangiomas, and organized subretinal hemorrhages. Melanomas exhibit a sharper drop to baseline on A-scan (increased attenuation on B-scan) with higher
frequencies. Metastatic carcinomas are usually solid at all frequencies (i.e., maintain internal echoes on A-scan). Hemangiomas vary according to size but are
usually solid at all frequencies, and organized subretinal hemorrhages are usually anechoic at all frequencies. Even the different cytologic types of choroidal
malignant melanoma may show different frequency-related variations, with mixed cell, or epithelioid tumors often showing hypoechogenicity with increased
frequency, unlike spindle cell tumors (140). These frequency differences are part of the reason that techniques for obtaining the frequency spectrum of tissue
echoes offer even greater tissue identification potential. Knowledge of frequency-tissue relationships as a means of augmenting the acoustic profiles of tumors
should develop further as instrument and techniques are improved and experience is gained.
Figure 3.115. A subretinal hemorrhage with the vector A-scan demonstrating low amplitude echoes from the
blood. (See also DVD.)
Figure 3.116. Acoustic discrimination within a melanoma caused by inhomogeneities, such as septae or blood
vessels. (See also DVD.)
Figure 3.117. B-scan ultrasonogram of a small melanoma at 10 and 20 MHz. Higher resolution at 20 MHz
allows thickness to be more accurately measured and gives a better measurement of the posterior ocular coats.
Even better resolution can be obtained with radiofrequency reconstruction, as seen in Figure 3.125.
P.99
Analytic Mathematical Modeling and Parameter Imaging
Tissue Characterization
Not only can gray scale be used to give texture to amplitude data on a B-scan, but analytic mathematical modeling can be used to analyze different frequencies of
reflected echoes from a power spectrum, that is, frequencies calibrated against a perfect reflector for a range of tissue frequency responses, using a single
frequency transducer. Analysis of the power spectrum is discussed in Chapter 2 for technical background.
Clinically, the resultant pixel representation of scatterer diameter or scatterer concentration relies on the B-scan. These parameter images of a tumor or tissue are
digitally reconstructed from radiofrequency data acquired from the tumor or tissue. With this technique, we can compare different areas of a tumor or a tissue to
determine probability of microarchitectural variations in the tumor that relate to such things as tumor lethality. The two parameters found most useful for these
variations are scatterer concentration and scatterer density (Figure 3.118). A scan through a melanoma demonstrating scatterer size and concentration with
pseudocolor is shown in Figure 3.119.
The areas that have been particularly useful with parameter image or tissue characterization are tumor identification with subclassification or stratification and
tissue identification (141).
We studied 117 patients with ocular melanoma with the cooperation of the University of Iowa (Boldt and Weingeist) and the University of Illinois-Chicago (Folberg,
Chen, and Vangveeravong). Patients were seen in Iowa prior to enucleation and the eyes were scanned and RF digitized data were collected. The enucleated eyes
were evaluated by Dr. Folberg and his group at the University of Illinois-Chicago for the histologic presence of high risk extravascular matrix patterns. The
ultrasound was analyzed independently at the Weill Cornell Ultrasound Lab by Dr. Silverman and Rondeau, using techniques previously described by Lizzi and
Coleman (Figure 3.120). The results showed that this noninvasive technique can identify high-risk melanomas with 80.1% cross-validated correct classification
(56).
Figure 3.118. Graph shows acoustic scatterer sizes as measured from tissues at different frequencies. By
comparing the frequency response to the interrogated values of the tissue, we can determine the scatterer size
and, in a similar way, concentration and density (see Chapter 2). (see color image)
Figure 3.119. A melanoma with gray scale on the left, and after processing the power spectrum for scatterer
diameter and concentration on the middle and right images. (see color image)
The value of this technique is being explored as a means of stratifying patients for treatment staging and monitoring of therapeutic modalities.
Tissue characterization is also of value in identifying tissue subgroups, such as the choroid and ciliary muscle. The identification of ciliary muscle is useful in
studying physiologic effects of drugs and the effects of treatments, such as radiation on tumors.
Small Melanomas and Nevi

An area of concern and controversy is the diagnosis and possible treatment of very small melanomas. Although nevi and congenital hypertrophy of the retinal
pigment epithelium (CHRPE) lesions are universally regarded as benign, the distinction between nevi and small melanomas can be subtle clinically. The distinction
between small, dormant melanomas and potentially high-risk melanomas can be both subtle and daunting.
Conventional 10-MHz ultrasound has not been particularly helpful in this distinction, because both tend to show hyperreflectivity but too little thickness to allow
conventional A-scan distinction. The use of higher frequencies, especially when complemented by parameter image analytic modeling, offers a better way of
distinguishing these three categories. Many authors
P.100
have discussed the clinical differences between dormant and high-risk melanomas, but Shields et al. (143, 144, 145, 146) have written most extensively on this
subject. Clinical signs of high-risk melanomas are orange (lipofuscin) pigment and subretinal fluid. These can be quantitatively augmented with high frequency
ultrasound. Figure 3.121 shows the clinical and ultrasonic features that can be used to differentiate these three classes. The ultrasonic distinction rests on three
quantitative features. These are choroidal replacement, subtumor or intratumor “fluid,” and precise measurement of growth.
Figure 3.120. Demonstrates the histology of a melanoma (lower left) compared to the ultrasonogram (upper
left). The histology demonstrates the extravascular patterns (EVM) of the melanoma. The two right scans show
the scatterers that correlate with these EVMs. The lower right shows the visual correlation of scatterers to EVMs
indicating a “high-risk” melanoma. (see color image)
The histologic difference between a nevus and a tumor is shown in Figure 3.122. Melanin is highly reflective and thus easily seen with ultrasound. No fluid exists,
and the choroid is normal. A dormant melanoma may be homogeneous in acoustic profile, has no fluid, and, generally, does not seem to replace the choroid. The
“high-risk” melanomas in our experience are those that show
1. Lipofuscin pigment, subretinal fluid, and drusen, as emphasized by Shields and Shields (147, 148)
2. choroidal replacement
3. fluid either subretinal or intratumor
4. growth of at least 0.1 mm in thickness over a 3- to 6-month period.
Figures 3.123, 3.124 and 3.125 demonstrate these features, and Figure 3.126 is the schema demonstrating our clinical management.
The mathematical analysis used to distinguish the choroidal layer and the placement is the same as described previously in the section on age-related macular
degeneration.
P.101
Figure 3.121. Clinical and ultrasound signs that have been shown helpful in differentiating nevi from dormant
and high-risk melanomas.
Figure 3.122. Histologic preparations demonstrating a nevus, which has high concentration of melanin and
relatively uniform vascular architecture, compared with a small melanoma showing uniform or homogeneous
tumor tissue, oft-associated subretinal fluid, and the concentrated vascular pattern (box). (Courtesy of Robert
Folberg, MD.) (see color image)
Figure 3.123. Patient with both a nevus, adjacent to the nerve, and a suspicious melanoma in the midperiphery.
The B-scan ultrasonogram is taken at 20 MHz, with parameter image tissue staining to outline the choroid; in
this case, presence is noted posterior both to the nevus and the melanoma. (see color image)
P.102
Figure 3.124. This patient also had a suspicious melanoma, although in this case, the choroid is replaced.
Unlike a high-risk small melanoma, there is no evidence of fluid or growth. (see color image)
Doppler and Kinetic Properties of Tumors

Ossoinig (12) has emphasized the value of using the A-scan probe as a means of “ballotting” tumors to stimulate changes in their acoustic patterns, as well as to
permit observation of their compressibility. This test is more useful in the orbit than in the globe, especially when dealing with cystic tumors. The detection of
vascular echoes can be enhanced with this technique during ocular examination.
Color flow Doppler can show vascularity in large tumors, not only in the tumor but in its underlying choroid and orbit (149).
Acoustic Characteristics (B-scan)

The acoustic profile of a tumor on A-scan translates into a B-scan display as variations in the appearance of tissue texture. The phenomena of acoustic quiet
zones, choroidal excavation, and acoustic shadowing are major sources of B-scan tumor differentiation.
Figure 3.125. This patient has what we regard as a high-risk melanoma, as a result of the presence of clinical
and ultrasound signs. These include: subretinal fluid, choroidal replacement. This class of lesion is followed at
short intervals for growth. (see color image)
Acoustic “Quiet Zone”

Malignant melanomas appear on B-scan as hyperechoic areas protruding into the anechoic vitreous cavity. Histologically, malignant melanomas are
homogeneously cellular with varying degrees of vascularity. With increasing vascularity of the tumor tissue, there are many internal acoustic interfaces, so that
more echoes are returned and the tumor thus appears hyperechoic. This hyperreflectivity in the more vascular uveal melanoma is apparent at transducer
frequencies of 5, 10, 15, and 20 MHz. Polypoid-shaped melanomas almost always exhibit these characteristics of acoustic solidity in the “button,” whereas the base
or “collar” remains relatively anechoic. In relatively avascular melanomas (most often those with convex shape), the homogeneous cellularity of the tumor and lack
of significant internal acoustic interfaces result in the appearance of an acoustic “quiet zone” or hypoechoic region within the tumor (Figure 3.127). This
phenomenon is accentuated on the B-scan, although the A-scan tracing registers echoes of moderate though declining amplitude
P.103
throughout the tumor. If the tumor were actually physically fluid filled, the A-scan would show absence of echoes after the initial leading echo (as seen in a retinal
detachment). This phenomenon of acoustic quiet zone, or hypoechoic region in relatively avascular melanomas, is most prominent at examining frequencies of 15
and 20 MHz. At 5 and 10 MHz, the tumors almost always appear echoic.
Figure 3.126. Schematic summarizing the clinical and ultrasonographic findings as they dictate separation of
suspicious and high-risk small melanomas, and the clinical management.
Choroidal Excavation
Involvement or replacement of the choroid by a melanoma can be shown dramatically by the “excavation” phenomenon. The area of tumor that has replaced the
surrounding choroid demonstrates a dish- or bowlshaped indentation into the smooth concave choroidal outline. It must be remembered that the choroid in the
living eye is a highly vascular erectile tissue that may be as thick as 500 mm or more at the posterior pole (Figure 3.128). Excavation has been frequently noted in
malignant melanoma, although not all melanomas exhibit this feature.
Figure 3.127. B-scan at 10 MHz of a homogeneous melanoma, showing very low amplitude or absent echoes in
the central part of the tumor.
Figure 3.128. A small melanoma seen at 20 MHz with the Quantel ultrasound apparatus, demonstrating the
typical choroidal replacement on B-scan. As noted earlier, this feature can be accentuated by use of tissue
staining. But in general, one sees a scaphoid indentation in the wall posterior to the tumor, of an accentuated
curvature relative to the curvature of the normal choroid.
P.104
The incidence of choroidal excavation in a series of 110 intraocular tumors was evaluated (130). Choroidal excavation was absent in all cases of metastatic
carcinoma and hemangioma and was noted only in malignant melanoma. Of the 89 malignant melanomas, 42% exhibited this characteristic and 58% did not.
Choroidal excavation was not seen in any melanoma anterior to the equator. Fuller et al. (150) have, however, reported that excavation was seen in metastatic
carcinoma in their series. The histologic similarity of metastatic carcinoma to certain melanomas, however, should perhaps lead us to expect a similar pattern.
Subretinal hemorrhage and disciform macular degeneration, in our experience, do not show choroidal excavation.
With high-frequency scans, the choroid can be measured even in very small nevi/melanomas by use of “midband fit” scans (as described previously) that can
differentiate sclera and choroid. Nevi, almost invariably, sit on residual choroid, whereas small melanomas displace the choroid.
Attenuation Defect or “Shadowing”

Attenuation of sound by one tissue mass can cause an acoustic attenuation defect or shadowing to appear in structures behind the mass. A solid mass will
sometimes attenuate sound to such an extent that the area of retrobulbar fat behind it will seem fainter than the rest of the orbit, or the sound beam will not
penetrate far into the orbit directly behind the tumor, causing a hypoechoic appearance (Figure 3.129). This absorption defect will not occur if the mass has good
sound transmission properties. Apparently, no significant variation occurs in the shadowing produced by melanoma and metastatic tumors with our techniques.
Hemangiomas show little evidence of shadowing, probably as a result of their lower attenuation coefficient.
Figure 3.129. A treated retinoblastoma, with considerable calcific change, produces a shadowing posterior to
the lesion.
Associated Ocular Changes

Pathology associated with intraocular tumors is also ultrasonically demonstrable. Asteroid hyalosis may mask a tumor in ophthalmoscopic exam. Vitreous
hemorrhages occur relatively infrequently with melanomas of the choroid but can be demonstrated ultrasonically (Figure 3.130). Retinal detachments secondary to
intraocular melanomas or metastatic carcinoma are clinically important (Figure 3.131). Melanomas will often have a fluid layer between the retina and the anterior
tumor surface. They appear on B-scan ultrasonograms as bullous retinal elevations with a sharp, high-amplitude leading edge. Serous retinal detachments can
also be associated with choroidal hemangioma.
Conditions Simulating Choroidal Tumors
Retinal Lesions
In retinal detachment, or retinoschisis, the ultrasonic pattern may show an elevated vitreoretinal interface echo, but because the subretinal space is anechoic, the
elevation is readily distinguishable from a tumor. Disciform macular degeneration also shows an elevated vitreoretinal interface. These hemorrhagic lesions show
low-amplitude internal echoes on A-scan but will appear hollow at 15 and 20 MHz (or with reduced gain at 10 MHz).
Figure 3.130. An ocular melanoma with surrounding hemorrhage that obscured the clinical view. This patient
had been treated with radiation, and the tumor regression could thus be followed with ultrasound only.
P.105
Figure 3.131. This ocular melanoma was obscured and in question, as a result of an overlying retinal
detachment. The tumor is easily seen with conventional ultrasonography.
In chorioretinitis, an area of elevated retina may be seen, but the subretinal space is acoustically clear. Retinal pigment epithelium (RPE) lesions (such as CHRPE)
that appear flat and highly pigmented ophthalmoscopically do not have sufficient elevation to allow ultrasonic detection, but the high reflectivity of melanin may
make them highly visible, both optically and acoustically.
Choroidal Lesions
Most benign choroidal nevi fail to show significant elevation and thus cannot be demonstrated on routine B-scan ultrasonograms, though as noted previously, high
frequency imaging of the choroid can be performed. Choroidal detachments present a typical convex circumferential elevation straddling the ora serrata with a
sonolucent area between the retina and sclera on B-scan. Organized choroidal hemorrhage may be difficult to distinguish acoustically from a tumor, but the internal
echoes are of lower amplitude. Lymphoid hyperplasia or lymphoma of the choroid (Figure 3.73) may be ultrasonically indistinguishable from an “en plaque”
melanoma but may be suspected because of greater sound absorption by inflammatory tissue. Lymphoma is difficult to diagnose because of its rarity and it
resembles inflammatory or metastatic disease. Vitreous debris on scanning may offer a clue to this diagnosis, particularly if the history indicates suspicion of such
an entity.
Vitreous Lesions
Vitreous hemorrhages that have undergone organization may appear as echoic masses. The internal echo amplitudes are usually lower than those seen in
melanoma. A posterior vitreoretinal interface appearing smoothly curved and in normal position may help in differentiating these hemorrhages. Repeated ultrasonic
evaluations may be necessary to distinguish a tumor that lies within a dense, vitreous hemorrhage, a situation that arises more often with retinoblastoma than with
choroidal tumors. The difficult diagnostic problem of organized hemorrhage in conjunction with retinal detachment occurs rarely.
Reliability and Limitations of Ultrasonic Differentiation

The reliability of ultrasonic diagnosis of ocular tumors in our laboratory has been reported as better than 96% for differentiation of neoplastic choroidal tumors from
benign subretinal hemorrhages, vitreous hemorrhages, and retinal detachments (151). Ossoinig (11) has reported a similar figure, using his techniques. To date we
have examined nearly 10,000 patients with ocular tumors, both neoplastic and benign. It has not always been possible to identify the tissue present on one
examination. Serial examinations are often requested to permit growth documentation as well as to repeat the evaluation. The methods described here, even when
absolute differentiation cannot be made, can direct the course of treatment, with small solid tumors being followed and patients with larger tumors referred for
metastatic workup.
In addition to the problems in identifying discrete lesions as discussed earlier, other difficulties in ultrasonic diagnosis of choroidal tumors exist that are related to
size or position.
First, very small lesions cannot be demonstrated ultrasonically. In general, lesions causing more than 1 mm of elevation of the retina can be detected and
demonstrated. When the tumor can be visualized and the ultrasonogram is performed under optimal conditions, tumors with only 0.5 mm of elevation can be
depicted. Smaller lesions can certainly be missed with ultrasonography. This is a problem of equivocal significance because there is a body of opinion that eyes
with small lesions should not be treated immediately, but rather be followed for demonstrated growth before being treated. Clinically, it is customary to follow a
small lesion to document a growth change.
Second, large lesions filling the vitreous may be confusing in that they may resemble vitreous hemorrhages. With massive necrotic melanomas, this is a particular
problem, and it also occurs with medulloepitheliomas, where cystic changes may be seen.
Finally, difficulties persist with optimal B-scan visualization of the ora serrata and pars plana regions with contact 10-MHz scans. Even with the immersion
technique, structures that lie perpendicular to the examining beam are well portrayed, but structures lying parallel to the beam, such as the ocular walls at the ora,
are not well outlined. Also, structures preceding a tumor will
P.106
tend to mask some of its acoustic profile. Very high frequency scans have made this area much easier to identify and should be the procedure of choice with areas
described later.
Newer tissue enhancement techniques have been found useful in enhancing the use of A- and B-scan techniques (137).
Ciliary Body Abnormalities

Ciliary body tumors occur less frequently than tumors of the choroid, representing approximately 10% of all ocular melanomas (110). Tumors of the ciliary body are
often difficult to diagnose clinically because they arise in an area of the eye not easily examined and are not usually amenable to fluorescein angiography. They
precipitate cataracts or secondary retinal detachments that may cause difficulty in clinical diagnosis. In addition, ophthalmoscopically visible masses may be
difficult to distinguish from cystic lesions of the ciliary body, with elevation of the nonpigmented epithelium of the ciliary body in this area. Thus, ultrasonography is
valuable in the diagnosis of such tumors, and the use of UBM or very high frequency ultrasound (VHFU) has greatly improved the initial diagnosis of ciliary body
tumors (54,152,153).
Attention to certain technical features of the ultrasound examination will improve imaging of ciliary body tumors. First, rotating the eye as much as possible is
important, bringing the mass perpendicular (in either an anterior or posterior position) to the transducer for best resolution. Second, small tumors in this region may
be missed, particularly at the 6:00 and 12:00 meridians with horizontal scans, so B-scans should be made in all meridians. Third, a range of transducer frequencies
should be used to optimize differentiation. Fourth, serial examinations at a later date are necessary in equivocal cases.
Ciliary Body Tumors

The ultrasonographic characteristics of ciliary body tumors will be discussed as choroidal tumors, in terms of both morphologic and acoustic characteristics. Solid
lesions are nearly always melanomas but may be other solid tumors, such as medulloepithelioma, in rare instances.
The location and size of ciliary body tumors can be well demonstrated by high frequency ultrasound (Figure 3.132). Secondary changes, such as retinal
detachment, hemorrhage in the vitreous, and cataractous lens changes, can also be shown.
As in tumors of the choroid, anechoic zones at the posterior part of a ciliary body tumor can be appreciated. The presence of cystic changes in the tumor is typical
of medulloepithelioma, but this pattern can also be seen in melanoma, although very rarely.
Lesions Simulating Ciliary Body Tumors

B-scan ultrasonography provides the differential diagnosis between ciliary body cysts and ciliary body tumors. The interior of the cysts is sonolucent, and the
A-scan trace remains at baseline throughout the cyst. High frequency B-scan ultrasonogram of ciliary body cysts are shown in Figure 3.133, demonstrating the
anechoic cystic structure of the lesion. In some instances, however, debris within a cyst may produce internal echoes.
Figure 3.132. Ciliary body tumors, as well as iris tumors, are best evaluated using 50-MHz scans. This figure
demonstrates the hyperreflective, internally solid mass of the ciliary body, consistent with a ciliary body
melanoma.
Choroidal detachments or effusions can simulate a ring melanoma of the ciliary body, but ultrasonography should differentiate them based on their hypoechoic or
hyperechoic internal structures, respectively.
Figure 3.133. This figure demonstrates a ciliary body/iris cysts, which appear similar clinically to the ciliary body
tumor seen in Figure 3.132. A clear cystic outline indicates a benign ciliary body cyst. Top: Ciliary body/iris cyst
with internal debris. Bottom: “classic” ciliary body cyst with no internal echoes.
P.107
Metastatic tumors are most unusual in a position anterior to the ora serrata, and the diagnosis of a metastatic tumor can, if so evaluated, usually be made clinically
on the basis of multiple ocular tumors and/or a known primary tumor.
Abnormalities of the Optic Nerve Head

Papilledema, papillitis, pseudopapilledema, drusen, and melanocytoma may be visualized ultrasonically as a protrusion of the intrascleral portion of the optic nerve
into the vitreous, with increased reflectivity from this region. Papilledema and papillitis usually cannot be distinguished by the ultrasonic appearance of the scleral
portion of the nerve alone, although subretinal fluid may, occasionally, be seen with papilledema (Figure 3.134). The appearance of the orbital portion of the nerve
may indicate the correct diagnosis ophthalmoscopically, in cases of both visible and nonvisible elevations of the optic nerve head. In cases of pseudopapilledema
as a result of drusen (Figure 3.135), the orbital echoes are normal, corresponding to the histologic findings that drusen in the nerve are not found posterior to the
lamina cribrosa. Drusen may be so large that shadowing or internal echoes can cause apparent enlargement or internal reflections in the anterior nerve segment.
In cases of optic neuritis, echoes from the nerve wall may produce a contiguous line, appearing to separate the nerve and the sheath. When optic neuritis is seen
in association with enlarged, inflamed rectus muscles, this “doubling of the wall sign” may indicate the diagnosis of Graves disease. When optic neuritis is
associated with pseudotumor of the orbit, the inflammatory character of these mass lesions can be suspected by the nerve changes as well as edema of normal
structures, such as Tenon's capsule.
Figure 3.134. 10-MHz ultrasonogram showing prominence of the optic nerve head. This usually indicates
papilledema, but differentiation from pseudopapilledema may be difficult. The presence of a fluid area may help
indicate true papilledema.
Figure 3.135. Drusen of the optic nerve head are very highly reflective and can help in the differentiation of
papilledema and pseudopapilledema. Conventional (left) and midband fit (right) are shown here in “stretched”
format.
We believe that shrinkage or atrophy of the optic nerve cannot be reliably appreciated with a B-scan 10-MHz ultrasound. Standardized echographers state that they
can measure the optic nerve diameter very accurately with A-scan techniques (154). However, given the orientation of the nerve to the interrogating beam, we
question the accuracy of this technique. When the optic nerve is invaded by tissue which is acoustically dissimilar to that of the normal optic nerve tissue (normally
anechoic as a result of alignment of nerve fibers), acoustic interfaces occur, and abnormal echoes are returned from within the nerve. As an example, when a
juxtapapillary melanocytoma extends into the nerve, many abnormal intraneural echoes can be identified. Melanocytomas are composed of polyhedral cells with
large amounts of pigment.
Optic nerve cupping is not usually demonstrable ultrasonically in its early stages because the beamwidth used in ocular evaluation is too wide to permit resolution
of a small depression with standard 10-MHz ultrasound. The standard beamwidth causes the nasal and temporal cup edges to merge, and the beamwidth artifact
causes merging of echoes from the floor of the optic cup, with echoes from adjoining tissue. Studies using precisely focused, very narrow beamwidth transducers,
can demonstrate cupping of the optic nerve head; higher frequency (20 MHz) focused transducers do allow such imaging (Figure 3.136).
Elevation of the optic nerve head may also be seen with intrinsic tumors of the optic nerve, such as melanocytoma, which, ultrasonically, may resemble drusen.
P.108
Figure 3.136. 22-MHz ultrasonogram of the posterior pole of a normal eye demonstrating resolution of the optic
nerve, the superior orbital vein, and the boundary between Tenon's and sclera, as well as the choroidal
measurements. Note the increasing thickness of the sclera adjacent to the optic nerve. Top image is in axially
“stretched” format.
OCULAR TRAUMA
Synopsis
Ultrasound scans can be made with sterile normal saline bath or a sterile sleeve over the transducer, even with lacerating injuries.
Rupture of the globe can be suspected by irregular wall outline.
A CT scan or plain film x-rays are recommended prior to ultrasound to optimize ultrasound examination. Foreign bodies can be better localized with ultrasound,
relative to ocular structures, but are detected easier with CT or plain film x-rays. Consideration of radiation exposure should be discussed when children are injured.
TABLE 3.5 Improvements in Management Aided by Ultrasound

Conditions Treatment
Vitreous Hemorrhage Vitrectomy
Retinal Perforation Prophylactic Cryopexy
Retinal Detachment Encircling Band
Lens Dislocation Observation or Removal
Lens Rupture Aspiration
Choroidal Hemorrhage Drainage
Scleral Rupture Repair
Foreign Body Localization Extraction
Foreign Body Magnetic Properties Extraction
The preceding sections of ocular diagnosis have been presented in a manner based on the anatomic divisons of the eye. Changes produced in these structures by
trauma have been alluded to; however, the importance of ultrasound in the evaluation of a traumatized globe merits a separate discussion so that specific changes,
examining techniques, and approaches to clinical management may be more adequately summarized.
Ocular trauma may be classified into three broad categories: contusion or concussion injuries, penetrating or lacerating wounds, and foreign body injuries. Eyes
subjected to any of these forms of trauma often exhibit cloudy media as a result of corneal or lens damage, hyphema, or vitreous hemorrhage. In these situations,
ultrasonography becomes essential for complete evaluation of the globe prior to primary repair and for evaluation prior to secondary repair, if required. Effective
medical and surgical therapy of the traumatized eye is thereby enhanced with ultrasonic examination. Table 3.5 summarizes some of the conditions subject to
evaluation by ultrasound and the modes of treatment that may be initiated or expedited.
Careful visual inspection and radiographic and/or CT scan examination for ocular foreign bodies should be performed first. Identification of an ocular foreign body
by x-ray or CT enables the examiner to more rapidly localize the foreign body in relation to ocular structures, shortening the total time required for the examination
and decreasing the chance of missing a small foreign body. MR imaging should be avoided because of the possibility of a magnetic foreign body (155,156). CT
exposure should be minimized in children.
In patients with recent ocular trauma, every attempt should be made to maintain sterile technique. We do not sterilize our transducer, but it is cleaned with alcohol
or immersed in an approved antibacterial wash prior to scanning. Alternatively, a sterile latex sheath can be placed over the end of the transducer. An antibiotic
solution can be instilled in the sterile normal saline bath, if immersion is used. With a severely traumatized globe, clinical judgment would determine whether
immersion B-scan or contact A- or B-scan is indicated. We generally
P.109
use a latex sheath cover (Figure 3.137) with saline between the transducer and the sheath to provide a cushion so that no significant pressure is placed on the eye
and a benefit of a standoff is obtained (Chapter 2) with contact B-scanning (Figure 3.138).
Figure 3.137. Use of a sterile sheath with normal saline can provide an ideal standoff for examining infants or
severely traumatized eyes. This can be used with A- or contact B-scan equipment and is particularly useful with
inexperienced examiners who have a tendency to push the probe harder to get a better image.
Contusion and Concussion Injuries
Hyphema
Hyphema may be noted ultrasonically. A moderately dense hyphema can be appreciated ultrasonically as echoes occurring within the anterior chamber, whereas a
relatively light, nonclotted hyphema may be anechoic, as discussed previously. Extension of the hyphema into the posterior chamber may be discerned. Figure
3.139 demonstrates hemorrhage into the anterior chamber as well as deepening, probable angle recession and hemorrhage into the posterior chamber.
Figure 3.138. 20-MHz B-scan of a normal eye using a latex sheath standoff.
Figure 3.139. 50-MHz B-scan of an aphakic eye following cataract extraction and a moderate hyphema. The
incision site is easily seen, and the hyphema is well outlined.
Angle Recession
High frequency ultrasound is able to depict anterior chamber geometry and can demonstrate an abnormal deepening of the anterior chamber and widening of the
angle as seen in recession. These findings, as seen in Figure 3.140, indicate the severity of the trauma to the anterior segment.
Dislocated Lens
In any form of severe concussion, the lens may become subluxated or totally dislocated from its usual lens position. Even minor variations in position may be
portrayed with the B-scan ultrasound display. Figure 3.141 shows an anterior displacement of the intraocular lens, and Figure 3.142 shows dislocation of the
intraocular
P.110
lens still in the capsular bag. In contusion injuries, the crystalline lens is usually not ruptured and maintains its normal configuration, but presence of capsular
rupture or cataract (Figure 3.143) can be seen and can indicate the need for lensectomy/vitrectomy.
Figure 3.140. 50-MHz B-scan following trauma showing recession of the angle.
Lens Injuries
The interior of the lens is normally anechoic. Intralenticular echoes can therefore indicate early cataract formation secondary to perforation. Identification of a
rupture of the posterior lens capsule with dispersion of lens material into the anterior vitreous can indicate the need for early lens extraction with anterior vitrectomy,
as noted earlier.
Vitreous Hemorrhage
Vitreous hemorrhage occurs frequently following ocular trauma and has already been discussed extensively. To reiterate, light diffuse vitreous hemorrhage is
usually anechoic, although coagulation and clotting accompanying more massive hemorrhage will appear as reflective aggregates within the posterior chamber. As
mentioned, the density, location, and extent of hemorrhage can be well demarcated with the B-scan display. In younger patients with formed vitreous, this
demarcation may orient the examiner to sites of stress and thus possible retinal tears.
Figure 3.141. Intraocular lenses are the most common form of foreign body examined at high frequency. This
50-MHz B-scan shows a lens displaced anteriorly, with the trail of echoes indicating the position of a folded
haptic.
Figure 3.142. An intraocular lens is outlined with its position dislocated temporally but still in the capsular bag.
(See also Figure 3.157 and DVD.)
Penetrating injuries into the vitreous nearly always produce vitreous changes visible acoustically. In young individuals, who suffer the majority of traumatized eyes,
the solid vitreous permits a track of hemorrhage to be traced through the entire vitreous compartment. If this path leads to the posterior globe wall, perforation must
be expected. Surgery can then be directed to the correct quadrant, minimizing unnecessary exploration and reducing the possibility of extrusion of ocular contents
through an unidentified posterior laceration. A posterior perforation
P.111
site can be localized with reference to its distance from the limbus as well as the correct meridian, so that during surgical exploration the wound can be examined
(often with vitrectomy rather than extended exploration) and appropriate therapy, such as endolase, silicone oil, cryopexy, and scleral buckling, can be instituted.
Figure 3.143. Example of a traumatized globe with shallow chamber and cataractous changes in the crystalline
lens.
Retinal Detachment
We have previously dealt with the ultrasonographic appearance of retinal detachment. To summarize, the vitreoretinal interface produces a high-amplitude echo
from the retinal surface, usually allowing identification of this surface in retinal detachment, distinguishing it from hemorrhage along the posterior vitreous
membrane caused by trauma. On rapid motion of the eye in kinetic scanning, a recent retinal detachment will usually move freely but maintain its points of
attachment at the ora and the disc. In most situations, lower-amplitude echoes characterize a vitreous veil versus a retinal detachment, but this feature becomes
less reliable in long-standing vitreal membranes, where the echo heights may approximate those from the vitreoretinal interface, and absolute differentiation may
no longer be possible.
Choroidal Rupture, Scleral Injury

Ultrasound can only rarely detect a scleral rupture, but by demonstrating the presence of a hemorrhage in the vitreous and the areas of contiguity between
hemorrhage and sclera, choroidal rupture or scleral injury may be deduced (Figure 3.144) (157,158). If rupture at the equator is suspected, the globe should be fully
rotated to permit perpendicular examination of this region.
Perforating or Lacerating Wounds
Anterior Segment
As in blunt trauma, hyphema or complete absence of anterior chamber often follows a perforating injury and can be delineated acoustically.
Figure 3.144. 10-MHz B-scan shows a ruptured globe with a break in sclera posterior to an area of subretinal
hemorrhage. Often it is the irregular outline of the globe that may be the only clue to a posterior rupture. The
actual scleral separation is only rarely, if ever, seen.
Scleral Rupture or Penetration

As noted previously, the presence of scleral rupture may not be ultrasonically noted, but distortion of the globe contour, as seen in Figure 3.145, or a path of blood
through the vitreous, can be used to identify the site of perforation.
Choroidal hemorrhage and anterior dislocation of the vitreous, secondary to surgical intervention, may also necessitate ultrasonic evaluation.
Figure 3.145. 10-MHz B-scan of ruptured globe shows distortion of the globe, detached retina, and hemorrhage
debris.
P.112
Foreign Bodies
A major use of ultrasound in ophthalmology has been the localization of intraocular foreign bodies and the determination of their physical properties. Bronson (159)
published extensively on the use of ultrasound for the localization of intraocular foreign bodies and described an intraocular forceps directed by ultrasound. The use
of ultrasound to identify and localize foreign bodies permits early and directed surgery on patients with an intraocular foreign body, with improved visual results
(160).
Radiopaque Foreign Bodies

As noted previously, in evaluating radiopaque foreign bodies, an available CT or x-ray report is useful to the ultrasonographer prior to performing the scan. The
number and position of foreign bodies as determined by radiography can be of inestimable help in directing and shortening the ultrasonic evaluation.
Because of the random size and material of foreign bodies, absolute criteria for their identification cannot be supplied. Their distance from the transducer,
orientation, and acoustic impedance variation from surrounding tissue all affect the reflected echoes. A rigorous, meticulous search for foreign bodies is thus
indicated in all cases where they are suspected. Careful B-scan serial sectioning of the globe and increased attention to the A-scan echo amplitudes are essential.
The localization of a foreign body and determination of its magnetic properties may therefore be more time-consuming than routine ocular diagnosis.
Several acoustic features of metallic foreign bodies are demonstrated by Figure 3.146. In this series of ultrasonograms, a metallic foreign body is seen on the
retinal surface at the posterior pole. By performing serial scans of the foreign body at a series of decreasing gain settings, the foreign body can be more readily
distinguished from surrounding hemorrhage or other tissues. This technique of repeating the serial sectioning at varying sensitivity settings is useful in foreign body
localization because it permits distinction of the high-amplitude echoes produced by the foreign body from lower-amplitude echoes produced by the surrounding
hemorrhage. This particular ultrasonogram demonstrates three other important acoustic characteristics of metallic foreign bodies:
1. The foreign body tends to reflect sound energy, so that the region posterior to it will appear shadowed, or anechoic. In this scan, the retrobulbar fat has a
wedge-shaped shadow resembling an optic nerve shadow in the region directly posterior to the foreign body. This shadow effect is a useful “pointer” to a
foreign body.
2. Sound travels faster through metal than through surrounding vitreous. Consequently, the region posterior to the foreign body shows a slight protrusion or
prominence of the retina, an artifactual result of the increased transmission velocity of sound through metal. This mound posterior to a foreign body on the
retina can be aligned with a shadowed area posteriorly to direct attention to a foreign body (Figure 3.147).
3. Reduplication echoes can act as another “pointer” to the foreign body and are a characteristic of BBs and gas bubbles.
Figure 3.146. Top: B-scan ultrasonogram demonstrating a metallic fragment at the back of the eye surrounded
by hemorrhage. Middle: Reduced gain (a lower sensitivity setting) shows the metallic fragment to lie anterior to
the retina. Acoustic absorption from the fragment produces shadowing in the orbit, a feature useful in localizing
the foreign body. Bottom: Lower sensitivity on both the B- and A-scans demonstrates the higher reflectivity of
the foreign body relative to the surrounding tissue.
These features, and additional acoustic characteristics, are helpful in identifying the position of a foreign body. Figure 3.148 shows a foreign body that has
penetrated the sclera. Because of the high-amplitude echoes in the
P.113
surrounding sclera, the foreign body itself cannot be seen here, even at lower gain. However, a path of hemorrhage through the vitreous leads to the presumed site
of foreign body penetration. A trail of reduplication echoes posterior to midvitreal or intrascleral foreign bodies will allow the examiner to trace back to the position of
the foreign body. In all cases, the foreign body will produce a high-amplitude spike on the A-scan, which will maintain its height even at reduced gain.
Figure 3.147. B-scan of an eye with a foreign body in the anterior vitreous. Absorption of sound by the foreign
body produces a defect in the sclera posterior to the foreign body along the acoustic path. These absorption
defects can often be useful in identifying foreign bodies or calcific lens fragments, or calcific changes seen in
such diseases as retinoblastoma, which absorbs ultrasound excessively.
Figure 3.148. 10-MHz B-scan of intraocular foreign body. Left: Highly reflective metallic foreign body seen
inferior to lens. Note trailing reverberation echoes. Right: Foreign body echo remains prominent at reduced gain
setting.
Magnetic Foreign Bodies
The magnet test during ultrasonic examination is one of the most useful preoperative studies in the evaluation
P.114
of foreign bodies. The test was initially described independently by Purnell (28) and by Penner and Passmore (22) and uses ultrasonic display of the motion of the
foreign body induced by a magnet; it is usually performed with a pulsed magnet and the A-scan as a simple visual correlation. (It could be used with B-scan as well,
if the magnet is not likely to magnetize the transducer scanning system. If this is not known, it is better to stay with the A-scan transducer.)
We use a Bronson-Magnion pulsed magnet, so that the easily recognized pulsating movement of the foreign body can be related to the lack of response from the
surrounding tissue structures (Figure 3.149). The magnet should be placed in position over the pars plana so that any induced motion of the foreign body will not
displace it into the lens or other delicate ocular structures. The magnet should be turned on while positioned well away from the eye, so that there is minimum
excursion of the foreign body. The magnet is drawn closer to the eye until motion of the foreign body is seen on the M-scan or the A-scan. The M-scan can
demonstrate the velocity of movement, the amount of excursion, and the recoil of the foreign body to its original position. A nonmagnetic foreign body will not
produce any motion on the M-scan. These graphs, in conjunction with the suspected mass of the foreign body as determined by x-ray, can indicate the likelihood of
successful magnetic extraction as well as direct the optimum position for surgical incision, whether at the pars plana or directly over the foreign body.
Radiolucent Foreign Bodies

Suspected foreign bodies of glass, plastic, wood, and other nonradiopaque materials require careful serial sectioning for ultrasonic localization. Glass or plastics,
particularly when discrete surfaces are present, can be well visualized ultrasonically. Figure 3.150 shows a piece of glass posterior to the lens, underlying the ciliary
body. In general, glass, plastics, or wood material (Figure 3.151) do not have the mass and velocity to penetrate deeply into the eye and are usually seen in the
anterior chamber, the lens, or anterior vitreous. We have found it difficult, or even impossible, to localize small pieces of glass in the sclera, angle, or cataractous
lens. Occasionally, when previous radiographic localization has been performed, a foreign body can be located within the lens, but, because of the layered
structure of the lens, traumatic separation of planes may make it difficult to absolutely distinguish tissue planes from intralenticular foreign bodies. Except for wood,
these materials are usually inert, and the reduced efficacy of ultrasound in these situations is less critical than it would be in the identification of metallic materials.
Figure 3.149. M-scans of an intraocular foreign body showing the velocity and rate of motion as well as the rate
of recoil to the initial position. The magnet should be positioned relatively far from the globe at the initiation of
the test, so that introduction of the magnetic field will not pull the foreign body unexpectedly into the ocular wall
or into the lens.
The uses of ultrasound in foreign body management are summarized in Table 3.6.
Very high frequency ultrasound has added significantly to the ability to discern small fragments and foreign bodies in the anterior segment. Residual
perfluorocarbon or silicone can be seen as tiny, reverberating foreign body images, as seen in Figure 3.83. IOL haptics now enjoy the status of the most frequent
foreign bodies that need to be localized. This can be important for determining the need for corrective surgery, as shown in Figure 3.152.
NEWER IMAGING MODES

Synopsis
The future of ophthalmic ultrasound will be enhanced with new transducer arrays, increase in computer generated imaging, and fusion techniques using synergies
with other imaging modalities.
P.115
Figure 3.150. Left: 10-MHz B-scan image shows glass foreign body temporally. Note trailing reverberation
artifact. Right: 50-MHz image shows foreign body to be resting on lens at the level of zonular insertion.
Figure 3.151. Small wood fragment seen in the iris at 50 MHz. These fragments are almost impossible to see at
lower frequencies.
Figure 3.152. An IOL haptic displaced posteriorly into the ciliary processes, causing symptoms that required
repositioning of the lens.
TABLE 3.6 Uses of Ultrasound in Intraocular Foreign Body Management
Foreign Body Localization
Axial Length Measurements to Augment X-ray Localization
Assessment of Associated Globe Damage
Determination of Magnetic Properties Using Pulsed Magnet
Extraction of Nonmagnetic Foreign Bodies Using Ultrasonic Data
P.116
20-Megahertz Imaging
Conventional ultrasound examination of the eye and orbit is performed at a frequency of approximately 10 MHz. Very high frequency ultrasound, or ultrasound
biomicroscopy, involves frequencies of 25 MHz or higher, and, because of the effect of attenuation, VHF ultrasound is restricted to the anterior segment. There is,
however, a midrange of frequencies that has recently been introduced to clinical practice. At 20 MHz, spatial resolution is double that attainable at 10 MHz and
attenuation, although significant, remains small enough to permit imaging of both the posterior and anterior segments. We use an immersion 20-MHz imaging
system for evaluation of the posterior segment. Examples of images produced using this system are provided in Figure 3.153. Commercial ultrasound systems
using 20-MHz transducers have been developed for imaging of both the anterior and posterior segments. Innovative Imaging, Inc., produces a system designed for
wide-angle imaging of the anterior segment. In this sector scan system, the transducer may be coupled to the eye either with a fluid standoff established with a
scleral shell or by enclosing the tip of the transducer in a fluid-filled sheath (tono-tip), which is then placed in contact with the globe. Quantel Medical produces a
20-MHz enclosed sector scan probe (Figure 3.154) as an option with their 10-MHz B-scanner, and a similar system is produced by Optikon. Although 20-MHz
images of the anterior segment do not provide the resolution of VHF systems, they can, in many instances, provide clinically significant information in situations
where 10-MHz systems are inadequate, allowing, for instance, assessment of IOL placement, glaucoma syndromes, hypotony, tumors, and cysts. Imaging of the
posterior segment at a frequency of 20 MHz allows improved assessment of pathologies, such as macular degeneration, cystoid macular edema (Figure 3.155),
retinal holes, and small tumors.
Figure 3.153. A very high frequency 22-MHz image at the posterior pole of the normal eye demonstrating the
sclera-Tenon's boundary as well as the retina-choroid thickness. Note the excellent resolution of the thickened
sclera near the optic nerve.
Figure 3.154. 20-MHz image of the posterior pole made with the Cinescan from Quantel demonstrating the
separation of sclera and Tenon's and the thickness of the posterior coats of the eye.
P.117
Figure 3.155. Left: 20-MHz image using the Quantel Cinescan demonstrating a vitreous traction membrane in
a patient with cystoid macular edema. Right: A high frequency posterior pole image of a patient with cystic
macular edema, demonstrating separation of sclera and Tenon's, as well as accentuation of the optic nerve
sheath. This indicates that there is often a tenonitis accompanying the cystoid macular changes.
The usefulness of 20-MHz imaging must be placed in the context of other technologies, including conventional 10-MHz B-scan and optical coherence tomography
(OCT). An ultrasound of 20 MHz is superior to that of 10 MHz in situations where resolution is crucial and penetration is less important. The OCT-3 (Zeiss
Humphrey Systems) provides an axial resolution of 10 microns and a lateral resolution of 20 microns, far higher resolution than even 20-MHz ultrasound. However,
OCT suffers greatly from attenuation as it passes through optically absorbing tissues, and a penetration of less than 1 mm is obtained. OCT images provide fine
detail of retinal pathology but little information at the level of the choroid or deeper. In small (<1.5 mm thick) pigmented lesions, OCT is of little value because of
light attenuation by melanin. In such lesions, the improved resolution of 20 MHz becomes invaluable for assessment of size, choroidal involvement, and the
presence of extraocular extension. In situations where opacities exist along the optic axis (cataract, hemorrhage), OCT cannot be used, and 20-MHz ultrasound
becomes the sole imaging method.
I-scan, C-scan, 3-D, Swept-scan Vascular Imaging, and Image Fusion Technology
In the evaluation of ophthalmic ultrasound, several imaging modes have shown promising improvements in diagnostic potential that adumbrate changes to follow.
Digital analytic imaging was mentioned earlier in relation to identification and separation of echoes in the cornea and the choroid. Certainly, accuracy is vastly
improved, especially with very high frequency scans.
The use of 3-D ultrasound with coronal multiplanar reconstruction adds new means of evaluating ocular pathology. We first used 3-D in ophthalmology to
demonstrate volume (Figure 3.156) and aid diagnosis (Figure 3.157; see also DVD) (142). Fisher has shown remarkable definition in scans of the posterior pole
with this method, as shown in Figure 3.158 (see also DVD). Silverman and Ferrara (44) have demonstrated that swept scan technology (Chapter 2) can allow
vascular patterns and flow characteristics to be both demonstrated and measured (Figure 3.159; see also DVD).
Image fusion and multispectral techniques allow us to combine data from imaging techniques, such as OCT and ultrasound, to improve diagnosis of retinal and
P.118
choroidal pathology, such as nevi and small melanomas (Figure 3.160).

Figure 3.156. A three-dimensional scan of a ciliary body melanoma, which can allow accurate volume
measurements to be calculated. (See also DVD.)
Figure 3.157. Three-dimensional reconstruction of a partially dislocated lens in a patient complaining of

photophobia and diplopia. The lens shown in red pseudocolor is noted to have slipped posterior to the posterior
capsule (yellow). Repositioning of the lens eliminated the symptoms. (See also DVD.) (see color image)
All of these techniques dramatize the continually evolving technology and instrumentation that lead to greater accuracy of measurement and diagnosis that
exemplify the ever expanding field of ophthalmic ultrasound.
Figure 3.158. Three-dimensional ultrasonogram from the OTI. (Courtesy of Yale Fisher, MD.) (See also DVD.)
Figure 3.159. Swept-scan analysis of the arterial circle of the iris in a rabbit, demonstrating a method of
measuring blood flow in small vessels. (See also DVD.) (see color image)
P.119
Figure 3.160. Demonstration of fusing of OCT and high resolution posterior pole images to demonstrate the use
of the OCT for confirming the separation retina and choroid and in providing an alternate measurement of retinal
thickness. (see color image)
The next chapter will pursue high frequency ultrasound examinations of refractive elements of the anterior segment.
REFERENCES
1. Mundt G, Hughes W. Ultrasonics in ocular diagnosis. Am J Ophthalmol 1956;41:488-498.
2. Oksala A, Lehtinen A. Diagnostic value of ultrasonics in ophthalmology. Ophthalmologica 1957;134:387-395.
3. Jansson F. Measurement of intraocular distances by ultrasound and comparison between optical and ultrasonic determination of the depth of the anterior
chamber. Acta Ophthalmol 1963;41:25-61.
4. Jansson F. Measurements of intraocular distances by ultrasound. Acta Ophthalmol 1963;74:1-51.
5. Oksala A, Lehtinen A. A measurement of the velocity of sound in some parts of the eye. Acta Ophthalmol 1958;36: 633-639.
6. Sorsby A, Leary GA. A Longitudinal Study of Refraction and Its Components During Growth. Special Report Series/Medical Research Council (Great
Britain) 1969;309: 1-41.
7. Coleman DJ, Carlin B. A new system for visual axis measurements in the human eye using ultrasound. Arch Ophthalmol 1967;77:124-127.
8. Giglio EJ, Ludlam WM, Wittenberg S. Improvement in the measurement of intraocular distances using ultrasound. J Acoust Soc Am 1968;44:1359-1364.
9. Oksala A. The echogram in the diagnosis of eye diseases. Klin Monatsbl Augenheilkd 1960;137:72-87.
10. Bronson NR. Techniques of ultrasonic localization and extraction. Am J Ophthalmol 1965;60:596-603.
11. Ossoinig K. Clinical echo-ophthalmology. In: Blodi F, ed. Current Concepts of Ophthalmology. St. Louis: Mosby, 1972. v.III.
12. Ossoinig K. The evaluation of kinetic properties of echo signals. In: Oksala A, Gernet H, eds. Ultrasonics in Ophthalmology. S. Karger, 1967.
13. Buschmann W, Staudt J. Fundamentals of echographic differential diagnosis. In: Ossoinig K, ed. Ultrasonographia Medica (SIDUO III). Vienna: Verlag
der Weiner Med Akad, 1971.
14. Gernet H. Ultraschall-Biometrie des Auges. Klinische Monatsblatter fur Augenheilkunde 1967;151:853-871.
15. Massin M, Poujol J. Clinical value of time-amplitude ultrasonography in 1,000 patients. In: K. Gitter et al., ed. Ophthalmic Ultrasound. St. Louis: Mosby,
1969.
16. Francois J, Goes F, Yobbagyi P. Ultrasonic echography in ophthalmology. Ann Ocul (Paris) 1968;201:609-645.
17. Vanysek J, Preisova J, Obraz J. Ultrasonography in Ophthalmology. London: Butterworth-Heineman, 1969.
18. Bertenyi A, Betko J, Greguss P. Comparison between ultrasound and isotope diagnosis of intraocular tumors. In: Ossoinig K, ed. Ultrasonographica
Medica (SIDUO III). Vienna: Verlag der Weiner Med Akad, 1971.
19. Gallenga R. The role of diagnostic ultrasound after nine years of routine clinic. In: Ossoinig K, ed. Ultrasonographica Medica (SIDUO III). Vienna: Verlag
der Weiner Med Akad, 1971.
20. Sarin L, Meyer D, Gitter K, et al. The clinical evaluation of the use of ultrasound (A-scan) in 500 patients. In: Vanysek J, ed. Diagnostica ultrasonica in
ophthalmologica (SIDUO II). Brno: University of Brno, 1968.
21. Keeney AH. Ultrasound in Clinical Ophthalmological Diagnosis. Transactions of the Pacific Coast-to-Ophthalmological Society Annual Meeting.
1967;51:89-97.
22. Penner R, Passmore JW. Magnetic vs nonmagnetic intraocular foreign bodies. An ultrasonic determination. Arch Ophthalmol 1966;76:676-677.
23. Cowden JW, Runyan TE. Localization of intraocular foreign bodies. Further experiences in ultrasonic vs radiologic methods. Arch Ophthalmol
1969;82:299-301.
24. Baum G, Greenwood I. The application of ultrasonic locating techniques to ophthalmology. Part I: reflective properties. Am J Ophthalmol
1958;82:475-479.
25. Baum G, Greenwood I. Ultrasonography—an aid in orbital tumor diagnosis. Arch Ophthalmol 1960;64:180-194.
26. Baum G. Use of ultrasonography in the differential diagnosis of ocular tumors. In: Boniuk M, ed. Ocular and Adnexal Tumors. St. Louis: Mosby, 1964.
27. Baum G. Aids in ultrasonic diagnosis. J Acoust Soc Am 1970;48(Suppl 2):1407+.
28. Purnell E. Ultrasound in ophthalmological diagnosis. In: R. Goldberg and L. Sarin, ed. Diagnostic Ultrasound New York: Plenum Press, 1966.
29. Purnell E. Ultrasonic interpretation of orbital disease. In: K. Gitter et al., ed. Ophthalmic Ultrasound. St. Louis: Mosby, 1969.
P.120
30. Purnell EW, Sokollu A, Holasek E. The production of focal chorioretinitis by ultrasound. A preliminary report. Am J Ophthalmol 1964;58:953-957.
31. Purnell E. Therapeutic use of ultrasound. In: Sarin LK, ed. Ultrasonics in Ophthalmology. Philadelphia: WB Saunders, 1967.
32. Holasek E, Sokollu A. Direct Contact, Hand Held, Diagnostic B-scanner. Proceedings IEEE Ultrasonics Symposium; 1972.
33. Coleman DJ, Konig WF, Katz L. A hand-operated, ultrasound scan system for ophthalmic evaluation. Am J Ophthalmol 1969;68:256-263.
34. Bronson NR. Development of a simple B-scan ultrasonoscope. Trans Am Ophthalmol Soc 1972;70:365-408.
35. Fisher YL, Bronson NR, Schutz JS, et al. Contact B-scan ultrasonography: clinicopathological correlations. Ann Ophthalmol 1975;7:779-786.
36. Coleman DJ. Reliability of ocular and orbital diagnosis with B-scan ultrasound. I. Ocular diagnosis. Am J Ophthalmol 1972;73:501-516.
37. Coleman DJ. Reliability of ocular and orbital diagnosis with B-scan ultrasound. II. Orbital diagnosis. Am J Ophthalmol 1972;74:704-718.
38. Coleman DJ, Lizzi FL, Jack R. Ultrasonography of the Eye and Orbit. Philadelphia: Lea & Febiger, 1977.
39. Coleman DJ, Katz L. Color coding of B-scan ultrasonograms. Arch Ophthalmol 1974;91(6):429-431.
40. Coleman DJ, Katz L, Lizzi FL. Isometric, three-dimensional viewing of ultrasonograms. Arch Ophthalmol 1975; 93: 1362-1365.
41. Coleman DJ, Weininger R. Ultrasonic M-mode technique in ophthalmology. Arch Ophthalmol 1969;82: 475-479.
42. Coleman DJ, Jack RL, Franzen LA. Ultrasonography in ocular trauma. Am J Ophthalmol 1973;75:279-288.
43. Coleman DJ. Measurement of choroidal pulsation with M-scan ultrasound. Am J Ophthalmol 1971;1:363-365.
44. Silverman RH, Kruse DE, Coleman DJ, et al. High-resolution ultrasonic imaging of blood flow in the anterior segment of the eye. Invest Ophthalmol Vis
Sci 1999;40: 1373-1381.
45. Erickson SJ, Hendrix LE, Massaro BM, et al. Color Doppler flow imaging of the normal and abnormal orbit. Radiology 1989;173:511-516.
46. Ardouin M, Urvoy M, Sabouraud O, et al. L'effet Doppler. Quels renseignements peut-il fournir a l'ophthalmologiste? Bull Soc Ophthalmol Fr 1975;75:
91-100.
47. Ardouin M, Urvoy M, Herve C. Effet Dopler et arteve ophthalmique. Bull Soc Ophthalmol Fr 1976;76:1099-1102.
48. Yamamoto Y. Doppler examination of blood flow in the ocular fundus. Bibl Ophthalmol 1975;83:32-40.
49. Yamamoto Y, Hirano S, Manabe T, et al. Supersonic observations of eyes in premature babies. Part 2: Hemodynamics at eye ground. Nippon Ganka
Gakkai Zashi 1976; 80:1194-1203.
50. Lieb WE, Merton DA, Shields JA, et al. Colour Doppler imaging in the demonstration of an orbital varix. Br J Ophthalmol 1990;74:305-308.
51. Lieb W, Flaharty P, Gobel W. Color Doppler imaging—a technique to study the hemodynamics of ocular and orbital veins. Invest Ophthalmol Vis Sci
1992;33:1047.
52. Finger PT, Romero JM, Rosen RB, et al. Three-dimensional ultrasonography of choroidal melanoma: localization of radioactive eye plaques. Arch
Ophthalmol 1998; 116:305-312.
53. Lizzi FL, Astor M, Feleppa EJ, et al. Statistical framework for ultrasonic spectral parameter imaging. Ultrasound Med Biol 1997;23:1371-1382.
54. Silverman RH, Lizzi FL, Ursea BG, et al. High-resolution ultrasonic imaging and characterization of the ciliary body. Invest Ophthalmol Vis Sci
2001;42:885-894.
55. Silverman RH, Folberg R, Rondeau MJ, et al. Spectral parameter imaging for detection of prognostically significant histologic features in uveal
melanoma. Ultrasound Med Biol 2003;29:951-959.
56. Coleman DJ, Silverman RH, Rondeau MJ, et al. Noninvasive in vivo detection of prognostic indicators for high-risk uveal melanoma: ultrasound
parameter imaging. Ophthalmology 2004;111:558-564.
57. Coleman DJ, Rondeau MJ, Silverman RH, et al. Correlation of microcirculation architecture with ultrasound backscatter parameters of uveal melanoma.
Eur J Ophthalmol 1995;5:96-106.
58. Feleppa EJ, Lizzi FL, Coleman DJ, et al. Diagnostic spectrum analysis in ophthalmology: a physical perspective. Ultrasound Med Biol 1986;12:623-631.
59. Pavlin CJ, Foster FS. Ultrasound biomicroscopy. High-frequency ultrasound imaging of the eye at microscopic resolution. Radiol Clin North Am
1998;36:1047-1058.
60. Foster FS, Pavlin CJ, Harasiewicz KA, et al. Advances in ultrasound biomicroscopy. Ultrasound Med Biol 2000;26: 1-27.
61. Fisher Y, Hanutsaha P, Tong S, et al. Three-dimensional ophthalmic contact B-scan ultrasonography of the posterior segment. Retina 1998;18:251-256.
62. Coleman DJ, Silverman RH, Daly SM, et al. Advances in ophthalmic ultrasound. Radiol Clin North Am 1998;36: 1073-1082.
63. Reinstein DZ, Silverman RH, Sutton HF, et al. Very high-frequency ultrasound corneal analysis identifies anatomic correlates of optical complications of
lamellar refractive surgery: anatomic diagnosis in lamellar surgery. Ophthalmology 1999;106:474-482.
64. Reinstein DZ, Silverman RH, Raevsky T, et al. Arc-scanning very high-frequency digital ultrasound for 3D pachymetric mapping of the corneal epithelium
and stroma in laser in situ keratomileusis [Erratum appears in J Refract Surg 2001;7:4]. J Refract Surg 2000;16:414-430.
65. Pavlin CJ, Harasiewicz K, Sherar MD, et al. Clinical use of ultrasound biomicroscopy. Ophthalmology 1991;98: 287-295.
67. Coleman DJ, Fish SK. Presbyopia, accommodation, and the mature catenary. Ophthalmology 2001;108:1544-1551.
68. Coleman DJ, Rondeau MJ, Silverman RH, Lizzi FL. Computerized ultrasonic biometry and imaging of intraocular tumors for the monitoring of therapy. Tr
Am Ophth Soc 1987;LXXXV:49-81.
69. Coleman DJ, Silverman RH, Rondeau MJ, Woods SM. Ultrasonic evaluation of vitreous and retina. In: Tasman W and Jaeger EA, eds. Duane's Clinical
Ophthalmology. Philadelphia: JB Lippincott, 1992.
70. Ossoinig KC. Standardized echography: basic principles, clinical applications, and results. Int Ophthalmol Clin 1979; 19:127-210.
71. Byrne SF. Standardized echography. Part I: A-scan examination procedures. Int Ophthalmol Clin 1979;19:267-281.
72. Ossoinig KC, Byrne SF, Weyer NJ. Standardized echography. Part II: Performance of standardized echography by the technician. Int Ophthalmol Clin
1979;19:283-285.
73. Byrne SF, Ronald L, Green MD. Ultrasound of the Eye and Orbit. 2nd ed. St. Louis: Mosby, 2002.
74. Dibernardo C, Schachat AP, Fekrat S. Ophthalmic Ultrasound: A Diagnostic Atlas. 1st ed. Thieme Medical Publishers, 1998.
P.121
75. Oksala A, Varonen E. The influence of the eyeball on the ultrasonic field of the transducer and its diagnostic significance. Acta Ophthalmol
1965;43:268-271.
76. Kossoff G. Ultrasonic visualization system. In: K. Gitter et al., ed. Ophthalmic Ultrasound. St. Louis: Mosby, 1965.
77. Baum G. An evaluation of ultrasonic techniques used in measurements of eye size. Am J Ophthalmol 1967;64: 926-936.
78. Coleman DJ, Jack RL, Cardona H. Ultrasonic evaluation of eyes with keratoprostheses. Am J Ophthalmol 1972; 74:543-554.
79. Ray S, Khan BF, Dohlman CH, et al. Management of vitreoretinal complications in eyes with permanent keratoprosthesis. Arch Ophthalmol
2002;120:559-566.
80. Pavlin CJ, Foster FS. Plateau iris syndrome: changes in angle opening associated with dark, light, and pilocarpine administration. Am J Ophthalmol
1999;128:288-291.
81. Coleman DJ. Unified model for accommodative mechanism. Am J Ophthalmol 1970;69:1063-1079.
82. Koretz JF, Cook CA, Kaufman PL. Aging of the human lens: changes in lens shape upon accommodation and with accommodative loss. J Opt Soc Am A
Opt Image Sci Vis 2002;19:144-151.
83. van der Heijde GL, Beers AP, Dubbelman M. Microfluctuations of steady-state accommodation measured with ultrasonography. Ophthalmic Physiol Opt
1996;16:216-221.
84. Vilupuru AS, Glasser A. Dynamic accommodative changes in rhesus monkey eyes assessed with A-scan ultrasound biometry. Optom Vis Sci
2003;80:383-394.
85. Deng CX, Lizzi FL, Silverman RH, et al. Imaging and spectrum analysis of contrast agents in the in vivo rabbit eye using very-high-frequency ultrasound.
Ultrasound Med Biol 1998;24:383-394.
86. Rondeau MJ, Barcsay G, Silverman RH, et al. Very high frequency ultrasound biometry of angle and sulcus dimensions. J Refract Surg
2004;20(5):454-464.
87. Cibis P. Vitreoretinal Pathology and Surgery in Retinal Detachment. St. Louis: Mosby, 1965.
88. Cibis P, Yamashita T. Experimental aspects of ocular siderosis and hemosiderosis. Am J Ophthalmol 1958;48:465-480.
89. Regnault FR. Vitreous hemorrhage: an experimental study. I. A macroscopic and isotopic study of the evolution of whole blood and hemoglobin. Arch
Ophthalmol 1970;83:458-465.
90. Machemer R, Williams JM Sr. Pathogenesis and therapy of traction detachment in various retinal vascular diseases [Erratum appears in Am J
Ophthalmol 1988;105: 714]. Am J Ophthalmol 1988;105:170-181.
91. Machemer R, Sugita G, Tano Y. Treatment of intraocular proliferations with intravitreal steroids. Trans Am Ophthalmol Soc 1979;77:171-180.
92. Machemer R, Roters S. Ultrasound biomicroscopy of chronic hypotony after cataract extraction. J Cataract Refract Surg 2001;27:327-329.
93. Coleman DJ, Daly SW, Atencio A, et al. Ultrasonic evaluation of the vitreous and retina. Semin Ophthalmol 1998; 13:210-218.
94. Coleman DJ, Franzen LA. Vitreous surgery: preoperative evaluation and prognostic value of ultrasonic display of vitreous hemorrhage. Arch Ophthalmol
1974;92: 375-381.
95. Boldt HC, Brown DM, McGeorge AJ. Echographic diagnosis of degenerative retinoschisis facilitated by scleral indentation. Am J Ophthalmol
1994;118:123-124.
96. Sebag J. Imaging vitreous. Eye 2002;16:429-439.
97. Jaffe N. The Vitreous in Clinical Ophthalmology. St. Louis: Mosby, 1969.
98. Oksala A, Lehtinen A. Diagnostics of detachment of the retina by means of ultrasound. Acta Ophthalmol 1957;35: 461-467.
99. Baum G. Problems in ultrasonographic diagnosis of retinal disease. Am J Ophthalmol 1971;71:723-739.
100. Baum G, Greenwood I. The application of ultrasonic locating techniques in ophthalmology. Part II: ultrasonic slit lamp in the ultrasonic visualization of
soft tissues. Arch Ophthalmol 1958;60:263-279.
101. Coleman DJ, Jack RL. B-scan ultrasonography in diagnosis and management of retinal detachments. Arch Ophthalmol 1973;90:29-34.
102. Boniuk M, Zimmerman L. Problems in differentiating idiopathic serous detachments from solid retinal detachments. Int Ophthalmol Clin 1962;2:411.
103. Norton E. Differential diagnosis of retinal detachment. In: al WCe, ed. Symposium on Retina and Retinal Surgery. St. Louis: Mosby, 1969.
104. Boniuk M, Zimmerman L. Occurrence and behavior of choroidal melanomas in eyes subjected to operations for retinal detachments. Trans Am Acad
Ophthalmol Otolaryngol 1962;66:642.
105. Coleman DJ, Rondeau, MJ, Silverman RH, Lloyd HO, Lizzi FL. High-resolution ultrasound spectra and wavelet analysis of the choroid in AMD. Invest
Ophthalmol Vis Sci 2005;46:E-abstract 2432.
106. Reese A. Pigmented freckles of the iris (benign melanomas): their significance in relation to malignant melanoma of the uvea. Am J Ophthalmol
1944;27:217.
107. Coleman DJ, Lizzi FL. In vivo choroidal thickness measurement. Am J Ophthalmol 1979;88:369-375.
108. Coleman DJ. Evaluation of ciliary body detachment in hypotony. Retina 1995;15:312-318.
109. Schepens C. Discussion. In: Schepes C and Keagan C, eds. Controversial Aspects of the Management of Retinal Detachment. Boston: Little, Brown
and Company, 1965.
110. Reese A. Tumors of the Eye. 2nd ed. New York: Harper and Row, 1963.
111. Norton EW, Gutman F. Fluorescein angiography and hemangiomas of the choroid. Arch Ophthalmol 1967;78: 121-125.
112. Shields JA, Font RL. Melanocytoma of the choroid clinically simulating a malignant melanoma. Arch Ophthalmol 1972;87:396-400.
113. Naumann G, Zimmerman LE, Yanoff M. Visual field defect associated with choroidal nevus. Am J Ophthalmol 1966;62:914-917.
114. Tamler E. A clinical study of choroidal nevi. A follow-up report. Arch Ophthalmol 1970;84:29-32.
115. Lemke AJ, Hosten N, Bornfeld N, et al. Uveal melanoma: correlation of histopathologic and radiologic findings by using thin-section MR imaging with a
surface coil. Radiology 1999;210:775-783.
116. Kiratli PO, Kiratli H, Bozkurt F, et al. Scintigraphic imaging of uveal melanoma with 99Tcm-glutathione. Nucl Med Commun 2001;22:197-201.
117. Lemke AJ, Hosten N, Wiegel T, et al. Intraocular metastases: differential diagnosis from uveal melanomas with high-resolution MRI using a surface coil.
Eur Radiol 2001; 11:2593-2601.
118. Rebolleda G, Suarez Figueroa M, Munoz-Negrete FJ, et al. Magnetic resonance imaging in cavitary choroidal melanoma. Eur J Ophthalmol
2000;10:335-337.
119. Muscat S, Parks S, Kemp E, et al. Secondary retinal changes associated with choroidal naevi and melanomas documented by optical coherence
tomography. Br J Ophthalmol 2004;88:120-124.
P.122
120. Collaborative Ocular Melanoma Study. Comparison of clinical, echographic, and histopathological measurements from eyes with medium-sized
choroidal melanoma in the collaborative ocular melanoma study: COMS report no. 21. Arch Ophthalmol 2003;121:1163-1171.
121. Anonymous. The Collaborative Ocular Melanoma Study (COMS) randomized trial of pre-enucleation radiation of large choroidal melanoma III: local
complications and observations following enucleation COMS report no. 11. Am J Ophthalmol 1998;126:362-372.
122. Anonymous. Echography (ultrasound) procedures for the Collaborative Ocular Melanoma Study (COMS), Report no. 12, part II. J Ophthalmic Nurs
Technol 1999;18:219-232.
123. Anonymous. Echography (ultrasound) procedures for the Collaborative Ocular Melanoma Study (COMS), Report no. 12, part I. J Ophthalmic Nurs
Technol 1999;18:143-149.
124. Diener-West M, Earle JD, Fine SL, et al. The COMS randomized trial of iodine 125 brachytherapy for choroidal melanoma, III: initial mortality findings.
COMS Report No. 18 [See comment]. Arch Ophthalmol 2001;119: 969-982.
125. Oksala A. Echogram in melanoma of the choroid. Br J Ophthalmol 1959;43:408-414.
126. Oksala A. Ultrasound diagnosis in intraocular melanomas. Ann N Y Acad Sci 1963;100:18-27.
127. Ossoinig K, Till P. Methods and results of ultrasonography in diagnosing intraocular tumors. In: K. Gitter et al., ed. Ophthalmic Ultrasound. St. Louis:
Mosby, 1969.
128. Poujol J. Clinical echography in intraocular tumors. In: Bock J and Ossoinig K, eds. Ultrasonographica Medica (SIDUO III). Vienna: Verlag der Weiner
Med Akad, 1971.
129. Baum G. Ultrasonographic characteristics of malignant melanoma. Arch Ophthalmol 1967;78:12-15.
130. Coleman DJ, Abramson DH, Jack RL, et al. Ultrasonic diagnosis of tumors of the choroid. Arch Ophthalmol 1974; 91:344-354.
131. Greenwald MJ, Weiss A. Ocular manifestations of the neurocutaneous syndromes. Pediatr Dermatol 1984;2:98-117.
132. Blodi FC, Ossoinig KC. Modern diagnosis of orbital tumors. Trans New Orleans Acad Ophthalmol 1982;30: 129-149.
133. Byrne SF, Marsh MJ, Boldt HC, et al. Consistency of observations from echograms made centrally in the Collaborative Ocular Melanoma Study COMS
report no. 13. Ophthalmic Epidemiol 2002;9:11-27.
134. DiBernardo C, Pacheco EM, Hughes JR, et al. Echographic evaluation and findings in metastatic melanoma to extraocular muscles. Ophthalmology
1996;103:1794-1797.
135. Mackley T, Teed R. Unsuspected intraocular malignant melanomas. Arch Ophthalmol 1958;60:475-478.
136. Ferry A. Lesions mistaken for malignant melanoma of the posterior uvea: a clinico-pathologic analysis of 100 cases with ophthalmoscopically visible
lesions. Arch Ophthalmol 1964;72:463-469.
137. Read RW, Green RL, Rao NA. Metastatic adenocarcinoma with rupture through the Bruch membrane simulating a choroidal melanoma. Am J
Ophthalmol 2001; 132:943-945.
138. Silverman RH, Coleman DJ, Rondeau MJ, et al. Measurement of ocular tumor volumes from serial, cross-sectional ultrasound scans. Retina
1993;13:69-74.
139. Daftari I, Barash D, Lin S, et al. Use of high-frequency ultrasound imaging to improve delineation of anterior uveal melanoma for proton irradiation. Phys
Med Biol 2001;46:579-590.
140. Coleman DJ, Silverman RH, Rondeau MJ, et al. Correlations of acoustic tissue typing of malignant melanoma and histopathologic features as a
predictor of death. Am J Ophthalmol 1990;110:380-388.
141. Coleman DJ, Lizzi FL. Computerized ultrasonic tissue characterization of ocular tumors. Am J Ophthalmol 1983; 96:165-175.
142. Coleman DJ, Silverman RH, Rondeau MJ, et al. New perspectives: 3-D volume rendering of ocular tumors. Acta Ophthalmol Suppl 1992;204:22.
143. Shields JA, Shields CL, Donoso LA. Management of posterior uveal melanoma. Surv Ophthalmol 1991;36: 161-195.
144. Shields JA, Shields CL, De Potter P, et al. Diagnosis and treatment of uveal melanoma. Semin Oncol 1996;23: 763-767.
145. Shields JA, Rodrigues MM, Sarin LK, et al. Lipofuscin pigment over benign and malignant choroidal tumors. Trans Am Acad Ophthalmol Otolaryngol
1976;81:871-881.
146. Shields JA, Augsburger JJ, Brown GC, et al. The differential diagnosis of posterior uveal melanoma. Ophthalmology 1980;87:518-522.
147. Shields CL, Shields JA, Kiratli H, et al. Risk factors for growth and metastasis of small choroidal melanocytic lesions. Trans Am Ophthalmol Soc
1995;93:259-275; discussion 75-79.
148. Shields CL, Shields JA. Clinical features of small choroidal melanoma. Curr Opin Ophthalmol 2002;13: 135-141.
149. Ivekovic R, Lovrencic-Huzjan A, Mandic Z, et al. Color Doppler flow imaging of ocular tumors. Croat Med J 2000;41:72-75.
150. Fuller DG, Snyder WB, Hutton WL, et al. Ultrasonographic features of choroidal malignant melanomas. Arch Ophthalmol 1979;97:1465-1472.
151. Coleman DJ. Reliability of ocular tumor diagnosis with ultrasound. Trans Am Acad Ophthalmol Otolaryngol 1973; 77:OP677-OP686.
152. Pavlin CJ, McWhae JA, McGowan HD, et al. Ultrasound biomicroscopy of anterior segment tumors. Ophthalmology 1992;99:1220-1228.
153. Marigo FA, Finger PT, McCormick SA, et al. Iris and ciliary body melanomas: ultrasound biomicroscopy with histopathologic correlation. Arch
Ophthalmol 2000;118: 1515-1521.
154. Doro D. Optic neuropathies: diagnostic role of standardized echography. Metab Pediatr Syst Ophthalmol 1990;13: 67-71.
155. Gunenc U, Maden A, Kaynak S, et al. Magnetic resonance imaging and computed tomography in the detection and localization of intraocular foreign
bodies. Doc Ophthalmol 1992;81:369-378.
156. Ta CN, Bowman RW. Hyphema caused by a metallic intraocular foreign body during magnetic resonance imaging. Am J Ophthalmol 2000;129:533-534.
157. Oksala A, Lehtinen A. Diagnostics of rupture of the sclera by means of ultrasound. Acta Ophthalmol 1958;36:37-42.
158. Freyler H. Ultrasonic studies in experimental scleral rupture. In: Ossoinig K, ed. Ultrasonographica Medica (SIDUO III). Vienna: Verlag der Weiner Med
Akad, 1971.
159. Bronson NR II. Management of intraocular foreign bodies. Int Ophthalmol Clin 1974;14:129-150.
160. Coleman DJ, Lucas BC, Rondeau MJ, et al. Management of intraocular foreign bodies. Ophthalmology 1987;94: 1647-1653.
> Table of Contents > 4 - Very High Frequency Digital Ultrasound Scanning in LASIK and Phakic Intraocular Lenses
4
Very High Frequency Digital Ultrasound Scanning in LASIK and Phakic
Intraocular Lenses
VERY HIGH FREQUENCY DIGITAL ULTRASOUND ARC B-SCANNER

Digital signal processing of ultrasound backscatter was pioneered by Coleman and others at the Bio-Acoustic Research Facility in the Department of
Ophthalmology of Cornell University in the 1980s. In the early 1990s we began integration of very-high-frequency (VHF) probes originally designed for quality
control in the metallurgical industry into the Cornell University three-dimensional (3-D) ultrasound scanning prototype. Pavlin et al. (1), at the University of Toronto,
also produced a VHF ultrasound scanner but it was based only on conventional analog signal processing; the Toronto prototype became a commercial unit called
the Ultrasound Biomicroscope (UBM) (Humphrey Zeiss; Dublin, CA). The Cornell prototype became a commercial unit called the Artemis (2) (Figure 4.1) (Ultralink
LLC; St. Petersburg, FL).
The Artemis arc-scanner was designed to help ophthalmologists in all disciplines, but particularly in refractive, cataract, and presbyopic surgery, to improve
anatomic diagnosis for surgical planning and postoperative diagnostic monitoring. The Artemis's primary functions are to provide very high resolution ultrasound
B-scan imaging of the anterior and posterior segment, high-precision 3-D mapping of individual corneal layers, 3-D mapping of anterior segment dimensions, and
axial length by a combined additional immersion A-scan probe. The Artemis is designed to scan in an arc of adjustable radius, thus following the curved surfaces of
either the cornea, the iris plane, or the globe, and enabling wide segments (up to 15 mm) to be imaged in one scan sweep.
The resolution of the Artemis, when set to scan cornea, is sufficient to distinguish individual corneal layers, such as the epithelium, stromal component of the flap,
residual stromal bed, and others, all in 3-D, thanks to multimeridional scanning. The Artemis VHF digital ultrasound technology is able to consistently detect internal
corneal lamellar interfaces (such as the keratectomy track) because of the permanent “mechanical” interface present, even years after surgery, and despite total
optical transparency. Optical coherence tomography (OCT) has been shown to be capable of detecting the interface in LASIK in the early postoperative period, but
this ability diminishes with time as edema subsides in the cornea, and the optical properties of the corneal lamellar interface homogenize. We have scanned former
nonfreeze keratomileusis patients more than 10 years after surgery and have been able to clearly delineate, end-to-end, the stromal lamellar interface (Figure
4.4A).
In 1993, we reported the first confirmed measurement of the epithelium of the cornea in vivo, using VHF ultrasound, demonstrating that acoustic interfaces that
were being detected were indeed located spatially at the epithelial surface and at the interface between epithelial cells and the surface of Bowman's layer (3). We
also reported the first high-precision 3-D thickness mapping of the corneal epithelium and flap (4). This system, acquiring a series of parallel, rectilinear B-scans,
was capable of mapping the epithelial layer thickness within the central 3- to 4-mm area. By using digital signal processing techniques (the I-scan), a 2.0-micron
reproducibility
P.124
for epithelial thickness measurements was obtained (5). The I-scan is an A-scan like trace produced by digital processing of the stored radiofrequency ultrasonic
data. The trace represents the instantaneous energy intensity with time, as opposed to the average amplitude, as is represented by the conventionally used
A-scan. Previous studies demonstrated that the I-scan more than doubles the measurement precision afforded by the analog A-scan process (5). We further
improved epithelial thickness measurement precision to 1.3 microns by increasing the fidelity of the digitized signal (6). Measurement precision within the cornea in
LASIK has been formally tested and published. The axial measurement precision within 9-mm wide corneal scans is of approximately 1 µm (2). When scans are
expanded to include the entire anterior segment (15-mm width), the axial precision remains similar, whereas the lateral precision for measuring angle-to-angle is
0.15 mm and from sulcus-to-sulcus is 0.20 mm (7).
Figure 4.1. Artemis 2: VHF digital ultrasound 50-MHz 3-D arc B-scan (Ultralink, LLC). (see color image)
The VHF digital ultrasound system has been used to characterize central epithelial lenticular anatomy and to demonstrate that the power of the epithelium is not
constant from eye to eye (8). We have also examined the shape of Bowman's layer (9), the measurement of anterior corneal scars for planning therapeutic
keratectomy (10, 11, 12), the quantitative analysis of corneal scarring (haze) after photorefractive keratotomy (PRK) (13), and the measurement of the depth of
radial keratotomy incisions (14). In 1999, we were the first to publish on the analysis of epithelial and stromal changes after lamellar corneal surgery, demonstrating
significant epithelial changes after uncomplicated LASIK and the masking of stromal surface irregularities that were producing optical complications (6). This
chapter will be focused on this application.
ARTEMIS TECHNOLOGY
Details of the scanning and signal processing technology have been described comprehensively elsewhere (2, 3,10,15). A broad-band 50-MHz VHF ultrasound
transducer (bandwidth approximately 10 to 60 MHz) is swept by a high-precision arc mechanism to acquire B-scans as arcs that follow the surface contour of
anterior or posterior segment (with a reverse arc) structures of interest. The Artemis possesses a unique scan-arc adjustment mechanism that allows maximum
perpendicularity (and signal-to-noise ratio) to be obtained for scanning any of the different curvatures within the globe (cornea, iris plane, retina). Ultrasound data
are digitized and stored. The digitized ultrasound data are then transformed, using digital signal processing technology. Digital signal processing significantly
reduces noise and enhances signal-to-noise ratio. We have demonstrated that using digital signal processing on 50-MHz ultrasound data doubles resolution and
increases measurement precision by a factor of 3 when compared to conventional analog processing of the same very-high-frequency data (5). Scanners produced
by Paradigm (UBM), OTI (35 MHz), and others use only analog ultrasound processing. As a result of coaxial, simultaneous video image capture at each scan
position (Figure 4.2), a correlation of measurements made from the ultrasound scans can be formed into visible ocular landmarks (such as the corneal reflex) that
P.125
enable accurate 3-D reconstructions to be made from multiple meridional scans and to allow production of corneal mapping. Simultaneous optical and ultrasound
imaging also enables the anterior segment sulcus-to-sulcus distance to be determined in a verified plane, such as the visual axis for surgical planning in phakic
intraocular lens (IOL) surgery. For the first time, it also enables localization of the optimum implantation site for devices, such as scleral expansion bands for
presbyopia, which need to be positioned based on internal (invisible) landmarks. The Artemis possesses a software application that will give the surgeon external
landmarks, identifiable under the operating microscope, that identify the location of lens equator based on a caliper measurement from the corneal reflex (Figure
4.3).
Figure 4.2. Artemis advanced control display panel. The upper left panel shows an infrared real-time video
image of the eye being scanned in which eye position can be verified and monitored during scanning. The lower
left panel is used for scan motion control, whereas the upper right panel displays the raw ultrasound echo data.
In this screenshot, the lower right anterior segment scan was known to have been taken in the horizontal plane
when the eye was fixating on a light source coaxial with an alignment beam that is centered on the corneal
vertex (corneal reflex visible). The patient's angle kappa produces a geometric tilt of the anterior segment
compared to the visual axis (green line). The corneal reflex is an excellent landmark for correlating scans taken
before and after anatomy by subtraction imaging. (see color image)
Although Artemis scanning is a noncontact test, it does require an ultrasonic standoff medium, and thus provides the advantages of immersion scanning. The
Artemis 2 was designed specifically to enable quick setup of this immersion scanning by a novel reverse-immersion technique. The patient sits and positions his or
her chin on a three-point forehead and chin rest, while placing the eye into a soft rimmed eye cup, akin to a swimming goggle (Figure 4.4). The sterile coupling fluid
fills the compartment in front of the eye, and the scanning is performed via an ultrasonically transparent (sterile) membrane, without the need for a speculum. Thus,
there is no contact by the scanner probe with the eye. Performing a 3-D scan set with the Artemis requires 2 to 3 minutes for each eye.
Figure 4.3. Annotated arc B-scan ultrasound image showing all measurements required for the accurate
implantation of a scleral expansion band. The intersection of the cornea with the line-of-sight is indicated by
arrow and C. The lens equator plane is localized based on the ultrasound image, and the eternal intersection of
this plane at the scleral surface is localized. The distance from C to the equatorial plane is identified for exact
localization of the scleral implant to achieve maximum effect. The thickness of the sclera is provided to
maximize depth without intraoperative exposure of the choroid. (see color image)
CLINICAL USE
Two-dimensional B-scan Imaging

Figure 4.5 demonstrates an arc B-scan taken along the horizontal plane of the cornea of a patient 4 months after LASIK. The interfaces of saline-epithelium (E),
epithelium-Bowman's (B), the keratectomy interface (K), and the posterior surface (endothelial-aqueous) (P) are clearly visualized along the 9-mm chord-length of
the B-scan preoperatively. The keratectomy interface can be seen with an entrance track nasally (S), coursing
P.126
temporally to a stop at the hinge (H). Magnification of the keratome entrance position shows that the flap was not fully distended and Bowman's was not fully
apposed, potentially inducing astigmatism and/or increasing the risk of epithelial ingrowth. The interface track has a small irregularity (I) (magnified insert), perhaps
caused by a patient squeeze during passage of the keratome. The flap can be seen to be thicker temporally and thinner (T) nasally.
Figure 4.4. Patient demonstrating the simple setup of the reverse immersion scanning system. Head
stabilization is achieved by the patient resting against a tripod of support points; an adjustable chin rest and two
adjustable forehead rests. The eye rests comfortably in a sterile cushioned eye-seal that produces a separate
sterile compartment for the eye, from the fluid-filled scanner mechanism compartment.
Figure 4.5. Horizontal B-scan through the visual axis of a cornea 4 months post-LASIK. The interface is clearly
visualized throughout the length of the keratectomy. See text for annotations.
Three-dimensional C12 Diagnostic Display

This display configuration and format form the mainstay, and state-of-the-art, in anatomic diagnosis after LASIK. Figure 4.6 shows such a display created from
scans of the right cornea of a patient scanned before and 6 months after LASIK for myopia of -4.75, -0.25 × 55. Uncorrected visual acuity (UCVA) was 20/16 with a
residual subjective manifest refraction of plano. Videokeratographic examination showed the customary central flattening with a small surface with-the-rule
astigmatism. The lamellar interface was only faintly detectable in places by slit-lamp examination.
This display of 12 pachymetric maps was designed as a standardized layered pachymetric summary of corneal anatomic changes following LASIK. We have
chosen to name this presentation a C12 diagnostic display, for it consists of 12 corneal pachymetric topographic maps of the same cornea before and after LASIK.
Each map depicts the local thickness of a given corneal layer represented on a color scale in µm. The C12 display was designed as a layout of map groupings by
time, anatomic depth, and calculation. Columns 1 and 2 depict maps preoperatively and postoperatively, respectively. Within these two columns, the rows
represent depth within the cornea. Thus, the first column depicts the thickness profiles of the preoperative corneal epithelium (Figure 4.6; map 1), full stroma
(Figure 4.6; map 2), and full cornea (Figure 4.6; map 3), respectively. The second column demonstrates the postoperative thickness profiles of the corneal
epithelium (Figure 4.6; map 4), stroma (Figure 4.6; map 5), and full cornea (Figure 4.6; map 6). Epithelium, full stroma, and full cornea color scales are identical for
preoperative and postoperative stages to allow direct color (thickness) comparison. The third column consists of calculated maps representing topographic
epithelial change (Figure 4.6; map 7) (derived by subtraction of the preoperative from postoperative epithelial map), stromal change (Figure 4.6; map 8) (derived by
subtraction of the postoperative from preoperative stromal map), and (calculated) original flap produced at the time of surgery as a Flap Profile (16) (Figure 4.6;
map 9). The Flap Profile is calculated by adding the
P.127
stromal component of the flap (Figure 4.6; map 12) to the preoperative epithelial thickness. The fourth column represents postoperative corneal layers: the
thickness profile of the flap at 6 months (including epithelial changes) (Figure 4.6; map 10), the 3-D thickness profile of the residual stromal layer (stroma excluding
the flap) (Figure 4.6; map 11), and the postoperative stromal component of the flap (Figure 4.6; map 12).
Figure 4.6. C12 display of the cornea of a patient pre- and 6 months post-LASIK OS. All 12 maps are
pachymetric representations of particular corneal layers depicted on a color scale in microns. The preoperative
epithelial (1), stromal (2), and full corneal (3) thickness maps appear in the first column. To the right of each of
these maps (column two) is the post-LASIK pachymetric maps of epithelium (4), stroma (5), and full cornea (6)
on identical color scales for direct comparison to preoperative. The third column depicts calculated maps only.
The calculated epithelial change map (7) (column 3, row 1) is derived in point-by-point subtraction of the
preoperative from the postoperative epithelial pachymetric map. Thus, the epithelial change map shows on a
color scale the number of microns that increase as a result of surgery. Note that the pattern of epithelial
thickness change is such that it is greatest centrally, with a decrease in a symmetric centrifugal fashion, thus
producing an increase in outer curvature of the postoperative cornea. Note that the area of epithelial thickening
is confined to the ablation zone or the zone of surgical corneal flattening. The calculated stromal change map
(8) (column 3, row 2) is derived in point-by-point subtraction of the postoperative from the preoperative stromal
pachymetric map. Thus, the stromal change map shows on a color scale the number of stromal microns that
decrease as a result of surgery in a topographic fashion and hence represents the ablation volume of tissue.
The calculated map of the “original flap” (9) (column 3, row 3) is derived by addition of the preoperative epithelial
thickness profile (1) to the postoperative “stromal component of the flap” (12) (column 3, row 3). One must
perform a temporally displaced addition of epithelial and stromal components of the flap separately because of
the epithelial changes present post-LASIK, leading to a flap anatomy post-LASIK (10) (column 3, row 1) that is
different from that at the time of creation by the keratome. Finally, the pachymetric topography of the “residual
stromal layer” comprising all stroma beneath and around the flap is shown in map 11 (column 3, row 2). This
map can be critically important in the determination of adequacy of the stromal bed for further LASIK
enhancement surgery under the flap in that the thinnest point is not always located centrally and may be missed
by any form of intraoperative single-point measurement of the bed. Thus, the “C12” display is set out to be read
by temporal grouping (columns) or anatomic grouping (rows). See text for further descriptive analysis. (see color
image)
The profile map of the preoperative epithelium OS was approximately 9.25 mm in diameter (Figure 4.6; map 1). The epithelial change map (Figure 4.6; map 7)
shows the pattern of epithelial thickening and thinning. The epithelium thickened between 15 and 20 µm centrally, with a concentric decrease in thickening
progressing toward the 7.5-mm diameter zone. Interesting to note is that within a 1-mm annulus at the 8-mm diameter zone there was circumferential epithelial
thinning after LASIK. We also note in this case that the pattern of epithelial change increased anterior corneal power (greater tissue addition centrally), but the
patient had a plano refraction
P.128
postoperatively. This indicates that the optical power shift produced by the epithelium, in this case, was exactly as expected by the nomogram setting used.
The stromal change map (Figure 4.6) shows a wellcentered difference around the center (0,0 coordinate) of the cornea. The difference in stromal thickness prior to
surgery is 70 µm centrally, decreasing to zero at the 7.5-mm diameter zone. Thus, the zone depicted on the color scale from green to red represents the effective
volume of tissue change in the cornea (the predicted central ablation depth by the Nidek EC5000 readout was 73 µm for a 6.5-mm optical zone, transition to 7.5
mm). Within the peripheral 8- to 9-mm zone there is annular stromal thickening of between 10 and 20 µm. We were the first to publish this finding (2), and Roberts
has proposed a mechanism to account for it (17). Also interesting to note is that the annulus of stromal thickening coincides with the annulus of epithelial thinning
described previously, consistent with the Law of Epithelial Compensation (see later text).
Examination of the anatomy of the calculated original flap (Figure 4.6; map 9) by the Moria LSK-One keratome (predicted mean 160 µm) reveals a central thickness
of 158 µm. Within the 4-mm diameter zone, the flap thickness was generally homogeneous between 160 and 165 µm, although irregularity is evident. Note that
direct measurement of the flap thickness at 6 months (Figure 4.6; map 10) would not provide an accurate description of the flap anatomy at the time of creation
because of the epithelial thickness changes present after LASIK. The stromal component of the flap (Figure 4.6; map 12) can be seen to possess a thickness
profile of approximately 110 to 120 µm within the central 6-mm diameter zone, except for the quadrant superotemporally within the 4-mm diameter zone, where this
is decreased to approximately 95 µm. This area may have been thinner because of the presence of thicker epithelium preoperatively in the corresponding quadrant
and the passage of the keratome parallel to the surface of the cornea during applanation by the keratome head.
The 3-D thickness profile of the residual stromal layer (Figure 4.6; map 11) shows a thinnest point of 280 µm, approximately 1 mm inferior to the center of the
cornea. This is an example of why intraoperative handheld ultrasound residual stromal pachymetry can be misleading lateral position variations of only a few
hundred microns could completely alter the course of an ablation by providing a residual stromal thickness that is not the minimum.
PREOPERATIVE ASSESSMENT
Corneal Thickness Profile: Minimum Thickness and Screening for Keratoconus

The importance of accurate preoperative corneal thickness profile determination is now generally accepted as an aid in the determination of candidacy for safe
LASIK with avoidance of ectasia (18). Concentricity of the thickness profile around the corneal center is also a contributor in screening for keratoconus. (Because
of the significant, added expense to the patient for Artemis scanning, at present we offer to, but do not routinely, use this preoperatively in every patient.) Current
indications for Artemis scanning in our practice include a greater than 15-µm discrepancy between Orbscan and handheld ultrasound pachymetry and a predicted
residual stromal thickness of less than 300 µm, based on whichever is the thinnest of Orbscan or handheld-ultrasound pachymetry.
The accuracy of measurement is defined as the concordance between the measured and the true value. A theoretical error analysis to estimate the accuracy of
Artemis pachymetry has been published (2). The accuracy of Artemis thickness measurements within the cornea was found to be at worst ±1.8%. This means that
the 95% confidence interval for concordance between the measured and the true value is expected to be within ±5 µm for corneal thickness measurements [mean
thickness 515 µm by VHF digital ultrasound (5)].
Optical methodology for the determination of corneal back surface shape and hence 3-D corneal thickness mapping, although possessing the convenience of in-air
data acquisition, suffers from variable accuracy (19, 20, 21), almost certainly, as a result of the variable optical properties of the cornea before and after corneal
refractive surgery (22). But variations in refractive index of the cornea probably also exist between normal, unoperated individuals. To test the difference in
accuracy between Orbscan and 3-D VHF digital ultrasound scanning, we determined the thinnest point of the cornea in 52 eyes using the two devices. The
variance of pachymetry measurements was 25 µm greater for Orbscan measurements than for VHF digital ultrasound measurements. (95% confidence interval
±35 µm). This implies that VHF digital ultrasound measurements are 7% more accurate.
POSTOPERATIVE ASSESSMENT: TRUE DIAGNOSIS AFTER LASIK AND OPTIMAL TREATMENT

PLANNING WITH ARTEMIS TECHNOLOGY
Although today LASIK and PRK are already relatively safe procedures, we are constantly striving to make them even safer. Preventing all complications is not
possible, and when these do occur we need methods for correcting them and restoring visual function. In keeping with basic principles of surgery, accurate imaging
and biometry are the cornerstone of these goals, because accurate diagnosis enables optimal treatment planning.
Surface topography has been the mainstay of diagnostic testing in complicated LASIK. The introduction of aberrometry has greatly enhanced our diagnostic
capabilities in being able to understand in a quantitative way how irregular astigmatism and other shape irregularities
P.129
produce visual complaints. However, understanding neither the optical defect nor the surface shape of the cornea will necessarily provide a diagnosis for the cause
of the problem (6). The anatomic cause of a surface abnormality may only be understood at an internal corneal level, for example, irregularities in the flap versus
the stromal bed. With burgeoning surgical rates of PRK and LASIK worldwide, it is becoming increasingly evident that there is a distinct need for a method of
determining the layered anatomy of the changes induced. Without an accurate anatomic diagnosis, topography or wave front guided treatments may lead to a
suboptimal treatment plan.
The development of digital VHF ultrasound corneal scanning technology was first reported in 1991, where digital signal processing was used to identify and
analyze the epithelium and scar layers formed in an experimental rabbit model (10). In 1993, we reported the first confirmed measurement of the epithelium of the
cornea in vivo, demonstrating that the acoustic interface detected within the intact cornea was localized spatially at the interface between epithelial cells and the
surface of Bowman's layer (3). This system, acquiring a series of parallel, rectilinear B-scans, was further developed to enable mapping the thickness profile of the
epithelium (4) as well as the lamellar flap within the central 3- to 4-mm area. By using digital signal processing techniques (the I-scan), a 2-µm reproducibility for
epithelial thickness measurements was obtained (5). Subsequently, by increasing the fidelity of the digitized signal, flap thickness measurement precision was
further improved to 1.3 µm, with epithelial and corneal measurements attaining a reproducibility of under 1 µm (2). In clinical application, analysis of epithelial and
stromal changes after lamellar corneal surgery has demonstrated significant epithelial changes after uncomplicated LASIK and the masking of stromal surface
irregularities that were producing optical complications (6).
The importance of epithelial changes in corneal refractive surgery has probably been underestimated. Significant changes in epithelial thickness profiles in both
PRK (23,24) and LASIK (25, 26, 27) have been demonstrated and implicated in regression as well as the inaccuracy of topographically guided excimer laser
ablation (6). The curvature of Bowman's layer in the center of the normal cornea is, on average, greater than that of the epithelial surface (9). As the refractive
index of epithelium and stroma are sufficiently different (1.401 versus 1.377) (28), the epithelial-stromal interface constitutes an important refractive interface within
the cornea, with a mean power contribution estimated at approximately 3.60 D (9). Thus, unpredicted changes in the epithelial lenticule after surgery will result in
unplanned refractive shifts, which is one of the reasons why current ablation depths and profiles (“normograms”) differ from theoretic ablation profiles: they
incorporate the average change of epithelial power for a given level of stromal surface flattening (level of myopia treated). Thus, the understanding of epithelial
dynamics and their patterns begins to unfold (26,27), and these factors may potentially be used to improve the accuracy of corneal refractive outcomes.
Preoperative ultrasound scanning can significantly contribute to LASIK accuracy and safety. Accuracy in LASIK translates to the chances of an eye achieving target
refraction. Safety relates to achieving this target without loss of best spectacle-corrected visual acuities (BSCVA) or other visual disturbance.
Ectasia is one of the most devastating potential consequences of LASIK, and it behooves us to prevent it from happening in every possible way. The thickness of
the flap determines at what level stromal tissue removal commences, and hence is directly related to the amount of stromal tissue remaining in the posterior cornea
under the flap after surgery. The thinner the flap the more difficult it is to handle surgically; however, the thicker the flap, the less tissue remains for the correction of
ametropia by LASIK. Despite all of the advances in corneal topography and ocular wave front measurement, diagnosing the cause of subjective visual complaints
by these means alone is not always possible (6). The reason is because the internal corneal refractive interfaces (such as the epithelial-stromal interface) are not
being measured independently. In fact, topography is often not, strictly speaking, a diagnostic test, but rather a descriptive one. For the diagnosis and correction of
complications, identifying the anatomic cause of a corneal surface abnormality—front or back—may be possible only by understanding the layered internal corneal
anatomy. For example, the distinction between irregularities in the flap profile (keratome), flap positioning (surgeon), and the stromal bed (laser) will aid in planning
further surgical correction. In addition, further surgery on the cornea should always be based on a full knowledge of the remaining tissue available.
The next section describes several examples of cases within, and commonly referred to, our practice for ultrasound anatomic evaluation after complicated LASIK,
in which Artemis provided essential information for further treatment planning and some clinical examples demonstrating the importance of distinguishing
biomechanical from epithelial components of ametropia, after an initial treatment.
Microfolds
The occurrence of microfolds in the LASIK flap is often a visually compromising complication. Numerous suggestions have been made as to how to treat
microfolds, but not all have been based on an anatomic diagnostic classification.
We have studied the anatomic morphology of microfolds, while correlating the VHF digital ultrasound scans to clinical slit-lamp examination and functional impact
on vision. Founded upon these studies, we have devised
P.130
a classification system that is based on clinical management options. The Reinstein classification is shown in Table 4.1. Folds are initially classified as involving
either the stromal component of the flap or Bowman's layer alone. If involving the stromal component of the flap, these are classified as flap corrugations,
representing gross flap malposition that, in effect, lead to undulation and waviness of stromal lamellae within the entire flap substance. Flap corrugations must
clearly be managed by flap lifting and repositioning. If the folds involve Bowman's layer, a distinction must be made between true microfolds and Bowman's cracks
because their management is completely different. True microfolds are literally grooves in Bowman's layer (Figure 4.7) produced by redundancy as a result of flap
malposition or incomplete distension. Bowman's cracks are fractures in Bowman's layer with no grooving (Figure 4.8). Bowman's cracks are caused by trauma to
the flap, which may result from either the stretching of Bowman's on dragging the flap from the hinge with a spatula or from folding or bending of Bowman's on
returning the flap to the bed (especially if the flap has dried during the period of ablation). The distinction between true microfolds and Bowman's cracks is very
important because although it may be warranted to relift a flap with true microfolds to adequately reposition the flap and distend the grooves to improve refraction
and BSCVA, lifting a flap with Bowman's cracks will further traumatize the flap, causing additional damage and potentially delaying the return of BSCVA.
TABLE 4.1 Reinstein Classification of Flap Microfolds in LASIK.
Type Anatomic Loss of Fluoroscein Clinical Findings Anatomical Management

Location BSCVA Pooling Basis
Corrugation Stroma [check [check Gross folds, Flap slip Flap

mark] mark] differential repositioning
pooling of
gutters,
mixed-cylinder
True Bowman's [check [check Grooves in Grooves Flap

Microfolds mark] mark] Bowman's in repositioning
Bowman's and
microfold
distension
Bowman's Bowman's [check ? Gray line, no Fractures Observe

Cracks mark] grooves in only
Bowman's
These anatomic disturbances often change astigmatism and produce loss of BSCVA. Anatomic
localization within the flap is important in designating the optimal management plan.
Figure 4.7. True Bowman's microfold: horizontal VHF digital ultrasound corneal B-scan through the visual axis
of a patient 6 months after LASIK. The surface of epithelium (E), Bowman's (B), the keratectomy interface (I),
and the endothelium (P) are labeled. Inspection of the surface of Bowman's demonstrates a true microfold, with
Bowman's showing a groove, approximately 25 µm deep and 100 µm wide.
Flap Complications
The postoperative assessment of flap complications is greatly aided by VHF digital ultrasound scanning. Determination of the exact anatomy of the faulty lamellar
dissection will show at what depth the flap was created and whether the flap repositioning was optimal. This information is important in the planning of a
subsequent recutting of another flap. In addition, for cases with central flap dissections or irregularities, VHF digital ultrasound can determine whether the edges of
Bowman's layer were properly and adequately apposed to help avert epithelial ingrowth.
Buttonholed Flaps
In the following example, a 33-year-old nurse underwent LASIK with the Moria LSK-One and the Nidek EC5000 for a -5.50 sphere. The preoperative corneal
thickness was 509 µm and the predicted residual stromal thickness was 280 µm (based on a 160-µm flap). She lived at a distance and eventually presented almost
2 years later, requesting that one of the eyes, which had regressed in the first 3 months after surgery, be enhanced. General opinion, at the time, dictated recutting
flaps rather than lifting, ostensibly because flap lifting after 6 months was assumed to be difficult, and recutting was assumed to lead to less chance of epithelial
ingrowth as a result of the sharp edges of the new flap. The patient underwent recutting with the Hansatome, using the 160-µm head
P.131
and the 9.5-mm ring (aiming to go outside but superficial to the original flap to save residual stromal tissue). The result was a central buttonhole within a double flap
dissection that was replaced without performing laser ablation. Figure 4.9 shows the exact anatomic result 1 day after flap replacement. From the ultrasound scans
it appeared clear that the Hansatome flap had managed to stay superficial and within the original flap, only to then exit (Figure 4.9; X1) through Bowman's and
epithelium and then reenter (Figure 4.9; X2) the cornea to regain the plane superficial to the original flap interface. The scan reassured us that Bowman's was
properly apposed and that anatomic restoration had been achieved for a good prognosis. At 2 years thereafter, she had not developed epithelial ingrowth and
remained without loss of BSCVA.
Figure 4.8. Bowman's cracks. Horizontal VHF digital ultrasound corneal B-scan through the visual axis of a
patient 9 months after LASIK. The surface of epithelium (E), Bowman's (B), the keratectomy interface (I), and
the endothelium (P) are labeled. Inspection of the Bowman's interface demonstrates fractures or discontinuities
(*) that do not involve grooving or puckering of Bowman's.
Figure 4.9. Geometrically corrected horizontal VHF digital ultrasound corneal B-scan of a cornea 1 day after
recutting a second flap and obtaining a simultaneous separation of the original flap from the bed and a
buttonhole (A). An axially zoomed image in a different horizontal plane is shown below (B). The surface of
epithelium (E), Bowman's (B), the keratectomy produced by the secondcut Hansatome 160-head (H), and the
original Moria keratectomy interface (M) are labeled. The Hansatome interface is seen to coarse superficially to
the original keratectomy as intended from left to right, but superficializing, exiting through Bowman's (X1 ) and
the epithelium, then reentering the epithelium and crossing Bowman's (X2 ) to again find a plane superficial to
the original keratectomy. Exact anatomic apposition of Bowman's is confirmed by the scan, thus confirming
perfect flap repositioning and minimizing the probability of epithelial ingrowth in the visual axis.
P.132
Short Flaps
In 1994, we described The Law of Epithelial Compensation for irregular astigmatism (29): “Irregular astigmatism results in irregular epithelium.” The epithelium
often compensates fully for stromal surface irregularities, keratoconus being an excellent example of this. Generally accepted is that as cone formation in
keratoconus progresses, the epithelium overlying the cone becomes progressively thinner. This occurs because the epithelium becomes invaginated by the
underlying bulging stromal surface while its outer surface is smoothed by the action of 10,000 blinking events a day. (This is why keratoconus can be detected
earlier by looking at the back surface topography of the cornea rather than the front surface.) Possibly, examination of epithelial thickness profiles may provide an
even earlier and therefore more sensitive screening tool for keratoconus. According to the Law of Epithelial Compensation, if a patient presents with stable,
irregular astigmatism, by definition the epithelium has reached its maximum compensatory function.
In the following example, a 23-year-old patient underwent LASIK, in the left eye, using the Moria LSK-One microkeratome in which a short, nasal hinged flap was
obtained and the laser ablation was performed. VHF digital ultrasound scanning is shown in Figure 4.10. A large amount of epithelial compensation takes place in
cases like this, in which there are large steps in the shape of the stromal surface. This explains why neither topography-guided nor wave-front-guided ablations will
be sufficient to correct such complications. In this case, the stromal surface is asymmetric. The epithelium has compensated as much as it can but is still leaving
asymmetry, and the patient presents with topographic asymmetric astigmatism. If one were to base the corrective ablation profile on the topography or ocular wave
front now (70% epithelial surface shape dependent), there would clearly be ineffective correction of the stromal surface shape. Following such a case, the
epithelium may or may not compensate fully for the remaining stromal surface asymmetry. If it does, the topography would become regular, but the patient may still
have symptoms as a result of the significant refractive index difference between epithelium and stroma (9).
Figure 4.10. Horizontal VHF digital ultrasound corneal B-scan through the visual axis of the left cornea of a
patient in whom a slightly short flap was created and the ablation was carried out. The surface of epithelium (E),
Bowman's (B), the keratectomy interface (I), and the endothelium (P) are labeled. The abrupt termination of the
keratectomy producing a short hinge is shown (SH). Lack of ablation nasal to this has produced a large step in
the cornea. The stromal surface step is partially compensated for by epithelial remodeling; the epithelium
characteristically thins over the “bump” while thickening in the crevice produced. This cross section clearly
demonstrates why topography-guided ablations (or even wave front guided ablations, which are 70% biased to
the front surface) will not be fully successful in correcting the stromal irregularity.
The Topographic Diagnosis of Decentration: Is it Really a Laser Decentration?

Decentration is a diagnosis made postoperatively by inspection of topography. Decentration denotes off-center ablation. We have found that what appears to be
decentration by topography is not always a result of off-center ablation.
In the following example, a patient presented to us complaining of monocular double vision after LASIK. The initial refraction was -6.50 D. Treatment was carried
out with the Moria LSK-One microkeratome and the Nidek EC5000. Preoperative corneal thickness by Orbscan was measured as 516 µm. With an ablation depth
of 90 µm, the predicted postoperative residual stromal thickness was 266 µm. On examination, his UCVA was 20/70; manifest refraction was +3.00 - 3.75 × 96,
yielding a BSCVA of 20/40 + 2. Slit-lamp examination showed a clear cornea, with an unremarkable flap possessing a few very faint, faded shallow-appearing
vertical microfolds. Orbscan anterior best-fit sphere mapping is shown in Figure 4.11, providing a differential diagnosis of decentration of the ablation zone, or
ectasia. Figure 4.12 shows Zywave (Bausch & Lomb, St. Louis) aberrometry of the same eye, demonstrating comalike higher-order aberrations.
Horizontal 3-D VHF digital ultrasound B-scan cross section of the cornea revealed anatomic features that provided further diagnostic information. Figure 4.13
shows the B-scan demonstrating a flatter (F) nasal side of the cornea, with a raised (R) surface temporally, as found also on the Orbscan best-fit sphere surface
shape map. Beneath the raised (R) area the epithelial thickness is seen to be reduced, as a result of invagination by the underlying Bowman's layer (B). Bowman's
(B) is highly irregular, showing three major ultrasonic discontinuities (*), representing either cracks or microfolds in the flap surface. 3-D pachymetric topography of
this cornea is shown in Figure 4.13. The epithelial thickness profile is seen to vary continuously, filling in and smoothing out the surface of Bowman's layer. The
thinnest point within the residual stromal bed, as determined by 3-D thickness mapping in a C6 graph (post-LASIK with no preoperative data for subtraction maps)
display (Figure 4.14), is 223 µm. The residual stromal layer thickness profile appears slightly asymmetric or decentered in the nasal direction. Inspection of the
stromal
P.133
component of the flap map (Figure 4.14; second column, second row) shows the reason for this: the stromal component of the flap was thicker temporally than
nasally. The central stromal component of the flap was 80 µm, thus implying that the central flap thickness was originally approximately 130 µm (80 + 50). The
original surgeon had calculated that the patient would still have 266 µm under the flap after treatment. Given that this is 43 µm less than observed, and that the flap
was 30 µm thinner than intended, that his preoperative pachymetry (by Orbscan) was underestimated by approximately 43 µm is probable, and the original corneal
thickness must have been closer to 473 µm.
Figure 4.11. Orbscan anterior best-fit sphere (default 10-mm zone fit) plot of the cornea in a patient presenting
with monocular diplopia and a topographic diagnosis of “decentered ablation,” proved incorrect by B-scan
imaging in the plane represented by the horizontal black line. Flatter (F) and raised (R) areas are correlated to
the ultrasound B-scan in Figure 4.13. (see color image)
Figure 4.12. Zywave aberrometry displaying the higher-order wave front plot of the eye represented in Figure
4.11 of a patient presenting with monocular diplopia and a topographic diagnosis of “decentered ablation.” There
is marked coma. Conventional wisdom would dictate ablation that would involve relatively more removal of
tissue in the yellow-to-red zones. B-scan imaging (Figure 4.13) proves this to be inappropriate for this case.
(see color image)
Figure 4.13. Horizontal VHF digital ultrasound corneal B-scan through the visual axis of the right cornea of a
patient presenting with monocular diplopia and a topographic and wave front diagnosis consistent with
“decentered ablation.” The upper image (1) shows the geometrically corrected image, whereas the lower image
(2) shows the raw ultrasound data with axial zoom to better appreciate the interfaces. The surface of epithelium
(E), Bowman's (B), and the keratectomy interface (I) are labeled. Clearly noted is that Bowman's surface is
highly irregular, with numerous true microfolds (*) that were only very faintly visible on slit-lamp examination, as
a result of the impressive epithelial compensation producing excellent smoothing of the corneal surface. The
diagnosis of “decentered ablation” is clearly less likely than that of an inadequately distended flap, producing
surface asymmetry. Appropriate management would most likely involve flap distension and repositioning, not
further laser ablation.
A diagnosis was made of flap malposition and possible asymmetric biomechanical shift. In addition, the residual stromal thickness was noted to be too thin for
further under-the-flap ablation, despite the fact that the preoperative parameters would have implied that there was room for further treatment.
This case clearly illustrates the importance of anatomic diagnosis as in contrast to a topographic description, in planning the management of the complications of
LASIK. By topography alone, this case may well have been diagnosed as a decentration. The eye may well have then undergone a topographically guided
treatment under the flap. Given the low residual stromal thickness, it is conceivable that further tissue removal would have led to further mechanical shifts and an
unpredictable result, with a high possibility of inducing progressive ectasia (30).
Flap Profile Irregularity

Flap profile irregularities can lead to irregular biomechanical shifts in refraction. A patient was able to undergo LASIK for the correction of -10.00 D OD, because
she had 5.5-mm scotopic pupils and adequate corneal thickness to leave 250 µm in under an assumed 160-µm
P.134
flap. The Moria CB microkeratome using the “110” head, which is designed to cut an average of 140 µm, was used in manual mode, with a deliberate “fast pass” of
about 1 second to attempt to obtain a thinner flap. A proposal was made that VHF digital ultrasound scanning would be used postoperatively to measure flap and
residual stromal thickness to decide whether it would be possible to perform further enhancement, if required. The procedure was uneventful, and the flap
intraoperatively was noted to “feel” thin and was noted to be of “good quality for a future flap-lift.” Day 1 postoperatively the refraction was +5.00 D, with no loss of
BSCVA. This only regressed to +4.00 D by 3 months. Figure 4.15 shows a horizontal cross section of the flap 3 months after LASIK. The flap centrally was indeed
thin—the stromal component of the flap was in the region of 90 to 100 µm thick. However, peripherally the stromal component of the flap reached over 200 microns
(equating to an original flap thickness of over 250 µm). It is conceivable is that this patient's gross overcorrection was partly a result of a biomechanical cause:
Deep keratectomy peripherally may have produced excess flattening centrally through a mechanism of peripheral thickening, as proposed by Roberts (17),
combined with an effect similar to radial keratotomy—midperipheral bulging as a result of localized deep keratectomy causing central flattening. Clearly, in this
case, a knowledge of the flap anatomy created will be essential in planning further treatment; the relifting of the apparently good quality flap (as assessed
intraoperatively) for hyperopic ablation to be performed would further deepen the midperipheral keratectomy and would not be ideal, given the apparent
mechanisms producing this excessive overcorrection.
Figure 4.14. C6 corneal pachymetric map display of the thickness in microns (color scale) of the epithelium,
stroma, full cornea, stromal component of the flap, and residual stromal bed, in the case of monocular diplopia
with a topographic diagnosis of “decentered ablation.” The residual stromal thickness minimum is 223 microns
(row 3, column 2). Inspection of the epithelial thickness profile (row 1, column 1) demonstrates the error
introduced by epithelial compensation, if one were to attempt topography guided or wave front guided ablation
to correct the optical defect. B-scan imaging (Figure 4.13) confirms that laser ablation would be a less optimal
management strategy in this case, in which there is extreme flap bunching as a result of inadequate distension.
(see color image)
P.135
Figure 4.15. Horizontal VHF digital ultrasound B-scan through the visual axis of a cornea 3 months after LASIK
for -10.00, which produced a gross overcorrection to +4.00 D. The geometrically correct image is shown above
(A) with the raw scan data axially zoomed below (B). The flap is seen to be much thinner centrally (small arrow)
than peripherally (large arrow). This unplanned, extreme midperipheral deep keratectomy explains the gross
overcorrection; central flattening occurred as a function of three mechanisms acting together: 1. laser ablation.
2. RK-like peripheral bulging with central flattening. 3. Roberts' like peripheral corneal thickening as a result of
lamellar relaxation producing increased central flattening. Clearly, relifting of this flap and further midperipheral
ablation (as would have been carried out before this diagnostic test) would risk exacerbating the hyperopic shift
by further unpredicted biomechanical changes in the cornea.
Biomechanical Changes in the Cornea: Elastic Versus Ectatic

Elastic changes in the cornea are comprised of forward or backward bending of the central cornea as a result of surgically induced lamellar structural changes, the
influence of intraocular pressure, and other external forces to the cornea, but by definition should not be described as ectasia. Ectasia, as it relates to lamellar
refractive surgery (and keratoconus) should be defined according to the way it was done by the father of lamellar corneal refractive surgery: José Ignacio
Barraquer-Moner. Barraquer (31) defined ectasia as a progressive deformation of the cornea in which there is progressive corneal steepening and thinning. As
such, it describes a plastic and/or viscoelastic deformation.
In LASIK, it is important to determine the true anatomic diagnosis for secondary ametropia (i.e., epithelial or biomechanical corneal changes), so that accurate
cognitive surgical planning for enhancement surgery can be accomplished. Given the lack of consistency in flap thickness, when considering correction of
secondary ametropia (enhancement surgery), a knowledge of the whole residual stroma bed is paramount for maximum safety. Currently, only the central residual
stromal thickness (RST) is determined either from the assumed presurgical parameters, or it is measured manually intraoperatively, using a handheld ultrasonic
pachymeter.
In the following example, a 30-year-old woman underwent LASIK OS, with the Nidek EC5000 and the Moria LSK-One “130”-head microkeratome. The preoperative
corneal thickness by Orbscan was 555 µm. The predicted ablation depth for correction of -6.25 - 1.25 × 175 (BSCVA 20/20) in a 6.5-mm zone was 101 µm, and the
flap thickness used for predicting the residual stromal thickness was 160 µm. Therefore, the predicted residual stromal thickness was 293 µm (555 - 101 - 160 =
294). One week after treatment, the refraction was -1.00
P.136
- 1.00 × 170 and at 3 months, the refraction was -0.50 - 1.75 × 175 (BSCVA 20/20). Note that there was no correction of the original on-axis cylinder, and an
undercorrection of myopia. Given that there were 43 µm of residual stroma predicted still left over the 250-µm limit, an enhancement was performed lifting the
original flap 3 months later, using the Bausch & Lomb 217C (predicting a removal of a further 42 µm, bringing the predicted residual stromal thickness to a level of
251 µm). Intraoperative bed pachymetry (Sonogage II handheld ultrasound pachymeter) measured 305 µm in the central bed under the flap before ablation of 42
µm, as predicted by the laser. Six months following this enhancement refraction was +2.75 - 2.50 × 90 (BSCVA 20/20), and the patient was complaining of double
vision in that eye (even with best spectacle correction). Greater than 100% overcorrection of on-axis cylinder occurred. Given that no further tissue was left under
the flap for further treatment, the decision was made to enhance as a photorefractive keratectomy (PRK) over the original flap, 7 months later, by what is termed
“advanced surface ablation”—removal of the epithelium using 20% ethanol and ablation with the Bausch & Lomb 217C. The treatment performed was a positive
cylinder ablation of plano +2.50 × 180. Although the cornea remained clear, and no haze developed, the refraction 6 months after this second enhancement was
+2.00 - 2.75 × 175. Again, an overcorrection greater than 100% of the cylinder was made and little change occurred in the spherical equivalent.
Figure 4.16. Multiple Orbscan back surface best-fit-sphere plots before LASIK (top row), 3 months
postoperatively (middle row), and 6 months after enhancement (bottom row). The best-fit-sphere radius for all
time stages are user set to the radius of the central 4-mm zone of the back surface before surgery (6.37 mm) to
show back surface shifts in curvature relative to the preoperative state. The first column shows the 3-D maps, in
2-D, whereas the second column shows horizontal cross-sectional representation of the best-fit to the initial
6.37-mm radius. At 3 months postoperative the back surface was seen to have decreased a little in radius
relative to the preoperative state (relative bulging), and, after enhancement, this back surface bowing is
considerably increased. (see color image)
The patient was evaluated with very-high-frequency ultrasound. Electronic files for all Orbscan examinations performed before and after surgery from her original
surgeon were obtained. After importing these into our Orbscan workstation, we displayed the best-fit sphere (BSF) back surface maps for preoperative, 3 months
postoperative (pre-first enhancement), and 5 months postenhancement. All three back surface best-fit spherical maps were exhibited side-by-side for comparison.
We set the fit sphere curvature for all three time points to that of the preoperative back surface best-fit sphere curvature of 6.37 mm, as shown in Figure 4.16. The
Orbscan back surface maps demonstrate how, with each treatment the central radius of curvature of the back surface decreased incrementally. But why would this
be occurring, despite that more than 250 µm were predicted (and confirmed by intraoperative measurement) to have been left under the flap?
VHF digital ultrasound B-scanning was performed using the arc scan prototype. Figure 4.17 shows the 3-D thickness map of the residual stromal layer
demonstrating a thickness centrally of approximately 270 µm, which was more than estimated by calculation based on preoperative parameters, but close to what
was predicted by the intraoperative pachymetry (305 - 42 = 263 µm). However, the RST 1.5-mm nasal to the center was found to be only 216 µm (the first
enhancement; under the flap would have reduced the thickness of the RST at this location by approximately 26 µm, therefore the RST here before the first
enhancement was approximately 242 µm).
The cause of this RST bed asymmetry is evident from inspection of the thickness map of the stromal component of the flap; the stromal component of the flap
1.5-mm nasally was 139 µm compared to only 65 µm centrally. The original flap thickness nasally and centrally would have been 189 µm (139 + 50) and 115 µm
(65 + 50), respectively. (The PRK over the flap would not have removed some tissue from the nasal stromal component of the flap because it was a positive
cylinder ablation.)
The asymmetric flap thickness, with an RST below 250 µm in the nasal portion of the cornea may explain why this patient's cornea was behaving unpredictably on
repeated enhancements. We have shown that significant and measurable biomechanical changes occur in the cornea after LASIK, with a residual stromal
thickness below 290 µm (26,27), and this asymmetric RST bed may be responsible for unpredicted biomechanical shifts, as evidenced by the serial back surface
Orbscan exams shown in Figure 4.16.
P.137
Figure 4.17. Three-dimensional pachymetric mapping of the residual stromal layer under the flap and the
stromal component of the flap (i.e., excluding epithelium) created with a nasal hinge with the Moria LSK-One
using the 130-head. The central thickness of the stromal component of the flap is 65 µm, corresponding to an
original flap thickness of approximately 115 µm (65+50). However, 2-mm nasal to the center of the flap the
stromal component thickness was 139 µm, corresponding to an original flap thickness of 189 µm (139+50). This
nasal increased thickness was responsible for an unpredictably low residual stromal thickness (above), which
led initially to an unpredicted increase in cylinder in the horizontal plane and subsequent unpredicted
biomechanically based refractive shifts on repeated enhancement surgery. (see color image)
Had VHF digital ultrasound scanning been performed before the first enhancement, it would have been evident that the RST was already below 250 µm nasal to
the center after the first treatment. This may have alerted the surgeon to not lift the original flap and remove further tissue from under it, and may have avoided the
eventual induction of biomechanically mediated asymmetric astigmatism with monocular diplopia.
In another example, a 33-year-old woman underwent simultaneous bilateral LASIK (OS first, OD second) for -6.00 - 0.50 × 115 (20/15) OD and -6.00 - 0.50 × 20
(20/15) OS. Treatment was carried out with the Hansatome 160-µm head and the MEL70 excimer laser, with ablation depths of 99 µm for each eye (6.5-mm fully
corrected optical zone). Preoperative Orbscan thicknesses were 542 µm OD and 540 µm OS, yielding predicted RST of 283 µm OD and 281 µm OS. One month
UCVA was 20/15 -2 in both eyes. Refractions were plano -0.50 × 125 OD and -0.50 D OS 9 months postoperative; UCVA was 20/25 OD and 20/30 OS; refraction
was -0.50 - 0.50 × 150 (20/15) OD and -0.75 - 0.25 × 145 (20/15) OS. As per enhancement protocol in our practice, she underwent Artemis scanning for
determination of the RST before enhancement. RST maps for right and left eyes are shown in Figure 4.18. The minimum RST was 278 µm OD and 221 µm OS.
Central thickness of the stromal component of the flap was 85 µm OD and 135 µm OS, corresponding to original flap thicknesses of 135 µm OD and 185 µm OS.
We reported VHF digital ultrasound pachymetric mapping of Hansatome flaps created bilaterally using the same blade. The mean (± SD) central thickness for first
eyes was 139 (± 21.3) µm and for second eyes was 122 (± 22.4) µm. Second flaps were statistically significantly thinner than first flaps (32). Therefore, it is not
surprising that the right flap was thinner than the left. However, the left (first) flap was approximately 15 µm thicker than expected (185 versus 160). Because the
RST was 60 µm thinner than expected, it can be concluded that the Orbscan preoperatively overestimated corneal thickness by approximately 45 µm (originally
predicted RST - flap thickness over 160 - RST achieved = 281 - 15 - 221). Therefore, in this case, RST monitoring by direct measurement before what appeared to
be a relatively benign enhancement, with adequate tissue reserve, resulted in removing excess tissue from the bed of a cornea already biomechanically
compromised with an RST of 221 µm. Had ablation been carried out for the enhancement under the flap, assuming a residual of 283 µm as predicted, ablation in a
7-mm zone of 20 µm would have reduced the RST to approximately 201 µm, with the surgeon assuming that there was still 263 µm of RST. Conceivably, this
would have led to further biomechanical change and a myopic shift (26), possibly a second enhancement and the risk of inducing ectasia (30), with the surgeon
believing that more than 250 µm were being left under the original flap. In fact, many current publications on biomechanical changes and ectasia in the cornea after
LASIK are based on the predicted value for the RST based on preoperative parameters (33, 34, 35). Our mathematical modeling using VHF digital
ultrasound-based direct measurement of the RST led us to determine that ectasia probably occurs, on the average, at an RST of 180 µm (30). We have studied the
relative predictability of the RST in LASIK and found that with modern pachymetry (Orbscan and handheld ultrasound
P.138
pachymetry), microkeratomes, and laser ablation depths, the RST can be predicted within a standard deviation of 30 µm (36) and hence a range of approximately ±
45 µm (95% confidence interval). Therefore, it appears prudent to stay within the Barraquer rule of aiming to leave 250 µm under a lamellar flap (31) to avoid
breaching this 180-µm limit.
Figure 4.18. Three-dimensional residual stromal thickness maps of the right and left eyes of a patient who
underwent routine LASIK with predicted residual stromal thicknesses of 283 µm OD and 281 µm OS. As seen
here, the thinnest points of the residual stromal beds were 278 µm and 221 µm on the right and left,
respectively, perhaps accounting for a slightly greater myopic undercorrection on the left as a result of a
biomechanical corneal shift. The determination of this low residual stromal thickness on routine Artemis
scanning prevented the removal of a further 20 µm, which would have resulted in an RST of 201 µm and
probably a high risk of developing ectasia. Artemis scanning for verification of residual stromal thickness is a
powerful tool for increasing the safety of enhancement surgery. (see color image)
Given our experience in the subject, our current thinking is that ectasia will occur if either less than 200 µm are left under the flap or LASIK is performed in a cornea
with undiagnosed keratoconus. Therefore, accurate biometry before LASIK, after LASIK, and before enhancement, as well as cross-interpretation with Orbscan
front and back surface shape evaluation, should protect corneas from ectasia, with the exception of those that harbor undiagnosed (or unexpressed) keratoconus.
A NOTE ON PHAKIC INTRAOCULAR LENS SURGERY

No phakic intraocular lenses (IOLs) are currently approved by the United States Food and Drug Administration (FDA), and this is almost certainly contingent (for
angle supported and posterior chamber lenses) on the lack of adequate preoperative internal ocular biometry in surgical planning. Maximizing the safety of phakic
IOLs is the honorous responsibility of surgeons who expect these devices to remain in normal eyes for several decades without causing serious side effects.
One of the most unique contributions to ophthalmology by very-high-frequency ultrasound is in the sizing of intraocular lenses (IOLs), particularly phakic IOLs.
Incorrect lens sizing or positioning can lead to long-term complications. One of the main safety hurdles encountered in anterior chamber, angle-supported phakic
IOLs implantation has been defining the correct amount of haptic force in the angle; if the haptic diameter is too large, this can lead to ischemia of the iris, causing
iris stromal scarring and pupil ovalization. If too small, the lens may become displaced in the anterior chamber, risking endothelial damage or decrease the ability to
correct astigmatism with toric lenses. Issues relating to the sizing of posterior chamber lenses also exist. If the vault of such a lens in the posterior chamber is too
large, it can lead to narrowing of the anterior chamber angle; it can also increase the chances of pigment dispersion from the pigment epithelium of the iris, with
subsequent glaucomatous consequences. If the posterior chamber phakic IOL is too small, excessive contact between it and the crystalline lens may decrease
aqueous flow and lens nutrition, as well as directly traumatize the lens surface, leading to cataract.
By providing accurate sulcus-to-sulcus and angle-to-angle measurements, the arc scan has the potential to increase the safety of both anterior and posterior
chamber phakic IOLs by improving the accuracy of lens sizing, a crucial issue for long-term safety of these devices (Figure 4.19). Until recently, surgeons have
been using the external white-to-white measurement to estimate the internal or sulcus-to-sulcus or angle-to-angle diameters
P.139
(37,38). A recent study revealed either no or insufficient statistical correlation between the external ocular measurements (including white-to-white) and the internal
angle-to-angle or sulcus-to-sulcus measurements of the eye, even if other conventional measurements (such as sphere, axial length, anterior chamber depth) were
included (7). This means that the only alternative for ensuring the greatest sizing safety in phakic IOL surgery will be to determine angle-to-angle and
sulcus-to-sulcus dimensions by direct measurement. Arc scan is the only technology available, at present, that can provide both these measurements directly, in
3-D, and under direct visualization for positional confirmation of the location from where measurement is taken. Without this feature, it would be relatively easy to
measure internal ocular dimensions in the wrong plane (Figure 4.20).
Figure 4.19. Full anterior segment horizontal VHF digital ultrasound B-scan encompassing a 15-mm wide
sector. The anterior retina can also be seen within this scan plane. The angle-to-angle and sulcus-to-sulcus
diameters are easily measured directly. Anterior chamber and posterior chamber volumes and dimensions can
be studied before insertion of phakic IOLs to predict the separation of such implants from the endothelium of the
cornea, or the crystalline lens. Predictive effects on the angle as a result of posterior chamber phakic IOLs could
also be made prospectively to improve patient safety.
Figure 4.20. Screen capture from the Artemis during an anterior segment patient exam for direct measurement
of the sulcus-to-sulcus. Real-time horizontal B-scans are displayed on the upper right panel of the screen. The
infrared simultaneous video image shows that the position of the horizontal scanning plane is not central or
axial. The zoom window (Z) of the B-scan shows a cross-sectional anterior segment representation containing
pupil borders (P) that, in the absence of positional information, could have been interpreted as an axial scan,
producing a false-low sulcus-to-sulcus diameter. Similarly, the angle-to-angle would have been underestimated
falsely. Simultaneous video control is paramount for maximizing the safety of phakic IOL sizing, because
improper localization will lead to erroneous biometry and the potential for the oversizing of phakic IOLs. (see
color image)
P.140
Improving the safety of phakic IOLs by accurate anatomic surgical planning and postoperative monitoring could position phakic IOLs as a real alternative treatment
for correcting lower refractive errors where, currently, extraocular corneal refractive surgery is the first-line approach.
CONCLUSION
Orthopaedic surgery was practiced without preoperative and postoperative anatomic imaging until the discovery of x-ray imaging in 1895, by Wilhelm Konrad
Roentgen. Perhaps layer-by-layer anatomic imaging and biometry of the cornea and anterior segment will have a similar impact on refractive surgery.
REFERENCES
2. Reinstein DZ, Silverman RH, Raevsky T, et al. A new arc-scanning very high-frequency ultrasound system for 3D pachymetric mapping of corneal
epithelium, lamellar flap and residual stromal layer in laser in situ keratomileusis. J Refract Surg 2000;16:414-430.
3. Reinstein DZ, Silverman RH, Coleman DJ. High-frequency ultrasound measurement of the thickness of the corneal epithelium. Refract Corneal Surg
1993;9:385-387.
4. Reinstein DZ, Silverman RH, Trokel SL, Coleman DJ. Corneal pachymetric topography. Ophthalmology 1994;101:432-438.
5. Reinstein DZ, Silverman RH, Rondeau MJ, et al. Epithelial and corneal thickness measurements by high-frequency ultrasound digital signal processing.
Ophthalmology 1994;101:140-146.
6. Reinstein DZ, Silverman RH, Sutton HF, Coleman DJ. Very high-frequency ultrasound corneal analysis identifies anatomic correlates of optical
complications of lamellar refractive surgery: anatomic diagnosis in lamellar surgery. Ophthalmology 1999;106:474-482.
7. Reinstein DZ, Silverman RH, Lloyd OH. Estimation of angle-to-angle or sulcus-to-sulcus from white-to-white and conventional ocular measurements: are
there adequate correlations for safe phakic-IOL surgery? European Society of Cataract and Refractive Surgery Annual Meeting, September 7-11, 2002; Nice,
France.
8. Reinstein DZ, Aslanides IM, Patel S, et al. Epithelial lenticular types of human cornea: classification and analysis of influence on PRK. Ophthalmology
1995;102(Suppl):156.
9. Patel S, Reinstein DZ, Silverman RH, et al. The shape of Bowman's layer in the human cornea. J Refract Surg 1998;14:636-640.
10. Reinstein DZ, Polack PJ, McCormick S, et al. High frequency ultrasound scanning of corneal scar formation in vivo. Invest Ophthalmol Vis Sci
1992;33:1233.
11. Reinstein DZ, Silverman RH, Trokel SL, et al. High-frequency ultrasound digital signal processing for biometry of the cornea in planning phototherapeutic
keratectomy [Letter] [Published erratum appears in Arch Ophthalmol 1993;111:926]. Arch Ophthalmol 1993;111(4):430-431.
12. Aslanides IM, Reinstein DZ, Silverman RH, et al. High-frequency ultrasound spectral parameter imaging of anterior corneal scars. CLAO J
1995;21:268-272.
13. Allemann N, Chamon W, Silverman RH, et al. High-frequency ultrasound quantitative analyses of corneal scarring following excimer laser keratectomy.
Arch Ophthalmol 1993;111:968-973.
14. Lazzaro DR, Aslanides IM, Belmont SC, et al. High frequency ultrasound evaluation of radial keratotomy incisions. J Cataract Refract Surg
1995;21:398-401.
15. Silverman RH, Reinstein DZ, Raevsky T, et al. Improved system for sonographic imaging and biometry of the cornea. J Ultrasound Med
1997;16:117-124.
16. Reinstein DZ, Sutton HFS, Srivannaboon S, et al. Microkeratome efficacy: 3D thickness assessment of corneal lamellar flap accuracy and reproducibility
by arc-scanning very high-frequency digital ultrasound. J Refract Surg 2005 In press.
17. Roberts C. The cornea is not a piece of plastic. J Refract Surg 2000;16(4):407-413.
18. Holland SP, Srivannaboon S, Reinstein DZ. Avoiding serious corneal complications of laser assisted in situ keratomileusis and photorefractive
keratectomy. Ophthalmology 2000;107:640-652.
19. Boscia F, La Tegola MG, Alessio G, et al. Accuracy of Orbscan optical pachymetry in corneas with haze. J Cataract Refract Surg 2002;28:253-258.
20. Prisant O, Calderon N, Chastang P, et al. Reliability of pachymetric measurements using Orbscan after excimer refractive surgery. Ophthalmology
2003;110:511-515.
21. Iskander NG, Anderson Penno E, Peters NT, et al. Accuracy of Orbscan pachymetry measurements and DHG ultrasound pachymetry in primary laser in
situ keratomileusis and LASIK enhancement procedures. J Cataract Refract Surg 2001;27:681-685.
22. Patel S, Alio JL, Perez-Santonja JJ. A model to explain the difference between changes in refraction and central ocular surface power after laser in situ
keratomileusis. J Refract Surg 2000;16:330-335.
23. Gauthier CA, Holden BA, Epstein D, et al. Factors affecting epithelial hyperplasia after photorefractive keratectomy [See comments]. J Cataract Refract
Surg 1997;23:1042-1050.
24. Lohmann CP, Reischl U, Marshall J. Regression and epithelial hyperplasia after myopic photorefractive keratectomy in a human cornea. J Cataract
Refract Surg 1999;25:712-715.
25. Srivannaboon S, Reinstein DZ, Sutton HF, et al. Effect of epithelial changes on refractive outcome in LASIK. Invest Ophthalmol Vis Sci 1999;40:S896.
26. Reinstein DZ, Srivannaboon S, Silverman RH, et al. Limits of wave front customized ablation: biomechanical and epithelial factors. Invest Ophthalmol Vis
Sci 2002;43:E-Abstract 3942.
27. Reinstein DZ, Srivannaboon S, Silverman RH, et al. The accuracy of routine LASIK: isolation of biomechanical and epithelial factors. Invest Ophthalmol
Vis Sci 2000:S318.
28. Patel S, Marshall J, Fitzke FW. Refractive index of the human corneal epithelium and stroma. J Refract Surg 1995;11:100-105.
29. Reinstein DZ, Aslanides IM, Silverman RH, et al. Epithelial and corneal 3D ultrasound pachymetric topography post excimer laser surgery. Invest
Ophthalmol Vis Sci 1994;35:1739.
30. Reinstein DZ, Srivannaboon S, Sutton HF, et al. Risk of ectasia in LASIK: revised safety criteria. Invest Ophthalmol Vis Sci 1999;40(Suppl):S403.
P.141
31. Barraquer JI. Queratomileusis y queratofakia. Bogota: Instituto Barraquer de America, 1980.
32. Srivannaboon S, Reinstein DZ, Sutton HS, et al. Hansatome flap consistency analysis by 3D VHF ultrasound pachymetric topography. Invest Ophthalmol
Vis Sci 1999;40:S327.
33. Seitz B, Torres F, Langenbucher A, et al. Posterior corneal curvature changes after myopic laser in situ keratomileusis. Ophthalmology
2001;108:666-672; discussion 73.
34. Pallikaris IG, Kymionis GD, Astyrakakis NI. Corneal ectasia induced by laser in situ keratomileusis. J Cataract Refract Surg 2001;27:1796-1802.
35. Seiler T, Koufala K, Richter G. Iatrogenic keratectasia after laser in situ keratomileusis. J Refract Surg 1998;14:312-317.
36. Reinstein DZ, Cremonesi E. Ectasia in routine LASIK: occurrence rate is reduced by one third when consistently using a thinner flap. Invest Ophthalmol
Vis Sci 2001;42:S725.
37. Zaldivar R, Oscherow S, Ricur G. The STAAR posterior chamber phakic intraocular lens. Int Ophthalmol Clin 2000;40:237-244.
38. Baikoff G. Intraocular phakic implants in the anterior chamber. Int Ophthalmol Clin 2000;40:223-235.
> Table of Contents > 5 - Orbital Diagnosis
5
Orbital Diagnosis
Evaluation of the orbit with ultrasound is part of a continuum of imaging techniques whose other primary components are magnetic resonance (MR), computed
tomography (CT), and routine x-rays. In the hands of the skilled practitioner, orbital ultrasound using multiple frequencies (including 20-MHz B-scan of the optic
nerve) and color Doppler imaging can provide most, if not all, diagnostic information available with other imaging techniques and may also be complementary to
other imaging modalities. For reasons as diverse as reimbursement rates and the relative steepness of the learning curve with ultrasound, the bulk of orbital
imaging is now performed with readily accessible CT and MR (Figures 5.1 and 5.2).
Even so, ultrasound remains a highly useful adjunct, and its strength within the diagnostic methodology is a result of its cost-effectiveness and ease of repeatability.
The higher spatial resolution of ultrasound is also particularly sensitive to detecting early and low grade inflammatory changes in the sclera, optic nerve, and
extraocular muscles. Ultrasound complements MR for monitoring treatment of Graves disease as well as imaging periocular drug delivery. Cone occupying lesions,
such as hemangiomas and lymphangiomas, can be imaged and diagnosed with ultrasound as well (Figure 5.3). But most extraocular tumors, such as meningiomas
or other solid tumors, are best imaged with MR. MR imaging of the orbit is optimal, if it includes the use of an appropriate surface coil, thin slices, contrast, and fat
suppression.
Metallic or questionable orbital foreign bodies are best detected with routine x-rays or helical CT (e.g., presence of metallic fragments, such as shotgun pellets, to
determine the number of foreign bodies [Figure 5.4] or unusual metal objects, such as nails [Figure 5.5], paper clips [Figure 5.6], or aerials [Figure 5.7]) because
MR should not be used when any possibility of ferrous foreign body exists. CT should be limited because of radiation hazards in children and young adults.
Ultrasound is useful in detecting low density foreign bodies, such as wood or some plastics in the proximal orbit, particularly if they are surrounded by exudates.
Color flow Doppler ultrasound remains in limited use by ophthalmologists outside of the hospital setting, where higher frequency small parts probes for general
radiology ultrasound units are available, and the test may be reimbursable through the radiology department. Flow studies are, however, helpful in quantifying blood
flow in the optic nerve and in differentiating suspicious masses by flow characteristics.
All imaging techniques continue to improve with increasing computer power and other technologic advances, particularly in ultrasound, where higher frequency
systems and digital analytic techniques permit improved higher resolution imaging of the posterior sclera and the optic nerve.
This chapter will describe ultrasonic examination methods and present a perspective in the complementary use of ultrasound and other imaging methods for orbital
evaluation.
TECHNIQUES
Ultrasound aids in the diagnosis of orbital abnormalities by providing information that may not be obtainable by any other examination technique and is noninvasive
and easy to perform in a serial manner. As discussed in Chapter 3, we prefer to use combined A- , B- , and M-scan ultrasonic techniques for evaluation of both the
globe and orbit. A water-bath standoff with the lids open permits maximum penetration of sound into the orbit.
P.144
Figure 5.1 A surface-coil T1-weighted image of an orbit and surrounding tissue shows excellent demonstration
of the globe, nerve, orbital muscles, and orbital fat.
In orbital diagnosis, the topographic outlining capabilities of the B-scan are used for localizing and delineating abnormalities. In our combined technique, after initial
localization with B-scan has been achieved, we use the A-scan for tissue evaluation and use A- or M-scan for determination of the pulsatile vascular properties of
tissues. Orientation difficulties (as a result of the lack of distinctive anatomic landmarks in the orbit) preclude total reliance on an A-scan-only method. Byrne and
Green (1) and DiBernardo and Schachat (2) have provided excellent and comprehensive reviews of
P.145
A-scan and combined A-scan/B-scan methods for evaluating the orbit. Their techniques, following concepts by Ossoinig (3), place greater emphasis on the A-scan
than do our techniques.
Figure 5.2 A CT scan of an orbit with a foreign body, demonstrating its use for suspected foreign bodies and its
excellent delineation of bone.
Figure 5.3 A 10-MHz A- and B-scan ultrasonogram of a globe and orbit showing the well-encapsulated lesion.
A-scan shows internal acoustic reflections of a typical hemangioma.
Figure 5.4 Plain film x-ray of shotgun pellets around the face and orbit. The orbital pellet had traversed the
globe. Shotgun pellets may be steel, lead, or composite material. Magnetic testing can be useful, using either a
similar shell supplied by the patient or family or, at surgery, using a pellet from a superficial nonocular site.
(Courtesy of Murk-Hein Heinemann, MD, New York, NY.)
Figure 5.5 Plain film x-ray of a nail that penetrated the orbit and cranial vault but did not puncture the globe.
(Courtesy of Alan Maberley, MD, Vancouver, BC.)
Figure 5.6 Plain film x-ray of an intraocular broken paper clip. The shape is outlined better than with any other
imaging modality, but the globe relationships are lacking and can be seen with ultrasound. (Courtesy of Gwen
Sterns, MD, Rochester, NY.)
Orbital evaluation with ultrasound is typically performed at 10 MHz, with higher frequencies limited to the posterior sclera, optic nerve, and proximal orbital region.
The ultrasonic orbital evaluation consists of serial tomographic sections, using both static and kinetic A-and B-scanning. In general, horizontal scans are made
serially across the eye and orbit at roughly 2-mm intervals, with the eye in the six major gaze positions. Ossoinig (3) has emphasized the importance of additional
meridional scanning of the orbit to portray structures partially obscured by the orbital rim. Lower frequency transducers (such as 5 MHz) may be used to outline the
posterior extent and apex of the orbit. A lower frequency may also be useful, if ocular pathology prevents adequate penetration at the normal examining frequency
of 10 MHz. However, these transducers are not generally
P.146
available in ophthalmic ultrasound equipment and have low resolution.

Figure 5.7 Plain film x-ray of a van radio antenna that penetrated the orbit and cranium, demonstrating the
value of x-rays in outlining the shape of the foreign body. The tip of the antenna was removed through a cranial
burr hole before the antenna was withdrawn. (Courtesy of Stephen Trokel, MD, New York, NY.)
Kinetic scanning, that is, having the patient move his or her eye from side to side and up and down, while fast sector scans are performed, is important in orbital
diagnosis. Kinetic scanning helps indicate the degree of adherence of a mass to the mobile anatomic features of the orbit, such as the optic nerve, and to the fixed
structures, such as the orbital wall.
Comparative evaluation of the patient's other orbit is not usually performed, because variations from the normal orbital pattern in most disease states are easily
discerned. However, in patients with subtle pathologic changes, such as disease of the optic nerve, a comparative evaluation of the companion orbit may aid in
differentiation.
When a lesion is palpable (e.g., a cyst or mass in the adnexa) it is useful to use direct contact A- and B-scan techniques. The contact probe can be more easily
manipulated to confirm the extent and location of the mass. In addition, the contact A- or B-scan can be used to identify a lesion that may be concealed by the bony
overhang of the superior orbital rim. These techniques are particularly useful with lacrimal gland tumors. A contact probe to compress various orbital lesions has
been described by Ossoinig (3) and is useful in characterizing the cystic or solid components of orbital masses.
A thorough knowledge of orbital anatomy and the pathologic situations likely to arise in the orbit greatly enhances the ability to interpret information that can be
obtained from the ultrasonic evaluation. For these reasons, ultrasonic orbital examination is best performed by an ophthalmologist or a technician who is well
trained for ophthalmic ultrasonography.
DIAGNOSTIC PARAMETERS
The ultrasound echo patterns seen in the orbital plane arise from the acoustic impedance mismatch between adjacent tissues. Measuring acoustic characteristics
of orbital tissues, Buschmann (4) reported sound velocities of 1,462 meters per second for fat compared with 1,615 meters per second for optic nerve and 1,631
meters per second for muscle. The difference in velocity between these tissue components is the acoustic impedance mismatch that causes partial reflection of an
ultrasonic wave as it meets the tissue interfaces.
Within a heterogeneous tissue, such as retrobulbar fat, are many smaller tissue elements, including vessels, nerves, and fat globules with many fibrous septa.
These multiple tissue interfaces produce individual echoes and result in a nearly uniform confluence of echoes representing the fat pad. The relatively uniform and
organized overall tissue structure of the optic nerve and extraocular muscles has markedly less internal impedance mismatches and produces only low-amplitude
echoes within the tissue substance. In addition, the optic nerve and, to a lesser extent, extraocular muscles and major structural components are organized in
tissue planes parallel to the ultrasonic beam, minimizing echoes that are detectable along the transmitter-receiver beam path. The same principle of reflections
from internal structural elements of a tissue applies to tumors and other abnormalities and is a major criterion for differentiation of tissue types. Orientation to the
beam is, thus, the major reason that amplitude of the tissue impedance differences alone cannot be absolutely reliable from the A-scan.
TYPES OF DIAGNOSTIC INFORMATION

Clinicians have long relied on radiographic techniques for orbital evaluation. Plain films and computed tomography depict bony abnormalities well—fractures,
erosion, or hyperostosis. Vascular contrast studies demonstrate arteriovenous lesions and intracranial abnormalities causing exophthalmos. Some indications of
orbital soft tissue lesions may also be gained from these studies, but the findings are often not definitive (5). Direct injection of contrast material into the orbit
(contrast orbitography) has been used but has fallen out of favor because of morbidity and diagnostic unreliability. The most common diagnostic tests performed
today for orbital imaging are plain film radiography, computed tomography, magnetic resonance imaging, and ultrasonography. Additional tests include plain film
angiography and venography, MR angiography (MRA), color Doppler imaging (CDI), and dacryocystography. Radiography methods
P.147
and computed tomography subject patients to the known risks of radiation exposure. Magnetic resonance imaging is contraindicated in the case of ferromagnetic
foreign objects in the body. CT and MR provide similar soft tissue details in the orbit. CT provides clearer bone details and is sensitive to calcification. MR is better
at defining the globe and intraocular structures as well as the optic nerve along its path.
MR is the best overall way of imaging abnormalities of the orbit. Definition and resolution are superb, but the test is still far more expensive than ultrasound.
Ultrasonography, in contrast, provides unique high-resolution information regarding tumors and inflammatory orbital changes using an inexpensive, reliable, and
easily repeatable technique. Bony changes and vascular abnormalities, however, are not well demonstrated, although soft tissue is generally well seen.
Ultrasonography can serve as an invaluable complement to the usual radiographic, CT, and MR studies, and can often demonstrate an abnormality (because of its
sensitivity and high resolution) when all other tests are negative.
Medical and surgical therapies are aided considerably by ultrasonographic findings. For instance, the effect of steroid administration in presumed pseudotumors
can be monitored by following the inflammatory signs ultrasonically. Surgical approaches to a tumor, for example, plaque therapy, can be planned with full
knowledge of the location, size, extent, tissue composition, and circumscribed or invasive character of the tumor. In general, ultrasonography, followed by contrast
MR or CT, should be the first test used for orbital evaluation. Ultrasound is a sensitive test and rarely misses any significant orbital abnormality. This inherent
sensitivity is occasionally misleading (in that inflammatory tissues can resemble neoplasms) so further imaging with MR or CT is desirable, if the ultrasonographic
findings indicate a pathologic situation.
INDICATIONS FOR ORBITAL ULTRASOUND

The indications for orbital ultrasonography are summarized in Table 5.1. Ultrasound is specifically useful in documenting clinically, evaluable pathologic states such
as myositis, Graves disease, or optic neuropathy. In general, ultrasound is indicated when orbital pathology is suspected. Ultrasound findings may aid in the
management and treatment of these orbital pathologic conditions, particularly inflammatory conditions.
TABLE 5.1 Indications for Ocular Ultrasonography
Unilateral or bilateral exophthalmos
Retinal striae
Unexplained optic atrophy
Papilledema without evident cause
Suspected orbital foreign body
B-SCAN
HORIZONTAL SCAN PLANE THROUGH THE OPTIC NERVE

With the ultrasonic scan plane passing through the optic nerve, the normal retrobulbar echo pattern is a W-shaped, acoustically opaque (white) area (Figure 5.8).
This opaque W-shaped area is bounded anteriorly by the globe and is indented posteriorly by an acoustically empty (black) notch that widens toward the orbital
apex. This notch, or triangle, is formed by the optic nerve and associated structures.
The source of the echoes that give rise to the relatively uniform, isoechoic W-shaped retrobulbar echo pattern is not definitely known. Purnell (6) postulated that
orbital fat lobules provide the major source of these echoes, and his view is generally accepted. Acoustic discontinuities occur throughout this loculated tissue
between intracellular lipids and cell membranes and between cell membranes and loose connective tissue septa. (Similar echo patterns can be produced
experimentally in fine-mesh silicone sponge.) M-scans of the orbit have shown marked pulsatile vascular activity, and acoustic discontinuities in the blood vessel
network in the muscle cone probably contribute to this echo pattern.
Figure 5.8 A typical 10-MHz horizontal B-scan of the eye and orbit through the lens and the optic nerve
demonstrating the typical W-shaped pattern of the orbital fat where the nerve is located (see also DVD).
P.148
In all normal orbits, the optic nerve consistently appears as an acoustically empty triangular notch in the retrobulbar fat pattern. The optic nerve appears
acoustically hypoechoic because of homogeneous tissue structure and because the nerve fibers, septa, and meninges usually lie parallel to the examining
ultrasonic beam, thus causing no reflections. The anterior angle of the optic nerve triangle or notch is usually less than 90 degrees (with a range of normally 40 to
70 degrees). Widening or any rounding of this angle may be an indication of pathologic enlargement of the optic nerve or its sheaths.
HORIZONTAL SCAN PLANE ABOVE OR BELOW THE OPTIC NERVE

When scans of the orbit are made inferior or superior to the optic nerve, the acoustically hypoechoic (black) optic nerve triangle is, of course, absent. The
retrobulbar pattern appears as a uniform, acoustically isoechoic (white) crescent lying immediately posterior to the globe (Figure 5.9). This crescent becomes
progressively wider as the center of the orbit is approached.
VERTICAL AND MERIDIONAL SCAN PLANES

Although horizontal scan planes are generally used, performing both vertical and meridional scans of the orbit is important, because areas of pathology that might
be missed in serial horizontal scans may be demonstrated more readily. The vertical scans closely resemble horizontal scans. On an axial scan, the optic nerve
notch is demonstrable, and orbital structures can be accentuated by having the patient look in different fields of gaze.
Figure 5.9 A horizontal 10-MHz B-scan directed above the nerve, giving the typical cup-shaped orbital fat
outline. The orbital walls (ow) are shown but the necessary high gain blurs the muscle (m) boundaries.
VARIATIONS WITH POSITIONS OF GAZE

The normal retrobulbar fat pattern has been described, with the eye in a straight-ahead direction of gaze. Gaze to the far right or left results in a foreshortening of
the retrobulbar pattern (decreasing its thickness) and bending of the optic nerve pattern toward the direction of gaze (shown dynamically on DVD). The movement
of the optic nerve with changing positions is diagnostically important. If a mass lesion is found in the orbit, a fast sector scan performed as the patient moves his or
her eye may demonstrate the relationship of the mass to the mobile optic nerve.
VARIATIONS WITH AGE

Age has relatively little effect on the B-scan ultrasonographic orbital pattern. The retrobulbar pattern in infants is, of course, smaller than in adults but is identical in
outline. Our impression is that the retrobulbar fat pattern in infants and children usually appears denser and more uniformly white than in adults, possibly indicating
fat distribution in smaller micelles.
EXTRAOCULAR MUSCLES AND ORBITAL WALL

The outer limit of the acoustically opaque (white) W-shaped retrobulbar pattern is formed by the rectus muscles and intermuscular septa. In B-scans made exactly
in the horizontal meridian, a rectus muscle outline can often be seen and traced forward to the globe at its insertion.
The ultrasonic patterns formed by the normal orbital walls in different horizontal scan planes are shown for reference in Figure 5.10.
A small portion of one orbital wall (usually the more perpendicular accessible lateral orbital wall) is usually seen on B-scan ultrasonograms, but, if a large portion of
the orbital wall or both medial and lateral walls are seen, edema or inflammation of the ocular muscles or periorbital tissue is suggested. The orbital apex is rarely
seen in clinical ultrasonograms but can be approximated by tracing the optic nerve and the rectus muscle/orbital wall echoes to their juncture.
Demonstration of the orbital wall is not a dependable ultrasonographic finding but is a feature dependent on the transducer frequency (more likely to be seen with
the lower-frequency transducers, i.e., 5 MHz); depth of ultrasonic beam penetration, for example, focal zone of the transducer; and receiver gain. The following are
imperative: (a) to obtain familiarity with the area of orbital wall seen in normal orbits with a given instrument, (b) to use a given transducer, and (c) to calibrate the
receiver gain before conclusions can be made, with regard to abnormal prominence of the orbital wall on one or both sides in a scan of a possibly abnormal orbit.
P.149
Figure 5.10 Ultrasonic sections taken horizontally through a demonstration skull to illustrate the typical acoustic
appearance of the orbital walls.
FREQUENCY-RELATED VARIATION
The orbital ultrasonographic appearance varies according to the transducer frequency selected. Figure 5.11 demonstrates a comparison of the normal orbital
patterns found, using 10- and 20-MHz transducers. As a rule, the lower the transducer frequency, the deeper the orbital penetration achieved.
ARTIFACTS ENCOUNTERED IN ORBITAL ULTRASONOGRAPHY

Numerous artifacts may occur during B-scan ultrasonography of the orbit, and knowledge of these possible artifacts will help avoid erroneous interpretation of
ultrasonograms of the orbit. As discussed in Chapter 3, artifacts may be classified into two groups: (a) reduplication artifacts and (b) absorption defects.
REDUPLICATION ARTIFACTS
These artifactual echoes (also known as axial multiple echoes) occur commonly and usually appear in the central orbit along the axis of the cornea and lens, often
in the region of the optic nerve triangle. They represent a “second bounce” of echoes, usually from anteriorly located ocular surfaces. A metal lid speculum can be
another common cause of reduplication echo artifacts, especially with immersion ultrasound. These echoes can be distinguished from “real” echoes by moving the
transducer either away from or toward the eye in the immersion system. This causes movement of the reduplication echo relative to the tissue, allowing them to be
readily identified (Figure 5.11). In CDI, reduplication effects from foreign bodies or other strong, reflective surfaces can cause the so-called twinkle artifact, where
false flow is seen (7).
ABSORPTION DEFECTS
Absorption of ultrasound energy by structures located in or in front of the eye may cause abnormal orbital
P.150
ultrasonic patterns. Because these absorption defects, or “shadows,” in the retrobulbar fat pattern may simulate orbital tumors, one must be aware of their
existence. Ocular tumors (Figure 5.12) or dense calcified lenses (Chapter 3, Figure 3.23) are commonly encountered ocular causes of such orbital defects.
Figure 5.11 A comparison of the orbital fat patterns with transducer center frequency. A: 20-MHz scan of the
normal orbit, with high resolution but inferior penetration. B and C: 10-MHz scan of the normal orbit,
demonstrating good resolution and good penetration. D: 7.5-MHz linear array scan of normal orbit
demonstrating poor resolution, but excellent penetration with angular resolution. Right: 7.5-MHz CDI scan of
normal orbit showing retinal and optic nerve flow. (see color image)
When an abnormal orbital B-scan pattern is encountered, the above artifacts should be ruled out. Recognition is aided by careful monitoring of the A-scan,
permitting detection of many electronic artifacts, movement of the transducer toward or away from the eye if a reduplication echo is suspected, and analysis of the
ocular B-scan for structures that may cause absorption defects in the orbital pattern.
Figure 5.12 A 10-MHz B-scan of a choroidal osteoma (arrow) showing an orbital shadow or artifact (A) as a
result of reflection from calcification.
GENERAL CLASSIFICATION OF ORBITAL ABNORMALITY

Ultrasonically, orbital abnormalities can be classified into structural anomalies, mass lesions, inflammatory or congestive changes, or foreign bodies. Each of these
categories has several well-defined subdivisions. A-mode findings in orbital pathology have been described extensively by Ossoinig (8) and Byrne and Green (1).
B-scan patterns were classified by Purnell (6) and elaborated on by Coleman et al. (9, 10, 11, 12, 13, 14, 15, 16). The flow diagram proposed by Coleman (17)
(Figure 5.13) is useful for the examiner in approaching an unknown orbital problem.
Pseudoproptosis of an eye may be accounted for by either a large globe or a shallow bony orbit (Figure 5.14). A-scan measurement of globe diameters will
establish any significant difference between the eyes. A posterior staphyloma may be completely outlined with B-scan. With progressively increasing
exophthalmos, even with a unilaterally large globe, complete ultrasonography of orbital structures should always be done to evaluate coincident disease.
MASS LESIONS
A distinct distortion of the retrobulbar fat, optic nerve, or rectus muscles by any sort of abnormal contour ultrasonically indicates a mass lesion in the orbit. Purnell
(6) first described ultrasound B-scan patterns of the orbit that we have reclassified into four general diagnostic patterns
P.151
of mass lesions that are identifiable with B-scan ultrasonography: cystic, solid, angiomatous, and infiltrative (Figure 5.15). Several features of the abnormal area
can be used to categorize the lesion, including its contour, sound transmission, internal echoes, and location (Figure 5.16). Close correlation exists between the
ultrasonographic findings and the morphologic and histologic characteristics of mass lesions.
Figure 5.13 Ultrasonic characterization of patients referred for orbital examination. Schematic flow chart of
orbital diagnosis, progressing from distinction of normal from abnormal orbits to differentiation of abnormalities
into tumor types or inflammatory tissue, based on criteria of transmission and morphology.
For example, the contour of an orbital tumor may be smoothly rounded, sharply defined, and compressing adjacent normal structures. If the lesion has good sound
transmission, its posterior wall will also be clearly evident. Internally, the lesion may be devoid of echoes, indicating no significant tissue interfaces within the lesion.
These findings are characteristics of a fluid-filled cystic lesion, such as mucocele or dermoid tumor.
In other cases, a mass lesion may have a contour similar to a cystic lesion. However, its posterior boundary may be indistinct because of poor sound transmission
through it, indicating a solid tumor. Low-amplitude echoes within the substance of a tumor indicate minor tissue interfaces, characteristic of a homogeneous solid
tumor. These findings indicate a well-circumscribed solid tumor, for example, neurogenic tumor, lacrimal gland tumor, and some metastatic tumors. Location of the
lesion within the orbit gives further clues to identification. A lesion of this type within the muscle cone and involving the optic nerve is
P.152
probably one of the neurogenic tumors (glioma or meningioma). A similar lesion located in the upper temporal aspect of the orbit is more likely to be a lacrimal
gland tumor.
Figure 5.14 A myopic globe can produce pseudoproptosis and a slightly reduced orbital volume, as
demonstrated in this 10-MHz B-scan. The globe outline can also be seen to be staphylomatous.
Figure 5.15 Schematic presentation of the division of orbital tumor patterns based on their outline and
transmission properties.
Figure 5.16 The location of a tumor in the orbit is an important aspect of acoustic evaluation. Tumors within the
muscle cone are particularly susceptible to ultrasonic detection.
An orbital tumor with an irregular contour suggests a different group of tumors. Angiomatous tumors show fingerlike protrusions extending into the orbital fat
pattern, usually with a larger mass more posteriorly. The tumor contour may or may not be well defined. The internal structure of angiomatous tumors presents
many dense acoustic interfaces from vessel walls and blood-filled spaces comprising the tumor. Ultrasonically, this heterogeneous interior appears as multiple,
irregular, high-amplitude echoes throughout the mass, with little sound attenuation. This irregular contour and heterogeneous internal structure are distinctly
different from the cystic and solid tumor patterns described previously.
Another irregular orbital tumor, having a more solid character, indicates a solid infiltrative tumor. Although the anterior lesion outline is jagged, it is also sharply
defined and distinct. As with other solid tumors, low-amplitude internal echoes are present and sound attenuation is marked, making the posterior tumor margins
indistinct. This pattern is associated with infiltrative tumors, particularly lymphomas and sarcomas. Metastatic tumors may also show these characteristics.
Idiopathic granulomas, or pseudotumors, often mimic invasive, solid tumors clinically and ultrasonically. Orbital hematoma may also appear ultrasonically as an
infiltrative mass lesion, although it is nonneoplastic.
A history of trauma can, occasionally, be misleading, because some tumors may bleed with minimal trauma. A contrast MR should be considered for more
definitive imaging of most of these lesions. Follow-up serial ultrasonography can be a useful way to document resolution of the lesion, in the case of hematoma.
In our experience, the infiltrative tumor category is the most difficult to identify with certainty, using ultrasound, because inflammatory changes in the orbit, as
described later, may produce a similar appearance, especially when edema involves the optic nerve and the sub-Tenon's space. Scarring from previous orbital
surgery may also be confused with this category.
Classifying orbital tumors into the four categories noted earlier is not always possible, because tumors do not have a consistent presentation. However, the pattern
and absorption characteristics outlined are consistent with the tumor types noted. Other ultrasonic techniques can augment the available information in these
cases. Ballottement of the tumor with a contact probe placed against the eye will demonstrate compression of cystic or angiomatous lesions and resistance to
compression with solid tumors. M-scan ultrasound or Doppler ultrasound provides a means of studying vascular pulsations within tumors. Dynamic or kinetic sector
scanning will demonstrate motion of orbital structures with changes in gaze position and often aids in categorization.
Individual tumor types, the variations of ultrasonic patterns associated with them, and the reliability of ultrasonic diagnosis with each tumor type will be discussed
more fully as separate topics.
INFLAMMATORY AND CONGESTIVE CHANGES

In contrast to orbital mass lesions, inflammatory and congestive processes tend to involve the normal tissues that are present in the orbit, causing subtle changes
in them. Inflammatory changes may be classified, ultrasonically, as diffuse or as localized to a particular area or tissue in the orbit, depending upon the specific
inflammatory process (Figure 5.17).
A generalized abnormal mottling of the orbital fat pad is indicative of a diffuse orbital inflammation, such as cellulitis. Echoes within the fat are more widely spaced
than normal and are of high amplitude, giving the fat pad a more heterogeneous appearance without circumscribed borders. A similar acoustic appearance is seen
in traumatic orbital hemorrhage. A localized area of such mottling within the fat is seen with an orbital abscess or focal granuloma. Other localized inflammatory
findings involve specific orbital structures. Expansion of the sonolucent (hypoechoic) space between the fat pad
P.153
and the orbital wall generally indicates enlargement of the extraocular muscles. The orbital wall may be accentuated in these cases, where the acoustic
transmission characteristics of extraocular muscles are modified by the disease state. These findings are particularly characteristic of thyroid-related disease or
endocrine exophthalmos, where edema and cellular infiltration of the muscles are present.
Figure 5.17 22-MHz scans of posterior pole with digital processing to enhance the boundary between the
sclera, nerve, and Tenon's capsule, demonstrating the value of higher frequency ultrasound to demonstrate
episcleritis.
Localized inflammation and edema may also involve the optic nerve or Tenon's space adjacent to the sclera. Accentuation of the optic nerve sheath without
enlargement of the nerve itself is indicative of optic neuritis or edema—a distinctly different ultrasound presentation from enlargement produced by a tumor. These
abnormalities usually resolve with regression of the inflammatory process. Although no pathologic specimens have been examined in the acute stage of retrobulbar
neuritis, the ultrasonic and MR findings are consistent with inflammatory edema of the dural sheaths surrounding the nerve. Similar optic nerve findings may be
seen with pseudotumor (idiopathic orbital inflammation), papilledema, or vascular anomalies.
The sub-Tenon's space surrounding the globe may become expanded posteriorly from edema associated with any type of orbital inflammation. Ultrasonically, this
appears as a sonolucent area adjacent to the globe wall and connecting to the optic nerve outline. Other inflammatory signs are usually present in the orbit along
with this finding. It may also accompany some orbital tumors, particularly granulomatous tumors or lymphomas. In a normal orbit, this becomes a potential space
and, as such, is not acoustically evident.
ORBITAL TUMORS
ACOUSTICALLY “CYSTIC” TUMORS

Acoustically “cystic” orbital tumors are those that have a rounded regular outline and good acoustic transmission. It should be emphasized again that acoustic
transmission is a relative phenomenon, and the ability to demonstrate transmission through orbital structures depends on a knowledge of the patterns obtained with
one's own equipment and transducers and with careful calibration of receiver gain.
Mucocele
Mucoceles are cysts lined by paranasal sinus mucous membranes. They enlarge slowly because of continued secretion and desquamation of lining cells and
frequently cause erosion of the bony walls of the sinus by pressure. They expand in the direction of least resistance, often into the orbit through the medial wall and
floor of the frontal sinus, forcing the globe in the opposite direction. Common clinical presentations are pain and periorbital swelling, diplopia, and proptosis.
Invasion of the orbit by mucoceles of the paranasal sinuses is an important cause of unilateral exophthalmos. In series compiled by radiologists, orbital mucoceles
are the most common cause of unilateral exophthalmos, with a figure as high as 15% in the series compiled by Zismor et al. (18). In series compiled by
ophthalmologists, orbital mucoceles are a less common but still significant cause of unilateral exophthalmos. In Reese's clinical series of 230 cases, 3% of
unilateral exophthalmos was a result of orbital invasion by mucoceles (19). In children, Shields et al. (20) note that mucoceles are the most common secondary
cysts that occur in children.
Ultrasonography gives accurate information as to the location and size of an orbital mucocele. Secondary changes, such as compression of retrobulbar fat and
indentation of the posterior pole of the globe, are well shown by ultrasonography (21).
Mucoceles ultrasonically appear rounded and smooth, or even spherical in contour (Figure 5.18). They are well demarcated from the surrounding normal orbital
tissues. Mucoceles demonstrate a sharply defined, rounded anterior acoustic border that indents the retrobulbar fat. They show definite acoustic hollowness, and
A-scans through them show few or no internal echoes (Figure 5.18). On B-scan, the interior of a mucocele appears as a solid, black cavity, that is, a sonolucent
space like the globe itself. Little sound energy is absorbed in these cystic structures, and, consequently, ultrasonic radiation can penetrate through the mass,
clearly outlining its posterior extent and the orbital walls and orbital apex.
Although MR diagnosis is reliable in orbital mucoceles, CT may better confirm the extent of bony erosion,
P.154
particularly in the case of paranasal sinus involvement (20). B-scan ultrasonography may add information as to the soft tissue configuration of the orbit and facilitate
the choice of surgical approach.
Figure 5.18 A 10-MHz B-scan of a mucocele of the orbit appearing as a sonolucent area, much like the globe
itself.
Dermoid Cysts
Dermoid cysts are relatively common tumors of aberrant ectodermal tissue. They usually occur in children (22). When dermoid or other epithelial cysts involve the
bony orbital wall, they may be diagnosed on x-ray, CT, or MR (23). However, even in this situation ultrasound is valuable, because it can often demonstrate the
exact limits of the orbital mass (24,25). In cases without bony involvement (up to 16% in one series) (26), ultrasound is particularly useful in diagnosing these
tumors.
Figure 5.19 10-MHz B-scan images of a cavernous hemangioma (H) of the orbit. These tumors are well
encapsulated and usually in the muscle cone. The internal A-scan echoes and MR are shown in Figures 5.20,
5.21 and 5.22.
Cavernous Hemangiomas
Hemangiomas are one of the most common of all orbital tumors. Reese (19) described the proportion of hemangiomas to all orbital masses as 12% in a clinical
study of 230 consecutive cases and as 15% in a histopathologic study of 877 cases. In a more recent series, children in a study of 1,264 patients have the
proportion of cavernous hemangiomas as 6%, the third most common behind lymphoid tumor, 11%, and pseudotumor, 11% (27). The cavernous hemangiomas of
adults are by far the most common orbital intraconal tumor (22).
The size and location of cavernous hemangiomas are well demonstrated with B-scan ultrasonography (Figure 5.19). All cavernous hemangiomas that we have
examined were located in the muscle cone.
The group of acoustic characteristics of all cavernous hemangiomas includes: (a) a rounded, regular outline; (b) a sharply defined and rounded anterior acoustic
border; (c) good demarcation from surrounding structures; (d) low-to-moderate acoustic absorption (i.e., fair-togood sound transmission); and (e) appearance
alteration with varying transducer frequencies.
Cavernous hemangiomas show a sharply defined anterior acoustic border as a result of the abrupt transition in acoustic velocities between retrobulbar fat and the
fluid-filled tumor. These tumors are ultrasonically well
P.155
demarcated from surrounding, normal orbital structures, also resulting from the marked acoustic discontinuity. Secondary changes of the globe, such as flattening
of the posterior pole, are also demonstrable.
Figure 5.20 Combined 10-MHz A- and B-scan image of the same tumor shown in Figure 5.19. A-scan shows
moderate reflectivity from the tumor.
Cavernous hemangiomas (Figure 5.20) are mainly blood-filled vessels that are characterized by good sound transmission but can contain low-amplitude internal
echoes from the vessel walls (Figure 5.21). Sound transmission is, of course, better than acoustically solid lesions with high sound absorption (such as optic nerve
tumors) but is not as high as in the completely acoustically hollow lesions with minimal sound absorption (such as mucoceles). Because sound absorption is low in
cavernous hemangiomas, ultrasound energy is able to penetrate, so that the posterior extent of the tumor and the outline of the wall and apex of the orbit are often
demonstrated (25,28).
Figure 5.21 An expanded A-scan through the cavernous hemangioma of Figures 5.19 and 5.20. Spacing of the
echo pattern suggests vascular channels. The histology of the tumor is shown at the bottom. Note the large
endothelial lined channels. (Courtesy of Herman Schubert, MD, New York, NY.)
Cavernous hemangiomas are best outlined at 10 and 15 MHz. The relatively good outlining of these tumors at 15 MHz helps differentiate them from the group of
solid, rounded orbital tumors, such as neurogenic tumors. CDI of cavernous hemangiomas may show spotty arterial slow flow within the tumor boundaries, but this
is not a diagnostic sign (25,29, 30, 31). CDI and flow measurements can also detect central retinal artery (CRA) disruption in cavernous hemangioma as well as in
other intraconal lesions. Cavernous hemangiomas are seen on MR as well as encapsulated lesions that are isointense on T1 images and hyperintense on T2
images (Figure 5.22). Rim enhancement of the lesion may also be seen. With gadolinium contrast and fat suppression, patchy or complete filling can be seen in
these tumors (32, 33, 34).
Hemangiopericytomas
Hemangiopericytomas are uncommon vascular tumors of the orbit. They contain mostly pericytes rather than endothelial cells. These tumors are blood-filled and
encapsulated, as are hemangiomas. Their ultrasonic characteristics are essentially the same as those described for cavernous hemangiomas.
P.156
Figure 5.22 Sagittal T1-weighted view shows a well-encapsulated lesion hypointense to fat, and axial MR image
with contrast shows enhancement of the cavernous hemangioma shown in Figures 5.19, 5.20 and 5.21.
(Courtesy of Michael Kasim, MD, New York, NY.)
Cystic Lymphangiomas
Lymphangiomas of the orbit are discussed in more detail in the section on angiomatous tumors. The usual presentation of lymphangioma of the orbit is diffuse;
however, some are cystic. Hemorrhage into a lymphangioma cavity may cause a hemorrhagic cyst to form. In these cases, and as discussed previously, the
ultrasonic pattern is an orbital mass lesion with a rounded outline and good sound transmission through the tumor, permitting outlining of its posterior wall (Figures
5.23 and 5.24).
ACOUSTICALLY “SOLID” TUMORS

Acoustically “solid” orbital tumors have a rounded outline, as do cystic-appearing orbital tumors; however, sound transmission is poor through these masses, and
multiple internal echoes are present.
Tumors of the Optic Nerve

Neurogenic tumors are relatively common causes of unilateral exophthalmos. Reese (19) described the proportion of neurogenic tumors to total orbital mass as
11% in a clinical study of 230 cases and as 10.6% in a histopathologic study of 877 cases. Silva (35) found 32 neurogenic tumors in a series of 300 consecutive
cases of orbital masses, a proportion of 10.9%. Shields et al. (27) found 8% neurogenic tumors in a series of 1,264 patients. This group of tumors may require a
neurosurgical approach for optimum removal and, thus, differentiation from other orbital tumors is clinically important. This differentiation is aided by B-scan
ultrasonography.
Secondary changes, such as flattening of the posterior pole of the globe and papilledema, can be shown (Figures 5.25 and 5.26) with B-scan ultrasound.
Figure 5.23 A-10 MHz combined A- and B-scan ultrasonogram of a cystic lymphangioma (bottom).
The lesion can be shown to have cystic components of varying echogenicity and internal septa (arrow).
P.157
Figure 5.24 MR image of cystic lymphoma shown in Figure 5.23. This noncontrast T1-weighted image does not
depict the lesion characteristics as well as ultrasound. (Courtesy of Kip Dolphin, MD, New York, NY.)
Figure 5.25 10-MHz B-scan of a meningioma of the optic nerve showing the sharp round border of the lesion,
expanding the optic nerve shadow. Although these tumors are solid, they absorb sound, and the posterior
border is usually absent, except along the orbital walls, as shown in Figure 5.26.
Acoustic characteristics of this group of tumors are as follows: (a) rounded configuration; (b) sharply defined anterior borders that are relatively well demarcated
from the surrounding orbital tissues, except in orbits that have undergone prior surgical exploration. The posterior extent of the tumor, which may extend to the
orbital apex and beyond, often cannot be acoustically delimited and should be studied with MR; (c) high acoustic absorption that does not permit good outlining of
the orbital wall; and (d) acoustic homogeneity, with few or limited internal reflections.
Figure 5.26 The same meningioma of Figure 5.25 with an accompanying A-scan. Orbital wall (o) is well defined
in this image.
In the interior of these solid masses, scattered low-amplitude echoes (or internal reflections) occur because of occasional acoustic discontinuities in the tumor
mass, which are best demonstrated on the A-scan presentation (Figure 5.27). The presence of interfaces between tumor cell planes, collagenous connective tissue
septa, and large blood vessels in the tumor provides a possible histologic basis for these acoustic discontinuities.
Tumors of the optic nerve often produce a smooth indentation of the anterior retrobulbar pattern of the nerve. Enlargement of the normal optic nerve shadow,
however, is the most important feature in the diagnosis. The normally acutely angled optic nerve shadow is converted into an obtuse-angled or smoothly convex
shadow.
These solid tumors show a similar appearance at examining frequencies of 5,10,15, and 20 MHz. In general, 10 MHz provides the best compromise of penetration
and resolution in outlining of these tumors.
CDI flow reduction in the central retinal artery has been reported in gliomas and meningiomas (36). Optic nerve gliomas show considerable enhancement on
gadolinium fat suppressed T1-weighted MR images. Optic nerve head meningiomas show a classic “tram track sign” on axial scans and the “donut sign” on coronal
scans, where the surrounding hyperintense nerve sheath is distinguishable from the nerve (37, 38, 39, 40) (Figure 5.28).
P.158
Figure 5.27 An expanded A-scan through the meningioma shown in Figures 5.25 and 5.26. This tumor is solid
but homogeneous, and only low amplitude echoes are seen.
Figure 5.28 A contrast MR of a large meningioma in the axial plane shows enhancement. The coronal
T2-weighted view shows a very extended “donut” sign.
ACOUSTICALLY “ANGIOMATOUS” TUMORS

Acoustically “angiomatous” orbital tumors have irregular outlines, as compared to the previously discussed tumors with rounded outlines. In these orbital tumors, as
in the cystic group of tumors, there is good sound transmission through the mass lesion.
Diffuse Lymphangioma (Venolymphatic Malformation)

Lymphangiomas in the region of the eye are uncommon (22). Reese (19) described the proportion of lymphangiomas to total orbital masses as 1.8% in a
histopathologic study of 877 cases, and Silva (35) found three lymphangiomas in a series of 300 consecutive orbital masses, a proportion of 1%. Shields et al. (27)
found them to comprise 4% of their 1,264 orbital masses. The clinical differentiation of lymphangiomas from other orbital tumors, which is difficult when lid or
conjunctival involvement is absent (42,43), may be aided by ultrasonography.
The location, size, and outline of lymphangiomas may be demonstrated with B-scan ultrasound. All diffuse lymphangiomas exhibit a highly irregular outline because
the tumor is not encapsulated and extends diffusely through
P.159
the orbit. Numerous fingerlike or lobate projections of the tumor into the retrobulbar fat pattern are seen ultrasonically. Occasionally, these projections are sectioned
transversely by the examining ultrasound beam, giving the appearance of a cyst in the retrobulbar fat pattern. The acoustic borders of these diffuse lymphangiomas
are moderately well defined, and they are acoustically well demarcated from surrounding orbital structures because of the abrupt acoustic discontinuity between fat
and muscle tissue and the fluid-filled lymphangioma. Lymphangiomas demonstrate definite acoustic hollowness. Sound transmission through these fluid masses is
good, and little sound attenuation occurs. As a result, the posterior extent of the tumor is well outlined acoustically. M-scans through lymphangiomas may show
pulsatile activity. However, lymphangiomas typically do not show flow on CDI (25). MR T1-weighted images with gadolinium contrast and fat suppression generally
show hypointense regions corresponding to blood and fluid-filled regions, along with isointense septa. On T2-weighted images these lesions are hyperintense with
isointense rims (41, 42, 43).
ACOUSTICALLY “INFILTRATIVE” TUMORS

Acoustically “infiltrative” tumors of the orbit demonstrate a highly irregular outline and very poor sound transmission, in contrast to the cystic and solid groups of
tumors. A number of tumors demonstrate this ultrasonic pattern. The pattern may also be seen in certain presentations of orbital pseudotumors (idiopathic orbital
inflammation).
Lymphomas
Malignant lymphomas are a relatively common cause of unilateral exophthalmos. Reese (19) described the proportion of malignant lymphomas to total orbital
masses as 10% in a clinical study of 230 cases and as 14% in a histopathologic study of 877 cases. Silva (35) found 21 lymphomas in a series of 300 consecutive
orbital masses, a proportion of 7%. Shields et al. (27) found a proportion of 11% in their series from the Ocular Oncology Service at Wills Eye Hospital. This group
of tumors is typically treated with radiation therapy through surgical excision, or chemotherapy may be appropriate in some cases. Tissue biopsy is essential for
definitive diagnosis, but preoperative ultrasonic differentiation of lymphomas from other orbital tumors can be useful clinically because a minimally traumatic
surgical approach can be planned because excision is not indicated.
Lymphomas of the orbit on pathologic examination are nonencapsulated and are poorly demarcated from the surrounding orbital structures. These properties
permit their ultrasonic distinction from rounded, discrete, encapsulated orbital tumors such as cavernous hemangiomas and orbital cystic lesions.
Lymphomas are randomly located in the orbit, and most are fairly large when examined. Secondary changes of the globe, such as flattening of the posterior pole,
are demonstrable by ultrasonography just as with cavernous hemangiomas.
The group of orbital acoustic features characteristic of all lymphomas (whether lymphocytic lymphomas or reticulum cell lymphomas) includes: (a) irregular contour;
(b) scalloped or lobulated outline; (c) demarcation from surrounding structures; (d) acoustic solidity, with high acoustic absorption; and (e) acoustic homogeneity.
All orbital lymphomas present ultrasonically as solid acoustic masses. Sound energy is readily absorbed in these masses and does not penetrate through them to
outline the posterior extent of the tumor, the orbital walls, or the apex of the orbit.
In the interior of these solid masses, acoustic echoes or internal reflections do not occur. The monotonous, highly cellular tumor pattern does not provide acoustic
discontinuities. Poor acoustic penetration of the orbit occurs with high ultrasound examining frequencies, and lymphomas are best outlined at 5 and 10 MHz. CDI
can reliably demonstrate localized flow in these lesions in all areas of the orbit (31). With gadolinium contrast MR mild enhancement of the tumor can be seen on
fat suppressed T1 images. On T2-weighted images the lymphoma appears isointense with fat (44, 45, 46, 47, 48, 49, 50).
Metastatic Carcinomas
Figure 5.29 is the ultrasonogram of a patient with disseminated breast carcinoma and known orbital metastasis. The ultrasonogram demonstrates the typical
pattern in metastatic carcinoma, that of irregular acoustically empty (black) spaces infiltrating the retrobulbar fat pattern. This patient exhibited unilateral
enophthalmos rather than exophthalmos. Adenocarcinoma commonly appears as enophthalmos. CDI also demonstrates flow
P.160
in metastatic lesions and may be more typically seen on the lesion periphery (25,31). In general, metastatic lesions show enhancement with gadolinium contrast
and are hypointense on T1-weighted images and hyperintense on T2-weighted images. Important to remember in the case of melanotic melanoma metastases is
that this pattern is reversed because of the paramagnetic properties of melanin (51, 52, 53).
Figure 5.29 Metastatic carcinoma of the orbit shown on B-scan. The tumor is shown compressing the optic
nerve (arrow).
Fibrous Histiocytoma
Fibrous histiocytomas are nonencapsulated, locally invasive tumors that tend to recur after surgery. They are the most common fibrous tumor of the orbit. These
tumors demonstrate a similar ultrasonic pattern to that seen in lymphomas and metastatic carcinomas. Flow can be demonstrated in these tumors with CDI and
may be dependent on lesion size. On MR it is difficult to differentiate these lesions from other fibrous tissue tumors of the orbit, although a more homogeneous and
isointense presentation on the T2-weighted image is more indicative of these tumors (54).
Rhabdomyosarcoma
Rhabdomyosarcoma is the most common malignant neoplasm of mesenchymal origin in the orbit (19). It occurs predominantly in Caucasians in the first decade of
life. These tumors ultrasonically appear similar to the previously described types of infiltrative tumors. They are irregular in outline and acoustically solid (Figure
5.30). They may become very large. Ultrasonography can demonstrate serial changes in rhabdomyosarcoma over time. In most CDI series orbital
rhabdomyosarcoma shows large areas of relatively high flow (25). Rhabdomyosarcoma can present with a wide spectrum of MR features. Although most lesions
enhance with contrast, the appearance on T1-weighted and T2-weighted images is variable, ranging from hypointense to hyperintense, with moderate homogeneity
to marked heterogeneity (55,56).
Figure 5.30 Rhabdomyosarcoma of the orbit shown with a small part linear array. This demonstrates the
acoustic solidity and irregularity of these tumors (Courtesy of Barrett Haik, MD, Memphis, TN.)
Pseudotumors (Idiopathic Orbital Inflammation)

The mass lesion type of orbital pseudotumor presents an ultrasonographic appearance that is similar to the infiltrative tumor pattern that has been discussed. This
area is difficult in ultrasonic differential diagnosis. The patterns in orbital pseudotumors will be discussed in later section on orbital inflammation.
LACRIMAL GLAND TUMORS

Tumors of the lacrimal gland can be benign or highly neoplastic. Benign pseudotumors are confined to the gland, although inflammation may invade adjoining
muscle. Adenocarcinoma or squamous carcinoma involving the lacrimal gland may invade other tissues. Because only the anterior tip of these tumors can be
palpated, ultrasonography is useful in determining the size, extent, and configuration of tumors in this area.
Lacrimal gland tumors generally appear as solid tumors with well-demonstrated contours, poor sound transmission, and few internal echoes. The size and shape
are highly variable, and, because of bony overhang from the superior orbital rim, may not be easily outlined to their full extent with B-scan. The most useful aspect
of the ultrasonic evaluation is in determining the posterior extension of the tumor and the lateral enlargement of invading orbital tissue (Figure 5.31). This finding is
an ominous indication of a probable malignant lesion. Color flow in lacrimal gland tumors is dependent on the tumor type
P.161
that ranges from vascular lesions to fibrous tumors. CDI can only be used to broadly categorize lesions as having or not having a discernible flow component.
Similarly, there is a broad spectrum of MR findings related to the type of lacrimal gland tumor (57, 58, 59, 60).
Figure 5.31 A well-encapsulated lacrimal gland tumor on high frequency B-scan. High frequency and 20-MHz
B-scan imaging can delineate these lesions in the anterior adnexa.
In all cases MR or CT are the best modalities to use to complete the localization of the tumor extent and to evaluate whether bone destruction is present.
ORBITAL INFLAMMATION
Ultrasonography can reliably show inflammation of the extraocular muscles, optic nerve, Tenon's space, localized inflammatory lesions (pseudotumors), and other
forms of localized and diffuse orbital inflammation (Figure 5.17). These different topographic types of inflammation may exist singly or together.
Presurgical differentiation of these inflammatory processes from mass lesions of the orbit is difficult on clinical criteria alone, and ultrasonic identification of an
inflammatory process may obviate the need for surgical biopsy or, alternately, direct its route.
INFLAMMATION OF THE EXTRAOCULAR MUSCLES (MYOSITIS)

In the normal orbit, only a small portion of either orbital wall is usually seen. The extraocular muscles are represented acoustically as hypoechoic or black spaces
between the retrobulbar fat and the occasional low-amplitude echoes from the orbital wall. The extraocular muscles in the normal orbit appear acoustically clear,
because of the highly ordered structure of muscle, and because muscle fibers and connective tissue septa lie at an acute angle to the examining ultrasound beam
so that no or few echoes are returned to the transducer.
Nonspecific Myositis
Although Graves disease is the most common cause of unilateral exophthalmos, a myositis of nonspecific etiology may cause the B-scan ultrasonogram to show
accentuation of the orbital wall and enlargement of only a portion of any one of the rectus muscles (Figure 5.32).
In general, such a localized inflammation would appear acoustically as a bulbous or globular enlargement of the muscle in its posterior portion, with less
involvement of the more anterior portion of the muscle. Indentation of the apical orbital fat pattern by an enlarged portion of an extraocular muscle may simulate a
rounded tumor when scanned at right angles. This common misinterpretation may be avoided by serially delineating the exact extent of muscle enlargement. If,
however, a single muscle is enlarged along its entire course, in cases of myositis, such as chronic granulomatous episcleritis, acoustic differentiation from Graves
disease cannot be made. Slow flow can, occasionally, be documented in enlarged muscles using CDI, and ophthalmic artery flow can be increased with active
inflammation (61). Changes in muscle conformation are easily seen with fat-suppressed T1-weighted images. With active inflammation, muscles are hyperintense
on T2-weighted images (62).
Figure 5.32 A 10-MHz B-scan of the enhanced outline of the ocular muscle (arrows) in a patient with
nonspecific myositis.
Graves Disease (Thyroid Orbitopathy)

The severe ocular changes that occur with Graves disease have been described as “progressive exophthalmos,” “malignant exophthalmos,” “thyrotropic
exophthalmos,” “dysthyroid ophthalmopathy,” and “endocrine exophthalmos.” The American Thyroid Association (22) has suggested the term “eye changes of
Graves disease,” but, generally, it is referred to as thyroid orbitopathy. Thyroid orbitopathy is the most frequent cause of unilateral exophthalmos and, by far, the
most common cause of bilateral exophthalmos (19).
The eye changes in hyperthyroidism result from an increase in volume in orbital tissues. The ground substance of orbital connective tissue increases in amount,
mast cells and lymphocytes become prominent, and extraocular muscles increase greatly in volume (up to eight times that of normal). Electron microscopy of
muscles so involved has shown marked infiltration of inflammatory cells and interstitial edema, with little muscle fiber involvement (63). One of the most important
uses of diagnostic orbital ultrasonography is the demonstration of changes in both orbital fat and extraocular muscles that occur in the active eye changes of
Graves disease.
Clinically, the inferior rectus muscle is most commonly involved in Graves disease. Ultrasonically, however, it is
P.162
easier to demonstrate the medial and lateral rectus muscles, (Figure 5.33), because the medial and lateral orbital walls are less obstructed by bony overhang than
the superior and inferior orbital walls. Ultrasonography may show muscle swelling without clinical evidence of impairment of muscle function and, indeed, with no
other signs or symptoms.
Figure 5.33 A 10-MHz B-scan of an enlarged and inflamed lateral muscle (arrows) in a patient with Graves
disease. The B-scan image is acoustically indistinguishable from the myositis shown in Figure 5.32.
In Graves disease, a large portion of the orbital wall is usually evident ultrasonically, and, occasionally, both orbital walls are demonstrated. The extraocular
muscles still appear acoustically clear, but the space between retrobulbar fat and the orbital wall is increased, indicating enlargement of the muscles in this
condition. This enlargement ranges from minimal to marked, and when the scan plane is adjusted to pass through the greatest diameter of the muscle, the
enlargement can be quantified, to some degree. In an optimal plane, tendon sparing can be shown with an enlarged muscle belly (Figure 5.34).
The posterior outline of the retrobulbar fat also often appears “scalloped” or indented because of compression of the retrobulbar fat by markedly enlarged
extraocular muscles. No obvious increase occurs in volume of the retrobulbar fat as well as any apparent difference in the echo pattern produced in its interior (i.e.,
its acoustic texture) as compared to the normal orbit.
Although MR imaging is the best way of demonstrating this total muscle involvement, changes in extraocular muscle enlargement during the disease course and
steroid treatment can be monitored with ultrasonography.
In advanced orbital involvement, edematous changes about the optic nerve may be shown ultrasonically (Figure 5.35). These perineural changes are demonstrated
ultrasonically as definite (though often subtle) duplication echoes of the optic nerve outline. Optic neuritis of origin other than Graves disease will be discussed in a
later section.
Figure 5.34 A 10-MHz B-scan with “Tendon sparing” showing an enlarged muscle belly (arrows) in Graves
disease.
Most patients with Graves disease and exophthalmos present no problem diagnostically. However, the appearance of unilateral exophthalmos may raise the
question of conditions, such as orbital tumors, orbital granulomas, and carotid-cavernous fistula.
A general increase in orbital arterial flow compared with published normal levels is seen with orbital inflammation in Graves disease with CDI (64).
T1-weighted MR contrast studies suggest that increased muscle enhancement may be related to disease severity (65, 66, 67). T2-weighted images may show fatty
degeneration of muscles as hyperintense foci. Although Graves disease inflammation is typically tendon sparing,
P.163
MR can demonstrate minimal tendon involvement in some patients (68).

Figure 5.35 A 10-MHz B-scan of enlarged lateral rectus with segmented inferior rectus enlargement seen at the
apex of the orbit. Vertical and meridional scans can outline each individual muscle.
In examining a large number of patients with various thyroid abnormalities with Dr. Sidney Werner (69) of Columbia University, we found that a certain number of
patients with nontoxic nodular goiter and myxedema showed the enlargement of extraocular muscles as seen in Graves disease. The reason for myositis to be
present in thyroid conditions other than Graves disease is the result of circulating autoantibodies and is becoming more fully understood. Experimental studies
have shown that the rectus muscles of both hypothyroid and hyperthyroid animals have numerous macrophages (70). This suggests the potential for localized
inflammatory response with varying thyroid functionality.
As a final note, it should be mentioned that patients with clinically obvious Graves disease may also have orbital tumors. Dallow (71) has had experience with a
small number of such patients. Ultrasonography and neuroimaging studies should not be neglected in these patients, because they may show an orbital tumor in a
patient who has been followed for many years with the diagnosis of exophthalmos as a result of Graves disease.
INFLAMMATION OF THE OPTIC NERVE

Inflammatory change around the optic nerve can be shown using B-scan ultrasonography (72). This condition can manifest itself as a single disease state, such as
optic neuritis, or as a secondary complication of pathology, such as orbital inflammatory disease or pseudotumor cerebri (idiopathic orbital inflammation).
The outline of the normal optic nerve shadow is single-walled, forming an acoustic line that blends smoothly with the retrobulbar fat pattern (Figure 3.136). In cases
of optic inflammatory change, there is doubling of the optic nerve shadow or acoustic accentuation of the optic nerve outline. Inflammatory states of the optic nerve
may produce a variety of anomalous echoes. We have noted three typical acoustic patterns. The first is accentuation of the optic nerve sheath, which may appear
as uniformly spaced linear echoes or as a more visually apparent configuration in which one or both sides of the nerve appear doubled (Figure 5.36). The second
finding, closely related to doubling, is the demonstration of a “half-ring” around the distal (globe) end of the optic nerve, but without a marked posterior extension,
which is an accentuation of Tenon's (detailed in later section) and Keeney (75) referred to as the “T” sign (Figures 3.79 and 3.153). Finally, random echoes may be
noted lying perpendicular to the path of the nerve, within the normally sonolucent nerve shadow. Although this pattern may be initially interpreted as an artifact
(Figure 5.11) on testing, the echoes are real and may be reproduced in varying scan angles.
In a study on the sensitivity of this ultrasound finding, Coleman and Carroll (73) found that 90% of patients with clinically confirmed retrobulbar neuritis
demonstrated ultrasonic evidence of optic nerve abnormalities.
Figure 5.36 A 20-MHz B-scan demonstrating accentuation of the meningeal sheath of the optic nerve.
The ultrasonic characteristics demonstrating inflammatory or edema-like changes around the optic nerve are not specific for optic neuritis. They may be noted in
Graves disease with optic nerve involvement, in pseudotumor with involvement of the optic nerve, in venous congestion of the orbit and optic nerve, and with
carotid cavernous fistulas. A somewhat similar pattern was also observed in a patient found to have an aneurysm of the ophthalmic artery.
In an attempt to ultrasonically portray these inflammatory accentuations of the optic nerve, positions of gaze, sensitivity settings, and transducer angulation should
be varied to establish the acoustic patterns. However, even in a normal nerve, extreme positions of gaze may bring portions of the nerve sheath more
perpendicular to the examining ultrasound beam and produce a doubling of the sheath that may be erroneously interpreted as a pathologic condition.
The precise source of redundant echoes outlining the optic nerve in these cases is not known. The acoustic pattern seen seems most consistent with inflammatory
edema of the meningeal space and fluid distention of the meningeal sheaths.
Ophthalmic artery peak systolic and end diastolic flow rates are increased compared to the contralateral eye in optic neuritis consistent with local inflammation (74).
The optic nerve swelling can be demonstrated with MR along with enhancement with contrast. Increased resistance in CRA and the posterior ciliary artery (PCA)
has been seen in optic neuritis with multiple sclerosis (76). Enhancement with contrast can be demonstrated with MR (77) along with optic nerve swelling in a small
number of cases (78). Typically, the ultrasound boundary size changes believed to be associated with nerve size changes are not seen with MR in optic neuritis.
P.164
Gadolinium enhancement has been shown to not be correlated with visual performance or recovery (79,80). Increased signal is seen in T2-weighted images. Some
studies suggest that water-suppressed (FLAIR) T2-weighted images be used to provide better delineation of the nerve (81,82).
INFLAMMATION OF THE RETROBULBAR FAT AND TENON'S SPACE

Primary inflammatory changes of orbital tissues have their own unique ultrasonic appearance. Episcleritis or panophthalmitis can produce an inflammatory edema
of the potential space of Tenon (Figure 5.37). A diffuse mottling of the orbital fat pattern may acoustically indicate orbital cellulitis (Figure 5.38).
Some cases of orbital cellulitis may have discrete cavities that can be acoustically outlined. Almost invariably, associated inflammatory changes are also seen.
PSEUDOTUMOR (IDIOPATHIC ORBITAL INFLAMMATION)

The term orbital pseudotumor is used to describe “inflammatory lesions of the orbital tissues of unknown cause, simulating in their clinical picture a neoplasm of the
orbit” (83). Pseudotumors are some of the most frequent causes of unilateral exophthalmos (16% in Reese's series) (19) and are a significant cause of bilateral
exophthalmos. Pseudotumors occur predominantly in older individuals and are less frequent in childhood and youth.
Orbital pseudotumors have been subdivided into specific types (i.e., those with a known, specific cause, such as sarcoidosis, collagen vascular disease,
xanthogranuloma, foreign body) and nonspecific types (i.e., those with no demonstrable etiologic agent). The nonspecific pseudotumors may be divided further into
pathologic types, such as posterior scleritis, myositis, vasculitis, lipogranulomas, and dacryoadenitis (84). Reese (19) used the term orbital pseudotumor for only
the nonspecific, idiopathic types of orbital inflammation and as a diagnosis of exclusion after known causes (especially the eye changes of Graves disease) were
ruled out. We prefer the term orbital pseudotumor to refer to the idiopathic orbital inflammatory lesions only. More recently, the term idiopathic orbital inflammation
has become the preferred designator for this catch-all diagnosis (51,85).
Figure 5.37 Inflammation of the optic nerve at 10 MHz with papilledema and doubling of the nerve (arrows).
Figure 5.38 Accentuation of Tenon's capsule, and the nerve is seen commonly with optic neuritis and with
episcleritis. Higher frequency scans can show subtle changes in Tenon's capsule that may be missed with
10-MHz ultrasound (Figure 3.174).
The following clinical features should suggest pseudotumor in cases of exophthalmos: (a) later age of onset than primary neoplasms, (b) more acute onset than
primary neoplasm, (c) sometimes bilateral, (d) pain and edema of lids or conjunctiva in 50% of cases, and (e) regression of exophthalmos with steroid therapy.
These features, however, are not diagnostic, although ultrasonography may be beneficial to the clinician in diagnosing these orbital conditions (19).
Orbital B-scan ultrasonography in cases of orbital pseudotumor may show two characteristic types of orbital abnormality: inflammatory mass lesions of the orbit and
inflammatory edema of normally present orbital structures.
The size and location of pseudotumors are determined, as in all orbital conditions, by serial acoustic sectioning of the orbit. Pseudotumors of the orbit usually
demonstrate many of the characteristics of lymphomas and metastatic carcinomas, including an irregular outline and acoustic solidity with high acoustic absorption.
Pseudotumors may appear as a dark area encroaching on the normal retrobulbar fat pattern from its periphery.
P.165
The dark area usually has an irregular anterior border but, on occasion, may have a rounded border.
Pseudotumors may also appear as acoustically empty (black) areas, diffusely infiltrating the retrobulbar fat.
Pseudotumors of the “en plaque” type, formed on the orbital wall, are best seen with MR, although they may be visualized ultrasonically.
Another morphologic variant of the orbital pseudotumor is a tumor that surrounds the area of the optic nerve. Readily differentiable from an optic nerve tumor, the
edges of the area are irregular, and signs of inflammatory edema of the optic nerve and Tenon's space are usually demonstrable.
Most pseudotumors present as acoustically solid. Sound energy is highly attenuated by these masses and usually does not penetrate through them to clearly
outline the posterior extent of the tumor or the orbital wall.
Solid orbital mass lesions of irregular outline (such as lymphosarcoma and metastatic carcinoma) are similar in appearance acoustically to pseudotumor and form
the category most likely to be misinterpreted as pseudotumor. Conversely, pseudotumor may be mistaken for a neoplastic tumor of this type. Changes as a result
of inflammatory edema may often be noted acoustically in adjacent orbital structures, thus indicating the inflammatory character of the mass lesion. These changes
may be noted in the optic nerve, the extraocular muscles, or Tenon's space by ultrasound.
Pseudotumors are the most difficult “tumors” of the orbit to diagnose acoustically and account for the majority of differential problems in orbital ultrasonic
evaluation.
This large group of orbital inflammatory lesions, which covers such a wide spectrum of pathologic manifestations, has a range of varying appearances in ultrasonic
B-scans. They usually appear as solid masses, with poor acoustic transmission and cause irregular indentation of the retrobulbar fat pattern. Thus, they may be
erroneously classified as “irregular, solid” tumors. Lymphoid hyperplasia or lymphoid pseudotumors may be identical to lymphomas on B-scan ultrasonographic
examination. Similarly, diffuse inflammation of the retrobulbar fat closely resembles the patterns produced by a metastatic carcinoma.
Figure 5.39 10-MHz B-scans are of a patient with an arteriovenous malformation in which positional change
modified the filling and hemodynamic function of the lesion. Right: Scan with the patient sitting. Left: Scan with
the patient lying prone.
In our experience, B-scan ultrasonography cannot differentiate between inflammatory mass lesions and “solid, irregular” neoplastic lesions unless acoustic signs of
inflammatory edema are found. Signs of myositis, sclerotenonitis, and optic nerve inflammation must be carefully sought whenever the “irregular, solid” tumor
pattern is seen, because we believe that these signs are usually indicative of a primary inflammatory process.
Pseudotumor is typically seen as isointense with muscle on T1-weighted MR images and isointense to minimally hyperintense to fat on T2-weighted MR images
(86, 87, 88). This pattern is similar to that seen with myositis and sarcoidosis.
ARTERIOVENOUS ORBITAL ANOMALIES

Patients with arteriovenous abnormalities exhibit several clinical signs. They may or may not have exophthalmos, and, if present, it may be intermittent or induced
by patient posture or environmental factors. Occasionally, the patient will note intermittent or more chronically apparent ectasia and/or discoloration in the ocular
adnexa. Bruits may be noted by either the examiner or the patient.
In many patients with these conditions, no B-scan ultrasonographic abnormalities can be discerned in the orbit. However, gross abnormalities, such as those with
poolings of blood, are readily demonstrable on B-scan but must be differentiated from simple cystic masses (Figure 5.39). The A-scan should be carefully
monitored
P.166
for vascular pulsations, which may then be further characterized with CDI studies. The combined Mueller-Valsalva maneuver may be implemented in certain cases
to markedly alter the venous flow, which thereby increases the size of the cystic spaces portrayed on the B-scan.
Most patients with arteriovenous anomalies will demonstrate only subtle ultrasonic changes. These are most often minimal irregularities in the posterior fat outline
and may appear in any quadrant. If they occur adjacent to the optic nerve, they may simulate an inflammatory or atrophic state of the optic nerve.
Vascular pulsations are not necessarily apparent on A- or M-scan and thus this valuable differential aid is frequently not available. If this type of abnormality is
clinically suspected and routine horizontal examinations are inconclusive, vertical scanning may detect an orbital varix obscured by the orbital rim.
High flow anomalies are easily seen with both CDI (localized high flow regions with turbulence and signs of retrograde flow) and MR (flow voids) (30, 31, 32,89,90).
Enlarged orbital veins can also be demonstrated with CDI, although complete evaluation of the angioarchitecture of arteriovenous anomalies with angiography or
MRA is often warranted (91).
OPTIC NERVE ABNORMALITIES

Optic nerve drusen give high reflectivity, and their identification can be helpful in evaluating optic nerve disease (Figure 5.40).
Melanocytoma is generally regarded as a benign hyperpigmented lesion surrounding the optic disc. Melanin is highly reflective and can be ultrasonically imaged as
it extends along the nerve sheath (Figure 5.41).
Figure 5.40 Optic nerve head drusen can produce changes in the optic nerve shadow, as shown in the 10-MHz
B-scan, including local absorption defects (arrow).
Figure 5.41 A 10-MHz B-scan of a patient with a melanocytoma showed accentuation of the nerve sheath,
indicating the extension of the lesion along the anterior aspect of the nerve (arrow).
ORBITAL TRAUMA
Orbital trauma causes a variation in echo patterns requiring careful ultrasonic analysis. This section on orbital trauma will discuss ultrasonic localization and
characterization of orbital foreign bodies, optic nerve damage, and orbital hemorrhage. The latter two types of traumatic change can occur with foreign bodies,
lacerations, or concussion.
ORBITAL FOREIGN BODIES

Ultrasonic detection of orbital foreign bodies, whether radiopaque or radiolucent, depends on their size, location, orientation, and the presence of associated
inflammatory change. Orbital foreign bodies are by no means always demonstrable by ultrasound. If a foreign body of any size lies along the posterior sclera or in
the retrobulbar fat, echoes from the foreign body may be lost in the retrobulbar echo complex. Very small orbital foreign bodies are not usually seen with present
ultrasonographic techniques.
Some metallic foreign bodies may be located by selectively decreasing the gain (Figure 5.42) and by studying high-amplitude spikes or ringing patterns (Figure
5.43). Smaller metallic orbital foreign bodies may be detected with use of more subtle ultrasonic techniques, if prior radiographic localization has been performed.
Vegetable material, on the other hand, usually returns lower-amplitude echoes that are not identifiable amidst the echoes from the fat. If, however, the vegetable
foreign body is surrounded by hemorrhage, edema, or purulence,
P.167
the cystic area may enhance the outline of the foreign body.
Figure 5.42 A 10-MHz B-scan of an orbital foreign body is seen adjacent to the optic nerve (arrow). It can be
emphasized by decreasing the gain while scanning.
We prefer a helical CT to be performed prior to ultrasound to detect the number and general location of expected foreign bodies (92, 93, 94). MR should never be
used in cases of suspected foreign body before CT has been performed to rule out metallic foreign body.
ORBITAL HEMORRHAGE
The ultrasonic patterns seen in hemorrhage of the orbit, whether traumatic or postsurgical, fall into two general groups. We have found the most common pattern to
be that of a mass lesion of the orbit, appearing much as a solid tumor. In this pattern, similar to that seen with an abcess, a hypoechoic area exists, replacing an
area of orbital fat (Figure 5.44). Its contour may be regular or irregular, the area is relatively well demarcated from surrounding structures, and moderate sound
transmission occurs through the area. In children with unilateral proptosis and a history of trauma, ultrasonography cannot definitely differentiate between
rhabdomyosarcoma and a massive orbital hemorrhage, because these often have a similar ultrasonic appearance. In this situation, CDI that shows a flow
component in rhabdomyosarcoma or MR imaging should be performed. Follow-up over a short period of time and repeated ultrasonography will also provide the
correct diagnosis, because the hemorrhage should resolve, and a rhabdomyosarcoma would enlarge rapidly. CDI can also be used to identify hemorrhages
secondary to orbital arteriovenous malformation (AVM).
Figure 5.43 A 10-MHz B-scan of a round foreign body lodged in Tenon's space that produces a ringing artifact
(arrow) that helps localize its position.
Figure 5.44 A 10-MHz B-scan of an orbital abscess is shown here as an irregular accentuation of fluid
extending along Tenon's space and posteriorly.
Another ultrasonic pattern seen in orbital hemorrhage is the “diffuse inflammatory disease pattern,” as is seen in certain cases of orbital cellulitis and orbital
pseudotumor. In our experience, this pattern is much less commonly seen than the mass lesion pattern described previously. If the hemorrhage is anterior in the
orbit, it may infiltrate Tenon's space, simulating inflammation. Small orbital hemorrhage will show minimal irregularity of the posterior fat pattern.
The MR pattern is dependent on the stage of the blood products in the hemorrhage. Acute hemorrhage is hypointense on regular T1-weighted and hyperintense on
T2-weighted; subacute hemorrhage is hyperintense on
P.168
T1-weighted and hypointense on T2-weighted. With cell lysis T2-weighted localized regions become hyperintense. Chronic hemorrhage is hypointense on both
T1-weighted and T2-weighted images (95,96).
OPTIC NERVE TRAUMA

We have made an attempt to determine whether ultrasonography can demonstrate any orbital changes in traumatic cases with avulsion of the optic nerve. In the
cases examined, the V-shaped optic nerve shadow in the retrobulbar fat moved well with change of gaze, and no acoustic changes were seen, unless an orbital
hemorrhage was present, in addition to the optic nerve damage. The optic nerve, when avulsed, may still be held in position by the meninges and surrounding
retrobulbar fat, and move on the kinetic scan in a fashion identical to that seen in the normal orbital ultrasonographic examination.
USEFULNESS, RELIABILITY, AND LIMITATIONS OF ORBITAL ULTRASONOGRAPHY

Orbital ultrasonography is a sensitive and reliable diagnostic technique for demonstrating soft tissue abnormalities of the orbit. Comments in this section refer to
high-resolution B-scan tomographic methods of ultrasound examination with water immersion.
Orbital mass lesions greater than 2 mm wide can be shown, if they are isolated and discrete. Larger masses usually present no problem for identification. Tumors
located at the orbital apex are difficult to recognize because of the attenuation of sound and confluence of optic nerve and muscles that are inseparable
ultrasonically. Tumors originating or extending along the bony wall of the orbit in an “en plaque” configuration, as with meningioma, osteoma, or pseudotumor, do
not present a reflecting surface perpendicular to the ultrasonic beam, and, consequently, do not produce distinct echoes. An enlarged space between retrobulbar
fat and bony orbital wall in a localized area may suggest tumor, but absolute demonstration may not be possible.
Ultrasonic signs of generalized inflammation and congestion are nonspecific and may appear with several types of disease processes, including cellulitis,
pseudotumor, and passive venous congestion from arteriovenous abnormalities. Inflammatory signs localized to one tissue element, such as optic neuritis, focal
granuloma, or orbital myositis, are, however, more specific. Orbital changes of infiltrative tumor and pseudotumor overlap considerably, making this the most
difficult tumor category to diagnose definitively.
Floor fractures and surgical defects of the orbital wall, as well as hyperostosis of bone, are not reliably detectable with ultrasonic techniques, and CT or MR
supplementation is required to augment the soft tissue information obtained with ultrasound.
B-scan ultrasonography of the orbit can accurately and reliably diagnose many soft tissue abnormalities of the orbit. MR and CT provide a more accurate and
anatomically detailed presentation of orbital disease, but inflammatory changes may produce morphologic tissue changes seen earlier with ultrasound.
Ultrasonography graphically indicates the location, size, extent, and basic tissue type of detected lesions, and is a proven invaluable aid to orbital diagnosis and
surgery when used in a complementary fashion with MR and CT scans of the orbit.
REFERENCES
1. Byrne SF, Green RL. Ultrasound of the Eye and Orbit. 2nd ed. St. Louis: Mosby; 2002.
2. Dibernardo C, Schachat AP, Fekrat S. Ophthalmic Ultrasound: A Diagnostic Atlas. 1st ed. New York: Thieme Medical Publishers; 1998:143.
3. Ossoinig K. Clinical echo-ophthalmography. In: Blodi FC, ed. Current Concepts in Ophthalmology. St. Louis: Mosby; 1972:101-130.
4. Buschmann W, Voss M, Kemmerling S. Acoustic properties of normal human orbit tissues. Ophthalmol Res 1970;1: 354-364.
5. Trokel SL. Radiology and ultrasonography in the diagnosis of orbital lesions. Trans Pa Acad Ophthalmol Otolaryngol 1974;27:10-14.
6. Purnell E. Ultrasonic interpretation of orbital disease. In: Gitter K, Kenney AH, Sarin LK, eds. Ophthalmic Ultrasound. St. Louis: Mosby; 1969:249-255.
7. Ustymowicz A, Krejza J, Mariak Z. Twinkling artifact in color Doppler imaging of the orbit. J Ultrasound Med 2002;21(5):559-563.
8. Ossoinig K, Till P. A ten-year study of clinical echography in orbital disease. In: Francois J, Goes F, eds. Ultrasonography in Ophthalmology. Basel,
Switzerland: Karger; 1975: 200-216.
9. Coleman DJ, Jack HL, Franzen LA. High resolution B-scan ultrasonography of the orbit. Part I: the normal orbit. Arch Ophthalmol 1972;88:358-367.
10. Coleman DJ, Jack RL, Franzen LA. High resolution B-scan ultrasonography of the orbit. Part II: hemangiomas of the orbit. Arch Ophthalmol
1972;88:368-374.
11. Coleman DJ, Jack RL, Franzen LA. High resolution B-scan ultrasonography of the orbit. Part III: lymphomas of the orbit. Arch Ophthalmol
1972;88:375-379.
12. Coleman DJ, Jack RL, Franzen LA. High resolution B-scan ultrasonography of the orbit. Part IV: neurogenic tumors of the orbit. Arch Ophthalmol
1972;88:380-384.
13. Coleman DJ, Jack RL, Franzen LA, et al. High resolution B-scan ultrasonography of the orbit. Part V: eye changes of Graves' disease. Arch Ophthalmol
1972;88:465-471.
14. Coleman DJ, Jack RL, Jones IS, et al. High resolution B-scan ultrasonography of the orbit. Part VI: pseudotumors of the orbit. Arch Ophthalmol
1972;88:472-480.
15. Coleman DJ, Jack RL, Franzen LA. B-scan ultrasonography of orbital lymphangiomas. Br J Ophthalmol 1973;57: 193-198.
16. Coleman DJ, Jack RL, Franzen LA. B-scan ultrasonography of orbital mucoceles. Eye Ear Nose Throat Mon 1972;51:207-211.
P.169
17. Coleman DJ. Reliability of ocular and orbital diagnosis with B-scan ultrasound. 2. Orbital diagnosis. Am J Ophthalmol 1972;74:704-718.
18. Zismor J, Fasano C, Smith B, et al. Roentgenographic diagnosis of unilateral exophthalmos. JAMA 1966;197: 343-346.
19. Reese A. Tumors of the Eye. 2nd ed. New York: Harper & Row; 1963:532-537.
20. Shields J, Shields C. Orbital cysts of childhood—classification, clinical features, and management. Surv Ophthalmol 2004;49(3):281-299.
21. Motaref B, Lorenz B. [Value of 10 MHz ultrasound in diagnosis and follow-up of retrobulbar mucocele]. Ultraschall Med 2001;22(2):100-103.
22. Ohtsuka K, Hashimoto M, Suzuki Y. A review of 244 orbital tumors in Japanese patients during a 21-year period: origins and locations. Jpn J Ophthalmol
2005;49(1): 49-55.
23. Abou-Rayyah Y, Rose G, Konrad H, et al. Clinical, radiological and pathological examination of periocular dermoid cysts: evidence of inflammation from
an early age. Eye 2002;16(5):507-512.
24. Junwu G, Jun W. B-ultrasound and computed tomography scan diagnosis of 32 cases of intraorbital dermoid cysts. Trop Doct 2001;31(1):49-50.
25. Neudorfer M, Leibovitch I, Stolovitch C, et al. Intraorbital and periorbital tumors in children—value of ultrasound and color Doppler imaging in the
differential diagnosis. Am J Ophthalmol 2004;137(6): 1065-1072.
26. Chawda S, Moseley I. Computed tomography of orbital dermoids: a 20-year review. Clin Radiol 1999;54(12): 821-825.
27. Shields J, Shields C, Scartozzi R. Survey of 1264 patients with orbital tumors and simulating lesions. Ophthalmology 2004;111(5):997-1008.
28. Yan J, Wu Z. Cavernous hemangioma of the orbit: analysis of 214 cases. Orbit 2004;23(1):33-40.
29. Pichot O, Gonzalvez B, Franco A, et al. [Color Doppler ultrasonography in the study of orbital and ocular vascular diseases]. J Fr Ophtalmol
1996;19(1):19-31.
30. Lieb W. Color Doppler imaging of the eye and orbit. Radiol Clin North Am 1998;36(6):1059-1071.
31. Hatton M, Remulla H, Tolentino MJ, et al. Clinical applications of color Doppler imaging in the management of orbital lesions. Ophthal Plast Reconstr
Surg 2002;18(6): 462-465.
32. Schick U, Dott U, Hassler W. Surgical treatment of orbital cavernomas. Surg Neurol 2003;60(3):234-244; discussion 244.
33. Puca A, Colosimo C, Tirpakova B, et al. Cavernous hemangioma extending to extracranial, intracranial, and orbital regions. Case report. J Neurosurg
2004;101(6): 1057-1060.
34. Scheuerle A, Steiner H, Kolling G, et al. Treatment and long-term outcome of patients with orbital cavernomas. Am J Ophthalmol 2004;138(2):237-244.
35. Silva D. Orbital tumors. Am J Ophthalmol 1968;65: 318-339.
36. Zhang W, Zhao H, Song G. [The value of color Doppler imaging ultrasound in diagnosis of orbital diseases]. Chung-Hua Yen Ko Tsa Chih [Chinese
Journal of Ophthalmology] 2001;37(6):447-450.
37. Chateil JF, Soussotte C, Pedespan JM, et al. MRI and clinical differences between optic pathway tumours in children with and without neurofibromatosis.
Br J Radiol 2001;74(877):24-31.
38. Parsa CF, Hoyt CS, Lesser RL, et al. Spontaneous regression of optic gliomas: thirteen cases documented by serial neuroimaging. Arch Ophthalmol
2001;119(4):516-529.
39. Liu J, Forman S, Hershewe GL, et al. Optic nerve sheath meningiomas: visual improvement after stereotactic radiotherapy. Neurosurgery
2002;50(5):950-955; discussion 955-957.
40. Jackson A, Patankar T, Laitt RD. Intracanalicular optic nerve meningioma: a serious diagnostic pitfall. Am J Neuroradiol 2003;24(6):1167-1170.
41. Zucker J, Assaad M, Levine MR. Orbital lymphangioma with intracranial extension. Ophthal Plast Reconstr Surg 1995;11(1):22-26.
42. Gurelik M, Ozum U, Erdogan H, et al. Orbital lymphangioma and its association with intracranial venous angioma. Br J Neurosurg 2004;18(2):168-170.
43. Lemke A, Kazi I, Landeck LM, et al. [Differential diagnosis of intraconal orbital masses using high-resolution MRI with surface coils in 78 patients]. Rofo
2004;176(10): 1436-1446.
44. Cytryn A, Putterman A, Schneck GL, et al. Predictability of magnetic resonance imaging in differentiation of orbital lymphoma from orbital inflammatory
syndrome. Ophthal Plast Reconstr Surg 1997;13(2):129-134.
45. Gufler H, Laubenberger J, Gerling J, et al. MRI of lymphomas of the orbits and the paranasal sinuses. J Comput Assist Tomogr 1997;21(6):887-891.
46. Buescu A, Teixeira P, Coelho S, et al. Orbital lymphoma misdiagnosed as Graves' ophthalmopathy. Endocr Pract 2001;7(2):110-112.
47. Dorey S, Clark B, Christopoulos VA, et al. Orbital lymphoma misdiagnosed as scleritis. Ophthalmology 2002; 109(12):2347-2350.
48. Billing K, Malhotra R, Selva D, et al. Orbital myositis in Churg-Strauss syndrome. Arch Ophthalmol 2004;122(3): 393-396.
49. Tovilla-Canales J, Tovilla y Pomar J, Ceron JR. Lymphoproliferative disorders of the ocular adnexa. Curr Opin Ophthalmol 2004;15(5):401-405.
50. Yan J, Wu Z, Li Y. The differentiation of idiopathic inflammatory pseudotumor from lymphoid tumors of orbit: analysis of 319 cases. Orbit
2004;23(4):245-254.
51. Scott IU, Murray TG, Feuer WJ, et al. External beam radiotherapy in retinoblastoma: tumor control and comparison of 2 techniques. Arch Ophthalmol
1999;117(6): 766-770.
52. Gotwald TF, Zinreich SJ, Schocke M, et al. CT and MR imaging of orbital metastasis from islet cell carcinoma of the pancreas. Am J Roentgenol
2000;175(2):475-476.
53. Mori H, Maekawa N, Satoda N, et al. [A case of primary lung cancer with initial symptoms due to orbital metastases]. Nihon Kokyuki Gakkai Zasshi
2003;41(1):19-24.
54. Dalley RW. Fibrous histiocytoma and fibrous tissue tumors of the orbit. Radiol Clin North Am 1999;37(1): 185-194.
55. Shields CL, Shields JA, Honavar SG, et al. Clinical spectrum of primary ophthalmic rhabdomyosarcoma [See comment]. Ophthalmology
2001;108(12):2284-2292.
56. Shields CL, Shields JA, Honavar SG, et al. Primary ophthalmic rhabdomyosarcoma in 33 patients. Trans Am Ophthalmol Soc 2001;99:133-142;
discussion 142-143.
57. Bajaj M, Pushker N, Kashyap S, et al. Cystadenoma of the lacrimal gland. Orbit 2002;21(4):301-305.
58. Gunduz K, Shields C, Gunalp I, et al. Magnetic resonance imaging of unilateral lacrimal gland lesions. Graefes Arch Clin Exp Ophthalmol 2003;241(11):
907-913.
P.170
59. Scheschonka A, Mosch M, Krieglsteiner S, et al. Pre- and posttreatment MR imaging in AIDS-related Kaposi sarcoma of the conjunctiva and lacrimal
gland. Am J Neuroradiol 2003;24(7):1327-1329.
60. Looi A, Gascoyne R, Chhanabhai M, et al. Mantle cell lymphoma in the ocular adnexal region. Ophthalmology 2005;112(1):114-119.
61. Boonman Z, De Keizer R, Graniewski-Wijnands HS, et al. Orbital myositis in scleritis. Br J Ophthalmol 2003;87(1): 38-42.
62. Fatterpekar G, Gottesman R, Sacher M, et al. Orbital lyme disease: MR imaging before and after treatment: case report. Am J Neuroradiol
2002;23(4):657-659.
63. Werner S. Classification of the eye changes of Graves' disease. J Clin Endocrinol Metab 1969;29:982.
64. Alp M, Ozgen A, Can I, et al. Colour Doppler imaging of the orbital vasculature in Graves' disease with computed tomographic correlation. Br J
Ophthalmol 2000;84(9): 1027-1030.
65. Antoniazzi F, Zamboni G, Cerini R, et al. Graves' ophthalmopathy evolution studied by MRI during childhood and adolescence. J Pediatr
2004;144(4):527-531.
66. Cakirer S, Cakirer D, Basak M, et al. Evaluation of extraocular muscles in the edematous phase of Graves Ophthalmopathy on contrast-enhanced
fat-suppressed magnetic resonance imaging. J Comput Assist Tomogr 2004;28(1):80-86.
67. Taoka T, Sakamoto M, Nakagawa H, et al. Evaluation of extraocular muscles using dynamic contrast enhanced MRI in patients with chronic thyroid
orbitopathy. J Comput Assist Tomogr 2005;29(1):115-120.
68. Ben Simon G, Syed H, Douglas R, et al. Extraocular muscle enlargement with tendon involvement in thyroid associated orbitopathy. Am J Ophthalmol
2004;137(6): 1145-1147.
69. Werner SC, Coleman DJ, Franzen LA. Ultrasonographic evidence of a consistent orbital involvement in Graves' disease. N Engl J Med
1974;290:1447-1450.
70. Schmidt E, van Hogerwou G, van der Gaag R, et al. Site dependent effects of experimental hypo- and hyperthyroidism on resident macrophages in
extra-ocular muscles of rats: a quantitative immunohistochemical study. J Endocrinol 1992;135(3):485-493.
71. Dallow R. Evaluation of unilateral exophthalmos with ultrasonography: analysis of 258 consecutive cases. Laryngoscope 1975;85:1905-1919.
72. Coleman DJ, Carroll FD. A new technique for evaluation of optic neuropathy. Am J Ophthalmol 1972;74: 915-920.
73. Coleman DJ, Carroll FD. A new technique for evaluation of optic neuropathy. Trans Am Ophthalmol Soc 1972; 70:154-163.
74. Karaali K, Senol U, Aydin H, et al. Optic neuritis: evaluation with orbital Doppler sonography. Radiology 2003; 226(2):355-358.
75. Kenney AH. Ultrasound in clinical ophthalmological diagnosis. Trans Pac Coast Otoophthalmol Soc 1967;48:89.
76. Akarsu C, Tan FU, Kendi T. Color Doppler imaging in optic neuritis with multiple sclerosis. Graefes Arch Clin Exp Ophthalmol 2004;242(12):990-994.
77. Rizzo JF III, Andreoli CM, Rabinov JD. Use of magnetic resonance imaging to differentiate optic neuritis and nonarteritic anterior ischemic optic
neuropathy [See comment]. Ophthalmology 2002;109(9):1679-1684.
78. Cornblath WT, Quint DJ. MRI of optic nerve enlargement in optic neuritis. Neurology 1997;48(4):821.
79. Kupersmith MJ, Alban T, Zeiffer B, et al. Contrast-enhanced MRI in acute optic neuritis: relationship to visual performance. Brain 2002;125(Pt
4):812-822.
80. Fazzone HE, Lefton DR, Kupersmith MJ. Optic neuritis: correlation of pain and magnetic resonance imaging. Ophthalmology 2003;110(8):1646-1649.
81. Jackson A, Sheppard S, Johnson AC, et al. Combined fatand water-suppressed MR imaging of orbital tumors. Am J Neuroradiol 1999;20(10):1963-1969.
82. Hickman SJ, Brierley CM, Brex PA, et al. Continuing optic nerve atrophy following optic neuritis: a serial MRI study. Mult Scler 2002;8(4):339-342.
83. Blodi F, Gass J. Inflammatory pseudotumors of the orbit. Trans Am Acad Ophthalmol Otolaryngol 1967;71: 303-323.
84. Hogan M, Zimmerman L. Ophthalmic Pathology Philadelphia: WB Saunders; 1962:763.
85. Rootman J. Why “orbital pseudotumour” is no longer a useful concept. Br J Ophthalmol 1998;82(4):339-340.
86. Atlas S, Grossman R, Savino PJ, et al. Surface-coil MR of orbital pseudotumor. Am J Roentgenol 1987;148(4): 803-808.
87. Hardman J, Halpin S, Mars S, et al. MRI of idiopathic orbital inflammatory syndrome using fat saturation and Gd-DTPA. Neuroradiology
1995;37(6):475-478.
88. Asao C, Korogi Y, Hotta A, et al. Orbital pseudotumors: value of short inversion time inversion-recovery MR imaging. Radiology 1997;202(1):55-59.
89. Secil M, Soylev M, Ada E, et al. Orbital varices: imaging findings and the role of color Doppler sonography in the diagnosis. Comput Med Imaging Graph
2001;25(3): 243-247.
90. Yoshimoto M, Matsumoto S. Orbital varix rupture during cataract surgery. J Cataract Refract Surg 2004;30(3): 722-725.
91. Huna-Baron R, Setton A, Kupersmith MJ, et al. Orbital arteriovenous malformation mimicking cavernous sinus dural arteriovenous malformation. Br J
Ophthalmol 2000;84(7):771-774.
92. Lakits A, Prokesch R, Scholda C, et al. Helical and conventional CT in the imaging of metallic foreign bodies in the orbit. Acta Ophthalmol Scand
2000;78(1):79-83.
93. Boncoeur-Martel M, Adenis J, Rulfi JY, et al. CT appearances of chronically retained wooden intraorbital foreign bodies. Neuroradiology
2001;43(2):165-168.
94. Papadopoulos A, Fotinos A, Maniatis V, et al. Assessment of intraocular foreign bodies by helical-CT multiplanar imaging. Eur Radiol
2001;11(8):1502-1505.
95. Moin M, Kersten R, Bernardini F, et al. Spontaneous hemorrhage in an intraorbital arteriovenous malformation. Ophthalmology 2000;107(12):2215-2219.
96. Atalla M, McNab A, Sullivan TJ, et al. Nontraumatic subperiosteal orbital hemorrhage. Ophthalmology 2001; 108(1):183-189.
> Back of Book > Appendices
Appendices
APPENDIX A: ULTRASOUND VELOCITY IN TISSUE

Techniques for measuring tissues to determine their inherent speed of sound conduction have been described by Jansson and by Oksala. A method that we have
found reliable uses a knowledge of the velocity of sound in saline to compute tissue velocity. A transducer is fixed in a saline bath, and the time for the sound beam
to traverse a given thickness of tissue is measured (Figure A.1). In this technique, a measurement of the sound transmission time from transducer to anterior tissue
surface is made (T1), as well as the time from the anterior surface of the tissue to the posterior surface of the tank (T2). When the tissue is then removed, the
transmission time is measured from the transducer to the posterior surface of the tank (T3). Subtraction of the initial measurement of transducer to anterior tissue
(T1) from the new total measurement (T3) will provide the total thickness of fluid that has “replaced” the tissue. With knowledge of the velocity of sound in saline
and the measured ratio of transit times, (T3-T1)/T2, the velocity of sound in the “unknown” tissue is easily derived. These three measurements thus allow
computation of the tissue velocity. (Velocity and speed of sound are often used interchangeably, but, technically, velocity is a vector and speed of sound is the
more correct scalar quantity.)
The identification and location of intraocular tumors are of the greatest importance diagnostically; therefore, many tumors initially require only a relatively gross
estimation of their size. Ultrasonography, however, can provide reproducible and highly accurate measurements of tumor elevation. The observation of intraocular
tumors is often enhanced by the use of ultrasound to determine the rate of growth or regression. In addition, calculation of optimum tumor radiation dosage and the
proper selection of external cobalt plaques are aided by accurate measurement of tumor height and cross section.
The height of a tumor can be adequately measured by either A- or B-scan methods. The two main problems in obtaining a proper measurement are (a) maximizing
the height of a tumor by aligning the transducer appropriately and (b) discerning the acoustic separation of tumor from underlying choroid or sclera.
The transducer must be carefully positioned to intercept the peak of the tumor while maintaining a normal or perpendicular orientation to the scleral base. The
examiner makes this visual correlation during the scanning procedure. On the B-scan, the photograph should be taken using gray scale to select the tissue plane of
tumorsclera separation. This plane is usually detectable by the high-amplitude leading edge of the sclera, but, occasionally, may be indistinguishable even on the
A-scan.
Because the accuracy required in measurements of tumor elevation prior to placement of external cobalt plaques is generally acceptable within half a millimeter, a
velocity correction factor for tumor tissue is not usually required. For determination of growth or regression, however, the comparative time for tumor traverse is the
essential fact, and conversion to millimeter measure is superfluous. Nevertheless, the tissue thickness can be computed by using a velocity constant. At present,
we use a velocity constant of 1,650 meters per second for choroidal malignant melanoma.
P.172
Figure A.1
P.173
APPENDIX B: AXIAL LENGTH MEASUREMENT FOR INTRAOCULAR LENS IMPLANT

Because most patients selected for intraocular lens implantation do not have good central vision with the eye being considered for surgery, visual axis
determination is not usually possible, and the optic axis measurement must be used. The special problems encountered in obtaining these measurements should
again be emphasized here.
First, the selection of suitable echoes requires practice and skill. The transducer must be aligned so that the amplitude of the anterior and posterior lens echoes is
maximized, while the distance from the cornea to the vitreoretinal interface spike is simultaneously maximized. Globe compression must absolutely be avoided
during this procedure.
Second, because the lens thickness is critical, a spurious, apparent anterior lens surface echo produced by the iris should be avoided by dilating the pupil, and the
true anterior lens echo should be carefully identified on the oscilloscope screen or photograph.
In calculating the axial length in cataract patients, we use the lens velocity of 1,629 meters per second because most cataractous lenses in our studies have a
lower density than normal lenses. If a thin lens is found on measurement, that is, a lens that measures less than 3.5 mm, it may indicate the more unusual clinical
situation of a “dense” or sclerotic lens in which the velocity of sound lies in a higher range of 1,660 meters per second. In this instance, we suggest supplying two
separate axial length measurements labeled “normal cataract” for the velocity computation using 1,629 meters per second and “dense cataract” for the velocity
computation using 1,660 meters per second. If the selection of the dioptric power of an intraocular lens would be affected by the difference in the computations,
then the surgeon, after examining the cataract at the time of surgery, could decide which value for the implanted lens is more applicable.
Third, after surgery the eye may have a flatter cornea than that in the presurgical state. Depending upon the surgical technique, this could indicate a corrective
factor in the selection of lens power.
Once the axial length is obtained, the selection of intraocular lens power can be determined by use of keratometry and a nomogram, such as that published by
Worst, or by calculation of the intraocular lens power using a formula, such as the SRK of Hoffer. The original Binkhorst formula for calculation of the intraocular
lens power required to render a given eye emmetropic is
where D = dioptric power in aqueous or vitreous (refractive index 1.336) of the intraocular lens
where r = radius of curvature of the anterior surface of the cornea in millimeters
where a = axial length in millimeters
where d = distance between the anterior vertex of the cornea and the intraocular lens in millimeters
With a programmable hand calculator, lens power for emmetropia or desired myopia or hyperopia can be quickly ascertained with one or more of the formulas
available.
P.174
P.175
APPENDIX C: SCAN REPORTS

Writing reports for ultrasound is similar to writing reports for other diagnostic modalities, such as radiology. Having a template is useful or, at least, an overall
concept of a general report. The report form that we use will be presented here because it may be helpful as a guideline.
One must recognize that writing a report and interpreting the scan are skills that require considerable experience to develop. In addition, and most importantly, an
absolute diagnosis or interpretation may not be possible with a single examination. Radiologists have developed a very useful, descriptive method that connotes
uncertainty, with such phrases as “consistent with but not diagnostic of” or “suggestive of,” “appears like,” and “______ cannot be ruled out.” We seldom use the
pathologist's terms of “pathognomonic,” although, some changes, such as retinal detachment, dislocated lens, or foreign body echoes, could merit such a term.
We list the tentative diagnosis or referral diagnosis at the top of the page under the patient's demographic information and date of examination. The interpretation
starts with an overall description of the eye as to its shape or dimension and whether or not it appears to be following normal contours. We then describe the
cornea and anterior chamber (at higher frequencies) and the presence or absence of lens or intraocular lens. The clarity of the vitreous is then described and
possible retinal or choroidal elevations. Pathology relevant to the cause of the referral is then described specifically as part of the ocular interpretation. If an
intraocular tumor is being examined, its height, cross section, and acoustic character from both the B-scan and A-scan are described in terms of being “most
consistent” with a specific type of tumor.
If the question is related to retinal detachment, hemorrhage, or a patient has had prior surgery with an encircling band or Ahmed valve or has had trauma, the
anatomic variations from normal are presented, including such specifics as location and extent of such change. Using measurements and meridional locations as
much as possible is important.
Following the ocular description, the retrobulbar area is then described. The appearances of the optic nerve shadow, the retrobulbar fat, extraocular muscles, and
orbital walls are described. The specific pathologies, the optic nerve dimensions, including the presence of inflammatory changes of the nerve, scleral-Tenon's
boundary, or muscles are noted and described. Apical lesions are often the most difficult to interpret, whereas intraconal lesions can be described in terms of their
consistency, shape, and internal characteristics. Following this descriptive we use interpretation, where we would briefly summarize the significant pathology and
attempt to relate the findings to the indication for referral. We always include a representative B-scan on the report form.
Two representative scan reports are shown on the next two pages.
P.176
ROBERT M. ELLSWORTH
OPHTHALMIC IMAGING AND ONCOLOGY CENTER
NEW YORK-PRESBYTERIAN HOSPITAL-WEILL CORNELL MEDICAL CENTER
525 EAST 68 TH STREET, NEW YORK, NY 10021, K810
(212) 746-2495; Fax (212) 746-8921
ULTRASOUND LABORATORY USG #: 12345
Date: 10-5-05
Name: RETINAL DETACHMENT History #:

Address:
D.O.B.:
Referred by: Dr. Coleman
Clinical History: Right eye: Vitreous hemorrhage R/O retinal

detachment
Interpretation of Ultrasonogram:
Contact B- and A-scan ultrasonography of the right eye reveals a phakic globe outline of normal overall contours and dimensions. The iris plane and posterior lens
capsule are identified; however, other aspects of the anterior segment are unable to be evaluated due to intervening lid echoes.
The vitreous is filled with low amplitude echoes that are mobile on kinetic scanning, consistent with the clinical impression of dense hemorrhagic debris. There are
interfaces of high reflectivity with attachment at the optic nerve head extending anteriorly that are damped on kinetic scanning, consistent with total retinal
detachment.
The retrobulbar echo pattern appears within normal limits with a normal optic nerve shadow and a normal fat outline.
Impression: Total retinal detachment, right eye. Dense vitreous hemorrhage.
D. Jackson Coleman, M.D.
P.177
ROBERT M. ELLSWORTH
OPHTHALMIC IMAGING AND ONCOLOGY CENTER
NEW YORK-PRESBYTERIAN HOSPITAL-WEILL CORNELL MEDICAL CENTER
525 EAST 68 TH STREET, NEW YORK, NY 10021, K810
(212) 746-2495; Fax (212) 746-8921
ULTRASOUND LABORATORY USG #: 12345
Date: 10-5-05
Name: MELANOMA History #:
Address:
D.O.B.:
Referred by: Dr. Coleman
Clinical History: Right eye: Elevated lesion
Interpretation of Ultrasonogram:
Contact B- and A-scan ultrasonography of the right eye reveals a phakic globe outline of normal overall contours and dimensions. The iris plane and posterior lens
capsule are identified; however, other aspects of the anterior segment are unable to be evaluated due to intervening lid echoes. Within the vitreous there are
scattered echoes of low reflectivity that are mobile on kinetic scanning consistent with mild debris within a posterior vitreous detachment.
There is an elevated lesion at the posterior pole inferotemporally that measures 5.02 mm in A-P dimension with a horizontal base of 13 mm and a vertical base of
10 mm. The lesion appears solid on B-scan and A-scan attenuation pattern demonstrates a high amplitude leading edge that retires quickly to baseline with low
internal reflectivity, acoustically consistent with melanoma. The retina appears to be in place. The retrobulbar echo pattern appears within normal limits with a
normal optic nerve shadow and a normal fat outline.
Impression: Elevated lesion at the posterior pole inferotemporally right eye, acoustically consistent with melanoma. Mild debris within a posterior vitreous
detachment. Retina appears to be in place.
D. Jackson Coleman, M.D.
P.178
P.179
APPENDIX D: TECHNIQUES OF ULTRASONIC EXAMINATION

Ultrasonic evaluation of the eye can be performed with either contact or immersion techniques. The contact method is the simplest and most direct way of viewing
the eye. In the contact method, the A- or B-scan transducer is applied to the closed lid, with methylcellulose gel to provide an acoustic couplant, and the globe is
systematically searched (Figure D.1; see also DVD). Use of proparacaine eye drops to anesthetize the surface of the globe can allow direct application of the
transducer to the eye for improved resolution. For more complete evaluation, an immersion method is preferred, and when the anterior segment is to be examined,
an immersion system is required. Regardless of the equipment used, a water bath technique may form part of the examination.
In patients with recent ocular trauma, every attempt should be made to maintain sterile technique. We do not sterilize our transducer, but it is cleaned with alcohol
or immersed in an approved antibacterial wash prior to scanning. Alternatively, a sterile latex sheath can be placed over the end of the transducer; we generally
use a latex sheath cover with saline between the transducer and the sheath to provide a cushion so that no significant pressure is placed on the eye, and a benefit
of a standoff is obtained with contact B-scanning (Figure D.2; see also DVD). This method can also be used to scan infants, without the need for anesthesia use.
We have found that a description of the technique to the patient allays any fears he or she might have and encourages cooperation. The explanation is of the
following type:
This test is called an ultrasonogram. We use sound waves that are much the same as my voice but have a higher frequency, to provide echoes from the tissues
within your eye, much like sonar is used to map out an ocean floor. The examination is painless. You do not feel the sound, and it does not cause any tissue
damage. We use a water bath around your eye to conduct the sound. The water is a sterile salt solution, and the feeling will be much the same as if you were to
open your eye under water while swimming. We will put a drop of anesthetic into your eye, so it will not be uncomfortable, and we will hold your eyelids open so
you do not have to exert any effort.
Figure D.1 Contact method of ultrasound examination. Methylcellulose coupling gel is used to provide an
acoustic couplant, and the transducer is gently placed on the globe (see also DVD).
Figure D.2 Use of a sterile latex sheath with saline can provide an ideal standoff for examining infants or
severely traumatized eyes. This can be used with A- or contact B-scan equipment (see also DVD).
With the patient supine on the examination table, we place a plastic Steri-drape around the eye to provide the container for the water bath. The Steri-drape is an
ocular drape that has a central opening so that no plastic comes between the transducer and the eye. The drape may be applied directly to the skin, but we have
found that perspiration, oily skin, or heavy makeup prevents a water-tight seal. Therefore, we routinely elect to paint a ring of collodion around the patient's eye,
using a cottontip swab. Leakage of saline occasionally occurs inside the nasolabial fold or between the eyebrows, and care should be taken to allow at least a
quarter-inch of plastic material to cover these areas. The collodion should never be passed directly over the patient's open eye, and, for extra caution, the patient is
instructed to keep his or her eyes closed.
Once the drape is in position, a metal hoop is centered over the patient's eye. The edges of the drape are brought up through the hoop, but the hoop is not
clamped. A drop of proparacaine is placed in the patient's eye, and a sterile Barraquer speculum is then gently inserted to hold the eyelids open. On occasion, the
patient will move his or her head during this process, and, if the drape has been clamped, it might be torn loose.
Following the insertion of the speculum, the clamp is used to secure the edges of the plastic, and sterile warmed
P.180
saline is poured into the edge of the drape but not directly onto the eye. We preheat the saline to body temperature to minimize discomfort to the patient. For the
experienced examiner, this initial patient preparation usually takes 2 minutes.
The transducer is then swung into position above the eye (Figure D.3; see also DVD), and the transducer is dipped to just below water lever. Horizontal scans are
then made across the eye, and the display is observed. Vertical scans are used in selected cases to better portray the situation. The two-dimensional pictures on
the oscilloscope provide a real-time monitor for the procedure. The examiner continually observes the A- and B-scan traces to abstract maximum acoustic
information.
By changing transducer frequencies, certain features more amenable to greater resolution (higher frequency and/or focused transducers) or greater sensitivity
(lower sensitivities and/or unfocused transducers) may be portrayed.
The total examination time varies according to the difficulty of the case. A simple precataract extraction screening procedure can take as short as 4 or 5 minutes,
whereas a thorough tumor differentiation study using multiple frequencies can last up to 20 minutes. The total examination time should not exceed this period,
because the anesthetic becomes ineffective, and the patient becomes uncomfortable.
Figure D.3 The plastic Steri-drape water bath system developed by Coleman for routine immersion scanning.
An ocular Steri-drape is placed around the eye, and approximately 400 cc of normal saline solution provide an
easily accessible acoustic view of the eye and orbit. The transducer is then placed above the patient's eye (see
also DVD).
Rarely does a patient not readily cooperate, but, if this is the case, contact scanning conducted through the lids on A- or B-mode alone is used. In the examination
of children, we prefer to use anesthesia up to age 5 years. Diazepam (Valium) 10 mg 1 hour before the procedure has been useful for children ages 5 through 8.
> Back of Book > Color Plates
Color Plates
Figure 2.22 (see black and white image)











Ultrasoungraphy of The Eye

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ultrasoungraphy of The Eye

Uploaded by

Copyright:

Available Formats

Ultrasonography of the Eye and Orbit

> Front of Book > Authors

D. Jackson Coleman MD, FACS

Ronald H. Silverman PhD

Frederic L. Lizzi EngScD*

Dan Z. Reinstein MD, FRCSC, DABO

Suzanne W. Daly BSN, RDMS, CRNO

CONTRIBUTORS FOR DVD

Residual Perfluorocarbon Bubbles

> Front of Book > Front Matter

Omar Ibn Al-Halif

> Front of Book > Preface

> Front of Book > Dedication

This book is dedicated to

Frederic L. Lizzi, EngScD

> Front of Book > Acknowledgments

> Table of Contents > 1 - Physics of Ultrasound

GENERATION AND DETECTION OF ULTRASOUND

Figure 1.2. Schematic representations of molecular configuration in a piezoelectric material illustrating

TABLE 1.1 Reported Mean Velocities of Ultrasound in Ocular Tissues

Cornea 1,639 m/sec 1,632 22 4 Chivers9

Sclera 1,744 22 4 Chivers9

Vitreous 1,532 m/sec 1,508 22 4 Chivers9

Lens 1,641 m/sec 1,548 22 4 Chivers9

Cataractous lens 1,629 37 — Coleman15

Water 1,524 37 — Willard 16

Aqueous humor 1,532 37 — Jannson12

Fat 1,476 24 — Frucht17

its compression regions and lengthening its rarefaction regions.

REFLECTION AND SCATTERING OF ULTRASOUND

axial resolution (mm) = 0.75 T

the assumed speed of propagation, and te is time after excitation.

BIOLOGIC EFFECTS OF HIGH-INTENSITY ULTRASOUND

FDA EXPOSURE INDICES

Spatial-peak pulse-average intensity: ISPPA.3

Spatial-peak temporal-average intensity: ISPTA.3

Mechanical index: MI = Pr.3/fc 0.5

Thermal index: TI = Wfc/210

MEASUREMENT OF ULTRASONIC BEAM PARAMETERS

1. Mundt GH, Hughes WF. Ultrasonics in ocular diagnosis. Am J Ophthalmol. 1956;41:488-498.

> Table of Contents > 2 - Ultrasonic Systems

SCAN CONVERTER AND DISPLAY

B-MODE IMAGE GENERATION

B-MODE IMAGE QUALITY

TABLE 2.1. Types of B-mode Artifacts

Velocity differences Displacement artifact Contour distortion

Multiple reflections Surface duplication

Saturation Obliteration of texture

Saturation (occurring with texture enhancement) “Swiss cheese” artifact

Inadequate superposition Blurring and duplication

DIGITAL IMAGE PROCESSING

VERY HIGH FREQUENCY ULTRASOUND/ULTRASOUND BIOMICROSCOPY

LINEAR ARRAY SYSTEMS

High frame rate

OTHER SCAN MODES

TISSUE HARMONIC IMAGING

SPECTRAL PARAMETER IMAGING

1. Mundt GH, Hughes WF. Ultrasonics in ocular diagnosis. Am J Ophthalmol 1956;42:488-498.

> Table of Contents > 3 - Ocular Diagnosis

ultrasound instrument), but it was never marketed (Figure 3.3) (32).

OCULAR DIAGNOSTIC TECHNIQUES

TABLE 3.1 Diagnostic Parameters

Gross Morphologic Features