GASTROENTEROLOGY 1989;96:1058-62

Experimental Evaluation of an
Endoscopic Ultrasound Probe: In Vitro
and In Vivo Canine Studies

F. E. SILVERSTEIN, R. W. MARTIN, M. B. KIMMEY, G. C. JIRANEK,

D. W. FRANKLIN, and A. PROCTOR
Departments of Medicine and Anesthesiology and Center for Bioengineering, University of
Washington, Seattle, Washington

We developed an endoscopic echo probe that can be endoscope. The systems are expensive (-$100,000).
passed via the biopsy channel of a flexible fiberoptic Rigid tip sections limit endoscopic maneuverability.
or video endoscope with ti 3.5-mm channel. The They are difficult to orient endoscopically because of
probe moves along the gastrointestinal wall under oblique or side-viewing optics, which also make
direct endoscopic vision. The translational scan- passage into the colon impossible. The ultrasound
ning action is sensed by a position potentiometer image can be difficult to interpret because the endo-
and combines with the ultrasonic B-mode echoes to scopist or ultrasonographer is uncertain about the
produce a cross-sectional image of the wall. The orientation of the image plane inside the body. If
system uses an ultrasound frequency of 20 MHz to either the optical or ultrasound components fail, the
produce high-resolution images. The device was instrument is no longer functional and must be
used to image canine gastrointestinal tissue in vitro repaired or replaced. This is especially a problem
and in vivo during endoscopy. Ultrasound images of because the combined ultrasound endoscope is ex-
the gut wall correlate with histologic structure. This pensive, thereby limiting the possibility of having
probe overcomes some of the problems associated multiple instruments available. These combined
with the combined ultrasound endoscopes now in ultrasound endoscopes have diameters of >l.O cm
use. Usi.? of the probe with video endoscopy allows and cannot be passed through strictures. Finally, if a
the endoscopic and ultrasound images to be dis- lesion is detected during routine endoscopy, the
played side by side, simplifying coordination of conventional endoscope must be exchanged for the
ultrasound endoscope.
application of the two techniques.
In this paper we report the initial use of an
ultrasound system that separates the ultrasonic func-

U ltrasound
doscopes
the intestinal
transducers
are increasingly
wall (l-13).
combined with fiberen-
being used to image
The transducer can be
tion from the endoscope,
(EEP). Initial studies are described
the endoscopic echo probe
that use this new
ultrasound system in excised animal gastrointestinal
placed adjacent to the mucosa, avoiding the need for tissue and at endoscopy in dogs.
deep penetration. Depth of penetration is inversely
proportional to the ultrasound frequency, whereas
Methods
resolution is proportional to frequency (14). By plac-
ing the transducer adjacent to the mucosa, higher System Description
frequencies can be used, improving resolution.
The ultrasound system consists of a probe, a probe
Ultrasound endoscopes currently incorporate either movement translator device for attachment to the endo-
an ultrasonic mechanical sector scanner or a linear scope, and an ultrasound generator processor and display
array of ultrasound transducer elements into the (15). The probe comprises a 1.8-mm-diameter XI-MHZ
endoscope tip (l-5). These combined systems can transducer element (PZTS piezoelectric material) mounted
demonstrate the depth of malignant invasion of
intestinal wall neoplasms and localize intramural
and extramural gastrointestinal mass lesions (6-13).
Abbreviation used in this paper: EEP, endoscopic echo probe.
There are problems in combining an ultrasound C: 1989 by the American Gastroenterological Association
imaging system with the other requirements of an 0016-5085/89/$3.50
April 198Y ENDOSCOPIC L~LT’KASOUNi) PROBE 1050

Figure 1. Miniature endoscopic ultrasound probe seen protrud-

ing from the channel of a standard endoscope.

Figure 2. A removable translator device is attached to the endo-

on a 2-m catheter with a 2.9-mm diameter, allowing scope biopsy channel. This device IS secured to the
catheter when the catheter tip is protruding from the tip
passage down a 3.5-mm or larger biopsy channel of a
of the endoscope. The linear position translator is
flexible endoscope [Figure 1). The transducer element is
located on this attachment and senses catheter move
recessed into the side of the catheter tip by 1 mm. The ment. The dial for rotation of the catheter tip is also
transducer sends and receives the ultrasound signal in an located on this attachment.
axis perpendicular to the axis of the catheter. The catheter
tip can be rotated through multiple 360” arcs both clock-
wise and counterclockwise to align the transducer so that
to acquire the ultrasonicinformation and format it into the
the ultrasound beam is perpendicular to the intestinal
image. The computer can be used to manipulate the image.
wall. The tip of the probe rotates but the carrying sheath
For example, the image can be expanded, compressed,
does not. Rotation is achieved by turning a cable inside the
stored. and structures on the image can be measured. The
carrying sheath. This rotation is manually controlled by
image itself is displayed on a video monitor [Panasonic
the endoscopist using a dial attached to the entrance of the
WV 5370) and videotaped (Panasonic AG 6300).
biopsy channel. A mark on the tip of the probe indicates
the transducer location to allow visual alignment of the
transducer toward the wall. In Vitro Imaging
The catheter is moved in and out of the channel under
endoscopic vision. The tip moves along the gastrointesti- Before in vivo use of the probe, studies were
nal wall over a distance of 2.0 cm. This movement is performed in vitro to learn how to interpret 20-MHz
sensed by a linear position translator attached to the ultrasound images of the canine intestine. The 20-MHz
catheter outside the endoscope (Figure 2). The translator ultrasound system was initially tested by mounting a
directs the x-axis displacement of the ultrasound image on 20-MHz transducer. identical to the EEP transducer, on a
the display monitor. The y-axis of the image is related to micropositioner. Tissue specimens examined included
the depth in the tissue at which an echo is produced. The canine esophagus, stomach, and colon. We compared
magnitude of the echo controls the brightness of each dot ultrasonic images of the specimens with the histologic
on the image. In this way a B-mode image is produced sections of the same tissue planes using a previously
when the probe is moved along the intestinal wall (Figure described method (16).
3). Field size of the image is 2.0 cm wide by 2.0 cm deep:
axial resolution is -0.21 mm and the lateral resolution is
In Vi\,0 Imaging
0.45 mm.
Individual images are acquired as quickly as 0.1 s. Each Fasting adult dogs were examined endoscopically
time the probe is passed in or out over tissue, the previous under general anesthesia after a colon cleansing prepara-
image is updated. Scans are made at several rotational tion. Eleven endoscopic sessions were performed in 5
orientations until the optimum image is obtained. The dogs. The esophagus, stomach. duodenum. and colon were
image can be recorded on v,ideotape. examined with a forward-viewing endoscope [Olympus
The ultrasound generator and processor system is con- TGF-2D). Ultrasound images of the gastrointestinal wall
figured around a frame grabber (Coreco Oculus 200, Coreco were obtained by passing the probe along mucosa under
Corp., Quebec, Canada) that is connected to a microcom- endoscopic vision. The catheter tip was rotated until the
puter (IBbI-AT). The frame grabber allows the acquisition beam was perpendicular to the mucosal surface. Acousti-
and storage of an image so that it can be displayed and cal coupling was achieved either by flooding the area of
recorded with standard television video equipment. An interest with water or by placing a water-soluble gel in the
electronic transmitter and receiver as well as other cir- l-mm recessed area of the transducer before passing the
cuitrv function with the frame grabber and the transducers catheter down the endoscope biopsy channel. In 1 dog, a
1060 SILVERSTEIN ET AL. GASTROENTEROLOGY Vol. 96, No. 4

ECHO PROSE+/ /

ENDOSCOPE

. --.

Figure 4. Images obtained in vitro with the ZO-MHz transducer.

The ultrasound image (top] shows the layers of the wall
as seen ultrasonographically: m is made up of two
layers, the first from the surface of the mucosa, the
second from the remainder of the mucosa; sm is the
submucosa; mp is the muscularis externa; and s is the
subserosal fat and fibrous tissue. The correlation with
the histologic section of the same tissue (bottom] is
evident. The bars represent 3-mm distance.

CROSS-SECTION OF ESOPHAGUS Images of normal esophagus, stomach, duodenum,

Figure 3. Use of the endoscopic echo probe at endoscopy is and colon were then obtained using the ultrasound
illustrated. The target is identified visually with the probe at endoscopy in anesthetized dogs (Figure 5).
endoscope. The probe is passed over the target to obtain
Images were of similar resolution as those obtained
the ultrasound image of the wall and underlying struc-
tures. in vitro. Several steps were necessary to obtain
optimal images with the endoscopic probe: mucus
and debris had to be cleaned from the target area;
video endoscope was used (Fujinon Video Endoscope; adequate covering of gel or water was needed for
Fujinon, Wayne, N.J.). The endoscopic monitor was placed acoustical coupling; the transducer had to be rotated
adjacent to the ultrasound monitor in direct view of the precisely so that the ultrasound beam was orthogo-
endoscopist. nal to the mucosal surface. In all organs studied the
20-MHz EEP was able to penetrate all of the normal
wall layers.
Results
The video endoscope was found to be ideally
Images obtained in vitro with the 20-MHz suited for use with the EEP system. Having the
system correlated with the histologic sections of the endoscopic and ultrasound images side by side fa-
same tissue. The resolution of the system is demon- cilitated the endoscopic localization of the area to be
strated with an image of canine gastric wall. In the imaged and the control of the EEP to orient the
dog stomach five layers of the gastric wall could be transducer toward the target (Figure 6). This allowed
imaged (Figure 4). With minor corrections for inter- the optimal ultrasound image to be obtained after
face echoes, the first echogenic layer represents the two or three passes of the probe.
interface of fluid bathing the tissue with the mucosa;
the second echo-poor layer represents the rest of the
Discussion
mucosa; the bright highly echogenic central layer
represents submucosa; the echo-poor deep layer rep- The ultrasound probe system offers several
resents the muscularis externa; and the outer echo- potential advantages over existing ultrasound endo-
genie layer represents echoes from subserosal fat. scopes for gastrointestinal imaging. The probe can be
Similar correlation of the ultrasound images with used with standard, large channel, upper endo-
histology of specimens of canine esophagus, stom- scopes, and colonoscopes and can be sterilized in
ach, and colon were obtained. ethylene oxide. It has the potential to be simpler and
April 1989 ENDOSCOPIC ULTRASOUNU PROBE 1061

less expensive than a dedicated ultrasound endo-

scopic system. Thus, several probes can be available,
making it possible to examine consecutive patients
without delay. If a probe requires gas sterilization or
if a probe breaks, another can be used and the system
remains functional.
Attachment of the ultrasound transducer to a re-
movable probe will allow the use of various ultra-
sound frequencies for different indications. High-
frequency probes (20 MHz) give excellent resolution
of the mucosa and the normal wall. However, a
lower frequency probe (10 or 15 MHz) will be
needed to achieve the deeper penetration necessary
for study of malignant invasion of the wall by a
sessile mass. Even though lower frequency probes
reduce resolution compared with the higher fre-
quency probes, the resolution should be better than
with transcutaneous ultrasound transducers. Cur-
rently available ultrasound endoscopes are reported
to detect malignant invasion with frequencies of 5-
10 MHz (6,7,10,12). Development of lower frequency
endoscopic ultrasound probes is feasible but re-
quires modification of the electronic ultrasound sys-
tem and possibly new probe tip designs.
It may take less time to do an examination of an
abnormal area with the probe system than with the
combined systems. The probe can be passed down
the channel of the diagnostic endoscope during
routine endoscopy to scan an abnormal area of the
Figure S. Images obtained in vivo in the esophagus [A), stomach gastrointestinal wall. Further, it is possible to actu-
(B), duodenum (C), and colon (D) during canine endos-
ally watch the area being imaged with endoscopy as
copy. m is the mucosa; sm the submucosa; mp the
muscularis propria; and s the subserosal fat and fibrous
the ultrasound is applied. With the commercial
tissue. Acoustical gel or water was used to couple the ultrasound endoscopy systems, it is not possible to
transducer to the mucosal surface. perform both functions simultaneously. Endoscopic

Figurr ti. Videoendoscopic image (right] is seen on a monitor next to the ultrasound image (left) made with the EEP in the car
esophagus. In the video image the line on the probe tip between the black and white areas is opposite the transdu cer.
Adjustment of this line permits orientation of the transducer so that the ultrasound beam is perpendicular to the target. On the
leit the ultrasound image is displayed twice. one above the other. The layers of the esophageal wall XC set:n (arrow).
1062 SILVERSTEIN ET AL. GASTROENTEROLOGY Vol. 96, No. 4

vision is lost when fluid is placed in the lumen or a to learn whether this potential will be fulfilled in
balloon is inflated around the transducer to obtain clinical and research applications.
acoustical coupling.
There are some disadvantages. The probe has a References
smaller acoustical field of view than the mechanical
1. DiMagno EP, Regan PT, Clain JE, et al. Human endoscopic
sector scanners, which have a 360” acoustical field of
ultrasonography. Gastroenterology 1983;83:824-9.
view. The penetration of the probe is limited to 2.0 2. Silverstein FE, Giuliani D, Daigle R. Endoscopic ultrasound.
cm, whereas the combined lower frequency systems Acta Endoscopica 1983;13:1-9.
penetrate 6-8 cm. Therefore, at 20 MHz, the main 3. Gordon SJ, Rifkin MD, Goldberg BA. Endosonographic eval-
uation of mural abnormalities of the upper gastrointestinal
indication for the EEP is high-resolution imaging of
tract. Gastrointest Endosc 1986;32:193-8.
the wall itself. Future development of lower fre- 4. Tio TL, den Hartog Jager FCA, Tytgat GN. Endoscopic ultra-
quency EEP devices may permit examination of sonography of non-Hodgkin lymphoma of the stomach. Gas-
deeper structures. troenterology 1986;91:401-8.
Some aspects of the probe system still need to be 5. Caletti G, Bolondi L, Labo G. Ultrasonic endoscopy-the
gastrointestinal wall. Stand J Gastroenterol 1984:19(Suppl
defined. Ways to better couple the transducer with
102):5-8.
the mucosa need to be developed especially in an 6. Rifkin MD, McGlynn ET, Marks G. Endorectal sonographic
independent area where it is difficult to pool fluid. prospective staging of rectal cancer. Stand J Gastroenterol
This includes determining whether a balloon is 1986;21:99-103.
7. Saitog N. Okui K, Sarashina H. et al. Evaluation of echograph-
needed over the transducer. The probe and com-
ic diagnosis of rectal cancer using intrarectal ultrasonic ex-
bined systems need to be compared in specific amination. Dis Colon Rectum 1986:29:234-42.
applications. It may prove that each has a unique set 8. Strohm WD. Classen M. Benign lesions of the upper GI tract
of indications. For example, the probe may prove to by means of endoscopic ultrasonography. Stand J Gastroen-
be useful for studying endoscopically visible lesions terol 1986;21(Suppl 123):41-6.
9. Takemoto T, Aibe T, Fuji T. Okita K. Endoscopic ultrasonog-
of the intestinal wall, whereas the dedicated system
raphy. Clin Gastroenterol 1986;15:305-19.
may be more useful for extraintestinal lesions such 10. Tio TL, Tytgat GNJ. Endoscopic ultrasonography in the as-
as liver and pancreas and for screening large areas sessment of intra- and transmural infiltration of tumours in
for abnormalities. the oesophagus, stomach. and papilla of Vater and in the
detection of extraesophageal lesions. Endoscopy 1984:16:
Finally, the EEP seems well suited for use with
203-10.
video endoscopes. With the monitors located side by 11. Tio TL, Tytgat GNJ. Endoscopic ultrasonography of normal
side, endoscopists can quickly move their eyes from and pathologic upper gastrointestinal wall structure. Compar-
the endoscopic image to the ultrasound image. This ison of studies in vivo and in vitro with histology. Stand J
Gastroenterol 1986;21(Suppl 123):27-33.
seems to facilitate the endoscopic placement and
12. Tio TL, Tytgat GNJ. Atlas of transintestinal ultrasound. Aal-
movement of the EEP and the orientation of the smere, the Netherlands: Mur-Kostverloren. 1986.
ultrasound transducer to optimize the image. Both 13. Yasuda K, Nakajima M. Kawai K. Endoscopic ultrasonogra-
the video endoscopic image and the ultrasound phy in the diagnosis of submucosal tumor of the upper
image can ultimately be recorded simultaneously to digestive tract. Stand J Gastroenterol 1986;21(Suppl 123]:59-
67.
document both the endoscopically visible surface
14. Wells PNT. Propagation of ultrasonic waves through tissue.
structures and the ultrasound appearance of the wall In: Fullerton GD. Zagzebski JA. eds. Medical physics of CT
layers. and ultrasound: tissue imaging and characterization. New
In conclusion, the EEP system may provide a York: American Institute of Physics, 1980:367-87.
15 Martin RW, Silverstein FE, Kimmey MB. A 20.MH, ultra-
relatively simple method of applying ultrasound
sound system for imaging the intestinal wall. Ultrasound Med
during endoscopy to examine the intestinal wall. It Biol 1989 (in press).
may prove useful diagnostically, for example to 16. Kimmey MB. Silverstein FE, Haggitt RC, et al. Cross sectional
determine the presence and extent of inflammation, imaging method: a system to compare ultrasound. computed
tomography, and magnetic resonance with histologic find-
scarring, ulceration, and neoplasia and to differen-
ings. Invest Radio1 1987;22:227-31.
tiate intramural vs. extramural masses. The system
should also be useful therapeutically to guide endo-
scopic therapy by providing cross-sectional images Received November 2, 1987. Accepted November 10, 1988.
of the area being treated: for example, to determine Address requests for reprints to: Fred E. Silverstein, M.D.,
the depth of coagulation during laser therapy of Division of Gastroenterology. RG-24, University of Washington,
Seattle, Washington 98195.
neoplasms and to confirm coagulation of a visible
This study was supported by grants RO 1 AM 34814 and 5 T 32
vessel after treatment with the heater probe in a AM 07 113 from the National Institutes of Health.
patient with ulcer bleeding. However, the device is The authors thank Mike Nessly and Yehuda Sabag for technical
still experimental and further studies will be needed assistance.