Comparison of ultrasound and IV-DSA for carotid evaluation.

W J Zwiebel, C M Strother, C W Austin and J F Sackett

Stroke. 1985;16:633-643
doi: 10.1161/01.STR.16.4.633
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1985 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/16/4/633

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click Request
Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Stroke is online at:

http://stroke.ahajournals.org//subscriptions/

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013

 Comparison of Ultrasound And IV-DSA For 633

Carotid Evaluation
WILLIAM J. ZWIEBEL, M.D., CHARLES M. STROTHER, M.D., CHARLES W. AUSTIN, M.D.,

AND JOSEPH F. SACKETT, M.D.

SUMMARY Sixty carotid bifurcations in 34 symptomatic patients were examined prospectively with
ultrasound (continuous wave Doppler and high resolution, B-mode imaging) and intravenous digital sub-
traction angiography (IV-DSA). The overall quality of examination was better with DSA than with ultra-
sound. Imaging of the external carotid artery was particularly difficult with sonography. For evaluation of
the common and internal carotid arteries, eight percent of IY-DSA studies were poor or inadequate as
compared with 12% for B-mode imaging. Overall for detection of atherosclerotic plaque, high resolution B-
mode sonography was 84% sensitive and DSA 81% sensitive. When only the common and internal carotid
arteries were considered, the sensitivity of high resolution sonography improved to 93% and the sensitivity
of IV-DSA increased to 86%. Ultrasound (combined high resolution, B-mode sonography and C W Doppler)
correctly identified all six internal carotid occlusions in the series. While IV-DSA correctly identifiedfiveof
the six occlusions, the sensitivity for detection of lesions causing 70% or more stenosis was 95% for both
ultrasound and IV-DSA. Sensitivity for 50% or greater obstruction was 79% for ultrasound and 85% for
IV-DSA. Ultrasound sensitivity for greater than 50.9% stenoses rose to 87% when only the common and
internal carotid were considered while IV-DSA sensitivity remained at 85%. Specificity was good at all
levels of obstruction.
It may be concluded from this study that the accuracy of ultrasound and IV-DSA are quite similar for
evaluation of the carotid bifurcation and that either test is a satisfactory screening method for carotid
bifurcation atheromatous disease.
Stroke Vo! 16, No 4. 1985

THE CAROTID BIFURCATION is the most common
extracranial location for stroke related atherosclerotic
lesions. Plaque formation at this site is estimated to be
etiologically related to 60% of ischemic brain infarcts,
either directly, through occlusion of the internal carot-
id artery, or indirectly through embolic obstruction of
more distal intracranial vessels.1 The carotid bifurca-
tion is also the most surgically assessible portion of the
cerebrovascular system and thus, this region has be-
come the principal focus for development of diagnostic
procedures designed to detect and characterize athero-
sclerotic changes in these arteries.
Over the last four years intravenous digital subtrac-
tion angiography (IV-DSA) and duplex ultrasonog-
raphy (high resolution, B-mode ultrasonography with
combined Doppler flow assessment) have come into
widespread clinical use for assessing the cervical por-
tion of the carotid arteries. Both procedures have been
reported to be reasonably accurate in detecting stenotic
carotid lesions.'"5 These reports suggest that the accu-
racy of IV-DSA is slightly superior to that of ultra-
sound (table 1). At the time this study was undertaken,
there were no published reports in which IV-DSA and
ultrasound were evaluated side-by-side in the same
patient population. We, therefore, prospectively stud-
ied IV-DSA and duplex sonography, using arteriog-
raphy as a standard, for the purpose of assessing the
relative accuracy of these techniques in evaluating the
extracranial of the carotid bifurcation. Subsequent to
the collection of these data, a study has been reported5
From the Department of Radiology, University of Wisconsin School
of Medicine. Madison, Wisconsin.
Address correspondence to: Charles M. Strother, M.D., Department
of Radiology. University of Wisconsin, Clinical Science Center, 600
Highland Avenue, Madison, Wisconsin, 53792.
Received July 3, 1984; revision #1 accepted December 13, 1984.
in which the same population was examined with ultra-
sound and IV-DSA, but only a portion of the popula-
tion had both non-invasive procedures.
In the study reported herein, the parameters evaluat-
ed were the presence or absence of atherosclerotic
plaque and if present, the degree of stenosis it was
causing in the involved vessels. Our reasons for choos-
ing only these parameters are as follows:
1) The presence or absence of atherosclerotic
plaque is a useful indicator of the sensitivity of both
IV-DSA and sonography and represents a reasonably
definitive arteriographic endpoint.
2) Occlusive disease is a major associated finding
both in completed stroke and in the transient phenom-
ena (TIA and RIND) that may precede stroke.^
3) Accurate characterization of obstructive lesions
is important in deciding whether or not a patient may
benefit from a surgical procedure (e.g. carotid endar-
terectomy).
4) Arteriography has been shown to be inadequate
as a standard for ulcer detection; 10 " therefore, we
chose not to study ulceration in this series.
Materials and Methods
Study Population, Instrumentation and Technique
of Examination
Sixty carotid bifurcations in 34 symptomatic pa-
tients were examined prospectively with ultrasound
and intravenous digital subtraction arteriography (IV-
DSA). The results of these studies were compared to
those of standard selective carotid arteriography.
The average patient age was 65. Twelve bifurca-
tions were studied for asymptomatic carotid bruits, 12
for ipsilaterals clearly defined carotid territory tran-
sient ischemic attacks (TIA) or transient unilateral
visual symptoms (amaurosis fugax), 11 for ipsilateral

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013

634 STROKE
TABLE 1 Results, Duplex Carotid Sonography and IV-DSA
Duplex sonography
Wolver-
Wetzner1 son2 Glover5 Glover 5
et al et al et al et al
Stenosis:
grade £60% >50% £30%
sensitivity 94% 86% 93%
specificity 88% 89% 77%
accuracy 92% 88% 85%
Occlusion accuracy
n (bifurcations) 53 47 85
neurological complaints other than carotid territory
TIA's (usually completed stroke) and 6 because of
poorly defined neurological symptoms, usually associ-
ated with bruits. In 19 instances, no symptoms or signs
of carotid disease were attributable to the specific ca-
rotid bifurcations under investigation, but signs or
symptoms relevant to the contralateral bifurcation
were present in all of these cases.
Sonographic evaluations were performed using a
high resolution scanner with incorporated pulsed,
range-gated Doppler (HRS) and a separate continuous-
wave Doppler device (CWD). The HRS instrument*
utilized a multi-element anular array transducer with a
6-10 MHz frequency range. The in-vitro axial resolu-
tion of this device was reported by the manufacturer as
< 0.3 mm and lateral resolutions as 0.5 mm. The
dynamic range was reported as 70 dB.
CWD instrumentationt included both 5 MHz, deep-
ly focused and 10 MHz, superficially focused trans-
ducers. The 5 MHz transducer was coupled with a
position sensing device to produce a crude, two-di-
mensional image used to localize flow signals in the
vessel being evaluated. Carotid flow signals obtained
with the 5 MHz transducer were evaluated aurally and
with on-line frequency spectral analysis.
All ultrasound examinations were performed by one
of three experienced diagnostic medical sonog-
raphers.t The HRS studies were carried out with the
patient in a supine position with the neck extended and
turned to varying degrees. Four HRS views of each
carotid bifurcation were attempted approximating an-
terolateral, posterolateral, far posterior and transverse
projections. It was not always possible to achieve all
four views (see Results section). The entire HRS ex-
amination was recorded on three-quarter inch video-
tape.
The HRS instrument was equipped with a range-
gaited, pulse Doppler that was used principally to de-
tect flow and to facilitate identification of the internal
and external carotid arteries. All of the Doppler results
•BIOSOUND1" Biodynamics. Indianapolis, IN.
tPrototype of DOPSCAN™. SONOCOLOR'" Frequency Spectrum
Analysis. Carolina Medical Electonics. King, NC.
tTotal experience of our laboratory exceeds 3.000 CWD studies and
1.000 HRS examinations.
Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013
VOL 16, No 4, JULY-AUGUST 1985
IV-DSA
Wood1 Chilcote4 Glover5 Glover5
et al et al et al et al
270% 3:60% >60% £30% >70%
(CC.IC)
93% 93% 89% 85% 70%
93% 94% 94% 90% 100%
93% 94% 93% 88% 89%
85 98 200 56 56
presented in this report were derived from examina-
tions done immediately following the HRS study with
a separate continuous-wave instrument which was
used concurrently with the HRS device. Indirect
Doppler studies consisted of posterior orbital12 and
common carotid13 flow evaluations. The periorbital ex-
amination14 was also used when the other indirect tests
were equivocal. Direct cervical Doppler examination
consisted of bifurcation imaging from the clavicle to
the mandible followed by tape recording of the Dopp-
ler-shifted signals along the course of the vessels. The
power frequency spectrum was monitored on line by
the sonographer during Doppler signal recording.
Hard copy images of frequency spectra were made
routinely at several locations in the common external
and internal carotid and at all areas of increased flow
velocity.
Ultrasound (HRS and CWD) examination time
ranged from 45 to 120 minutes with an average time of
60 minutes. This time included recording of the pa-
tient's history, blood pressure measurement, neck aus-
cultations, as well as performance of the HRS and
CWD examinations.
The findings of the HRS and CWD examinations
were reviewed by an experienced radiologist prior to
release of the patient from the Department. In approxi-
mately 50% of bifurcations, he personally scanned and
recorded one or more areas with the HRS device, and
in approximately 20% of cases, he personally re-exam-
ined areas with the Doppler instrument. The initial
interpretation of this individual was not used for study
purposes; instead, at a later time, the tape recorded
HRS and CWD studies were independently and blindly
reviewed by one of the physicians involved in the
study.
Digital subtraction arteriography (DSA) was per-
formed with one of two prototype instruments under
development at the University of Wisconsin.§ The pa-
tients were examined in the supine position using a C-
arm fluoroscopic unit. Attempts were made in all cases
to obtain two views of each bifurcation at 90° angles
from one another. In many cases, three views were
obtained (two 45° oblique views and a straight AP
SPrototype. Philips Corporation, Einhoven, The Netherlands.
 CAROTID EVALUATION/Zwiebel
view). Injection was through a 5 French pigtail cath-
eter inserted through an antecubital vein with its tip
located either in the innominate vein or superior vena
cava. Contrast medium (76% Renografin) was injected
at the rate of 14—16 cc/second for a total volume of 40
cc per projection. The standard imaging sequence was
two exposures per second over a 9-12 second interval.
All DSA studies were conducted by one of the mem-
bers of the Neuroradiology Section. As with the sono-
graphic studies, the initial report of this individual was
not used for study purposes. Instead, the hard-copy
films were independently and blindly reviewed at a
later time by one of the study physicians. The time
required for the usual IV-DSA examination was be-
tween 30 and 45 minutes.
Methods of Tabulation and Correlation
The carotid arteriograms were interpreted retrospec-
tively by a single observer, having no knowledge of
either the patient's clinical history or the results of the
IV-DSA or sonographic evaluations. To insure consis-
tency, the same individual interpreted all arteriograms.
For ambiguous or difficult-to-interpret cases, his im-
pression was verified by a second interpreter.
Because more than one occlusive lesion or area of
plaque deposition was often present in each bifurca-
tion, lesions of the common external, and internal ca-
rotid arteries were independently tabulated ignoring
the relationship to the bifurcation as a whole.
The following parameters were assessed:
Image Quality — The quality of HRS and IV-DSA
visualization of each arterial segment (common, ex-
ternal or internal carotid) was judged according to
the criteria listed in table 2 and illustrated in figures 1
and 2.
Sensitivity for Plaque — For HRS and IV-DSA,
each vascular segment was designated as positive or
negative for the presence of plaque and these results
were compared with arteriography. Doppler findings
were not utilized in this assessment due to recognized
limitations in detection of non-occlusive plaque.15
Stenosis and Occlusion — Obstructive lesions were
classified based on percent of projected original lumen
diameter as < 50%, > 50% but < 70%, > 70%
stenosis, and occlusion. The maximal stenotic lesion
detected in each arterial segment (common, external,
or internal carotid artery) was tabulated for sonography
(combined HRS and CWD), IV-DSA and standard
arteriography.
HRS evaluation of obstructive lesions was accom-
plished through direct measurement of the residual
lumen from the video display monitor (fig. 3). Doppler
estimation of occlusive lesions was based on peak fre-
quency elevation, diastolic frequencies, turbulence
(fig. 4), and indirect Doppler tests such as the posterior
orbital examination or common carotid resistivity.16
For minor stenoses which produced relatively little
disturbance of flow, the HRS results were favored;
whereas, in most major obstructive lesions the Doppler
results were more heavily relied upon.
For IV-DSA and arteriography, the extent of an
Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013
et al 635
obstructive lesion was determined by comparing cali-
per measurements made at the stenotic and normal
portions of the vessel. When measurements of residual
lumen diameter obtained from the AP and lateral pro-
jections of the arteriograms were different, it was as-
sumed that the lumen was elliptical and the two mea-
surements were averaged. The normal diameter of the
carotid bulb was visually estimated, based on the di-
mensions of calcified plaque, when present, or on the
interpreter's judgment (fig. 5).
Results
Quality of Arterial Visualization
Data for HRS and IV-DSA visualization of the ca-
rotid vessels are presented in table 3. Overall, visual-
ization of the cervical portion of the carotid arteries
was considerably better with IV-DSA than with HRS.
Excellent or moderate quality visualization (grades 1
& 2) was achieved more frequently with IV-DSA and
there were fewer inadequate studies (9.5% DSA vs.
18% HRS). With IV-DSA the quality of visualization
of the external and internal carotid arteries was similar,
while HRS performed especially poorly for the exter-
nal carotid artery. Examination quality obtained with
HRS compared more favorably with IV-DSA when the
external carotid was excluded from tabulations. Visu-
alization of the common and internal carotid arteries
was poor or inadequate (grades 3 or 4) in twelve per-
cent of the HRS studies and in 8% of the IV-DSA
studies.
Factors affecting the quality of visualization with
HRS and IV-DSA have previously been reported15 l7
TABLE 2 Categories of Visualization of Carotid Vessels
Category Description
1A Arterial walls clearly defined
Excellent gray-scale resolution
Minimal or no artifacts
Vessels imaged in two longitudinal views at approxi-
mately right angles
Internal carotid imaged at least 2 cm beyond the bi-
furcation
IB Same as I A. but only one longitudinal view
2A Arterial walls satisfactorily defined but less sharply
than on IA
Satisfactory gray-sca!c resolution but of lesser quality
than 1A
Artifacts present which do not significantly limit di-
agnostic potential
Vessel imaged in two longitudinal views at approxi-
mately right angles
Internal carotid imaged at least 2 cm beyond the bi-
furcation
2B Same as 2A, but only one longitudinal view
3 Artery imaged, but not at diagnostic levels due to
poor definition of walls, lack of gray-scale resolu-
tion, or artifacts
4 Artery not imaged (includes cases in which at least 2
cm of the internal carotid were not seen)
636
and will not be discussed in detail. Briefly, for HRS
these include: 1) an inaccessible location of the carotid
bifurcation high in the neck beneath the mandible. 2)
Ultrasound attenuation in deep lying vessels. 3) Arteri-
al tortuosity. 4) Reflection of the incident sound beam
by massively calcified plaque. For IV-DSA studies,
the primary causes of suboptimal image quality were
subtraction misregistration due to patient motion and
poor vascular opacification most often as a result of
inadequate cardiac output.
Plaque Detection
Data for HRS and IV-DSA evaluation of atheroscle-
rotic plaques are presented in tables 4-6. Except in
isolated instances, these data demonstrate no substan-
tial difference in the accuracy of plaque detection.
Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013
STROKE VOL 16, No 4, JULY-AUGUST 1985
FIGURE 1. Grades of HRS visualization: A) Grade I. The
artery wall (arrow) is well defined and adjacent plaque is clear-
ly seen. Little or nor artifact is present. B) Grade 2. The artery
wall is less clearly defined and the amount of artifact is greater.
C) Grade 3. A large area of shadowing (S) from echogenic
plaque obscures a major portion of the vessel. EC = External
Carotid. IC = Internal Carotid.
Sensitivity for plaque detection was good for both pro-
cedures (84% HRS, 81% IV-DSA). Negative predic-
tive values were poor for both HRS and DS A (49% and
36% respectively), but the high proportion of false
negative studies as well as a high prevalence of disease
in the study population accounted for the poor negative
predictive values.18 Nineteen of 25 HRS and 27 of 30
IV-DSA false negative examination resulted from fail-
ure to detect plaques of minimal thickness (<20%
lumen diameter). Deletion of studies did not affect
appreciably the results obtained for plaque detection
(table 6), indicating that examination quality could not
be evoked as a cause for the relatively poor perform-
ance of these tests in this regard. The sensitivity for
plaque detection improved substantially for HRS
(from 84 to 93%) and slightly for IV-DSA (81 to 86%)
when findings in the external carotid artery were ex-
cluded from the tabulation (table 6). Several inter-
related factors probably account for this observation:
1) minor plaques, which are inherently more difficult
to detect than are moderate or severe alterations, were
proportionately more common in the external carotid
artery than in the common internal carotid artery. 2)
The complicated configuration and relatively small
size of the external carotid artery made plaque detec-
tion difficult.
 CAROTID EVALUATION/ZiW<?6e/ el al 637

TABLE 3 ESS and IV-DSA, Quality of Visualization

Common carotid External carotid Internal carotid Total
Grade HRS DSA HRS DSA HRS DSA HRS DSA
1A 75.0% 25.0% 27.07c 17.07c 28.0% 17.0% 43.0% 19.07c
IB 5% 47.0% 7.0% 35% 9.0% 46.0% 7.07c 43.07c
2A 12.0% 2.0% 23.0% 8.0% 28.0% 7.0% 21% 6.0%
2B 6.0% 20.0% 12.0% 26.0% 13.0% 21.0% 11,07c 23.0%
3 2.0% 3.0% 23.0% 12.0% 17.0% 7.0% 14.0% 7.0%
4 0% 3.0% 8.07c 2.0% 5% 2.07r 4.07c 2.07c

Although exclusion of observations related to the
external carotid arteries improved the sensitivity of
both procedures for plaque detection, it sharply de-
creased the negative predictive values for plaque to
27% for HRS and to 16% for IV-DSA. This occurred
because of a decrease in the number of normal vessels
that was far out of proportion to the reduction of false
negative results. The very high prevalence of disease
in the series (98%) renders meaningless calculations of
negative predictive value for plaque.
HRS and IV-DSA specificity for plaque was rela-
tively poor (77% each) for the entire series, reflecting
the incidence of false positive examination. For HRS,
all false positives occurred in the external carotid and
were in vessels with plaques of minimal thickness. All
IV-DSA false positives also occurred in the external
carotid artery. Two of these were minimal disease and
three were in ones having plaques of greater thickness.
False positive errors for HRS, as previously discussed,
related to difficulties in imaging the external carotid
artery. For IV-DSA, the availability of only one pro-
jection was a factor in two of the three major over-
estimations of plaque. Specificity for evaluation of
atherosclerotic plaques for both HRS and IV-DSA was
not improved when poor quality studies were excluded
from considerations, but increased dramatically to
100% when no attempt was made to evaluate the exter-
nal carotid artery. However, calculations of specificity
as well as the positive predictive values for plaque
detection are not reliable because only three normal
vessels remained in the series after withdrawal of the
external carotid findings.
Occlusive Disease
The results of HRS and IV-DSA for characterizing
occlusive iesions, as compared with those obtained
with standard arteriography, are presented in tables 7
and 8. Statistical comparison of these two modalities
for luminal narrowing of ^ 50, ^ 70% stenosis and
occlusion are tabulated in table 9 for all vessels and in
table 10 for the common and internal carotid arteries.
The reader is reminded that ultrasound (US) results
listed in these tables represent the combined HRS and
CWD findings as outlined in the Materials and Meth-
ods section.
The sensitivity of both modalities for obstructive
lesions causing a 50% or more decrease in lumen di-
ameter was good (79% US, 85% IV-DSA), but was
not at the 90% level generally preferred for a screening
test. Sensitivity rose to excellent levels (95%, each
method) for lesions causing a 70% or greater stenosis.
HRS detected all occlusions, whereas, one false nega-
tive occurred with IV-DSA. (Misregistration artifacts
were misinterpreted as intravascular contrast.)
The specificity as well as positive and negative pre-
dictive values for US and IV-DSA were excellent over-
all for lesions causing 50% or greater stenosis and also
were excellent in those vessels with either a 70% or
greater stenosis or occlusion. The exceptions are note-
worthy, however: 1) the positive predictive values for
IV-DSA detection of lesions causing stenosis of >
50% and > 50% < 70% levels was suboptimal (83%
and 72% respectively) due to a tendency with IV-DSA
to overestimate stenoses within these ranges; and 2) the
positive predictive value of US for detection of occlu-
sion was poor (60%), due to failure with this method to
distinguish between very severe stenosis and occlu-
sion. Aside from these differences, the results obtained
with US and IV-DSA are comparable.
As with plaque detection, improvement in US and
IV-DSA performance, with few exceptions, occurred
when the external carotid was excluded from tabula-
tion (table 10). Nonetheless, sensitivity for lesions
causing a 50% or greater steonsis remained at subopti-
mal levels (87% US, 85% IV-DSA). The limited posi-
tive predictive value of iV-DSA for stenosis of 70% or
greater and of ultrasound for occlusion persisted in
spite of exclusion of the external carotid data. These
areas of difficulty are not felt to be related directly to
external carotid imaging problems.

TABLE 4 Plaque Detection HRS TABLE 5 Plaque Detection IV-DSA

All vessels CC& IC All vessels CC & IC
n = 180 n = 120 n = 180 n = 120
Arteriography Arteriography Arteriography Arteriography
HRS Normal Plaque Normal Plaque HRS Normal Plaque Normal Plaque
Normal 17 25 3 6 Normal 17 30 3 16
Plaque 5 133 0 111 Plaque 5 128 0 101

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013

638 STROKE
FIGURE 2. A) Grades of IV-DSA visualization. A) Grade 1.
The vessel margins are sharply seen and contrast density is
excellent. B) Grade 2. Vessel margins are less well seen and
contrast is not as great as grade 1. C) Grade 3. The vessel is
vaguely seen but is obscured by low contrast density and regis-
tration artifacts.
The exclusion of poor quality (Grades 3 and 4) stud-
ies increased overall ultrasound sensitivity for lesions
causing 50% or greater stenosis from 79-84% (CC,
IC, EC: 26 true positive, 5 false negative), but overall
sensitivity for IV-DSA in lesions of the same severity
declined minimally from 85 to 83% (CC, IC, EC: 29
true positive, 6 false negative). These changes were
not significant.
Discussion
Quality of Examinations
IV-DSA suffered from considerably fewer non-di-
agnostic studies than HRS (18% and 9.5% respective-
ly) and displayed overall better examination quality
(table 4). Particular difficulty was noted with HRS
when attempts were made to image the relatively
small, branching external carotid artery. When the ex-
ternal carotid artery was excluded from consideration
the incidence of non-diagnostic HRS studies decreased
to 12%. Considering that atherosclerotic lesions occur
Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013
VOL 16, No 4, JULY-AUGUST 1985
most often in the common and internal carotid arteries
and that the external carotid artery is of unproven im-
portance in stroke pathogenesis, it appears that both
procedures offer satisfactory image quality for a
screening study. The results of HRS and IV-DSA in
our study for image quality are comparable to those
reported by others.3- *• "
Findings for both plaque detection and assessment
of stenosis indicate that the accuracy of ultrasound is
considerably more dependent on examination quality
than is that for IV-DSA; i.e., it appears that, on the
whole, more reliable information may be extracted
from poor quality IV-DSA studies than can be obtained
from poor quality ultrasound examinations.
Plaque Detection
The performance of HRS and IV-DSA for plaque
detection was comparable with the following excep-
tions. Both HRS and IV-DSA were imperfect in plaque
detection, and this problem was more severe for IV-
 CAROTID EVALUATION/Ziv;>fc<r/ el al 639

FIGURE 3. Technique for HRS stenosis measurement. The

original lumen diameter is indicated by large arrows and the
residual diameter by small arrows.

is of less importance as a cause of stroke than is more

DSA, considering that sensitivity with this modality
remained below 90% even when the external carotid
arteries were excluded from consideration. Minimal
thickness plaque was overlooked in the majority of all
false negative studies (both HRS and IV-DSA). This is
reassuring in that minor plaque involvement probably
severe disease. Considering intrinsic limitations of the
technique (reduced spatial resolution and available
levels of vascular opacification), it is to be expected
that IV-DSA would be suboptimal in detection of mini-
mal thickness plaque. The failure to detect minimal
thickness plaque with HRS occurred principally in the
external carotid artery, the method providing better
results with plaque detection in the common and inter-
nal carotid arteries. Other authors have found HRS
more sensitive than arteriography for detecting minor

TABLE 6 Statistical Summary, Plaque Detection — HRS and IV-DSA

HRS IV-DSA
Exclude Exclude
grades grades
All CC & IC 3 & 4* All CC & IC 3 & 4*
(n = 180) (n = 120) (n = 138) (n = 180) (n = 120) (n = 145)
Sensitivity 84% 95% 84% 81% 86% 82%
Specificity 77% 100% 76% 77% 100% 76%
Positive predictive value 96% 100% 96% 96% 100% 97%
Negative predictive value 46% 33% 50% 36% 16% 38%
False ( - ) 19 minimum plaque thickness 27 minimum plaque thickness
3 moderate plaque thickness 3 moderate plaque thickness
3 severe plaque thickness
*Poor or inadequate examinations excluded.
Severity of the plaque: most of false — were minimum thickness range. Minimum thickness <% diameter.

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013

640 STROKE VOL 16, No 4, JULY-AUGUST 1985

FIGURE 4. Method ofDoppler stenosis estimation. A) Frequency spectrum in the stenotic zone. Peakfrequencies (arrowhead)
reach 12 KHz (vertical scale) and diastolic frequencies (arrows) range from about 6 to Th KHz. B) Sonogram in the immediate
post-stenotic zone demonstrates considerable turbulence. The spectrum is completely "filled in" and the upper margin of the
tracing is indistinct. Compare with 4A.

FIGURE 5. Method of arteriographic stenosis measurement. A) Subtraction film from standard arteriogram. B)IV-DSAfilm.
The dashed line represent the interpreter's estimation of the original lumen contour. The same carotid bifurcation is illustrated
in A and B, but the projections are slightly different.

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013

 CAROTID EVALUATION/ZuvV/*?/ et al
TABLE 7 Correlation of Arleriography and Ultrasound for Oc-
< lustre Lesions
Arteriography
> 50 Occlu-
None < 50% < 70% a 707c sion
None 77 15 2 1 0 riography whenever occlusion is suspected.
< 50% occlusion 8 37 6 0 0
& 50 < 70% occlusion 0 0 14 0 0
& 70% occlusion 0 0 1 9 0 noninvasive studies may be less reliable at this level of
occlusion 0 1 0 3 6 disease than at the > 70% level. Sensitivity of US or
Total 85 53 23 13 6 IV-DSA for > 50% obstruction did not exceed 85% in
(n = 180 vessels).
plaque deposits.2 '8 This was not the case in our study.
The level of specificity for plaque with both HRS
and IV-DSA (77%) is disappointing but in consider-
ation of a screening test is of less concern than is a lack
of sensitivity. Furthermore, it is of importance that all
observed false positives involved the external carotid
artery as this vessel is not felt to be of primary concern
in the pathogenesis of stroke.
Because of the high prevalence of disease in our
patient population, the positive and negative predictive
values for plaque detection are not of significance. It is
likely that the negative predictive values for plaque
with HRS and IV-DSA would be much better in a
population having a larger proportion of normal
patients.18
Occlusive Disease
The primary objective of evaluation of the carotid
bifurcation screening is to determine the severity of
carotid atheromatous disease and thereby determine
whether or not surgery is a therapeutic option. As
noted in the introductory material, the relative degree
of luminal narrowing provides an approximation of the
severity of disease that is useful for clinical decision
making. Once it is decided that major atheromatous
disease is present at the carotid bifurcation, standard
arteriography is generally considered appropriate to
define as precisely as possible the condition of extra-
and intracranial vasculature. Some surgeons, howev-
er, may operate based solely on the results of non-
invasive tests, particularly intravenous IV-DSA.
in view of the importance of occlusive disease eval-
uation in clinical management, it is of particular inter-
est that our data suggest no substantial difference be-
tween ultrasound and IV-DSA for evaluation of
stenosis and occlusion. Both procedures are highly
sensitive and reasonably specific for severe obstructive
lesions (70% stenosis and complete occlusion), indi-
cating that either IV-DSA or ultrasound may be used
with a high degree of reliability in selecting patients
who may be surgical candidates on the basis of severe
carotid occlusive disease. Caution is necessary in some
areas, however. The relatively low positive predictive
value of IV-DSA for > 70% stenosis means that the
severity of disease may be overestimated in some pa-
Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013
641
tients. For ultrasound, inability to distinguish between
severe flow reducing stenosis and occlusion may be a
problem. It is also not clear that IV-DSA can accurate-
ly distinguish between very severe stenosis and occlu-
sion, and it may be necessary to employ standard arte-
For occlusive lesions at the 50% level, there is great-
er potential for error from negative studies, hence
clinical decisions about surgical candidacy based on
spite of exclusion of poor quality studies as well as
results of the troublesome external carotid. This find-
ing is somewhat disappointing. Carotid screening tests
ideally should be highly sensitive for obstruction >
50% because this is the generally accepted level for
hemodynamic significance in carotid stenosis. Intra-
observability and other measurement difficulties may
well have contributed to the lack of precision at this
level of stenosis. These factors would be expected to
more strongly affect the 50% level as compared with
70% stenosis or occlusion, because the endpoints for
measurement are considerably less clear at 50% than
for higher grades. The results presented herein for IV-
DSA sensitivity at the 50% are not as good as those
obtained in other published series (table 1). For exam-
ple, Wood et al3 reported 93% IV-DSA sensitivity for
> 50% stenosis in all carotid vessels, and Chilcote et
al4 found IV-DSA 89% sensitive for > 60% common
and internal carotid stenoses. Ultrasound results from
our series are comparable to those presented by Co-
marota, et al' 9 and Wolverson, et al2 for 50% lesions
but are not as good as these reported by Glover et al
who observed 93% sensitivity for > 30 and for > 70%
stenosis (table 1).
It must be recognized that the accuracy of non-inva-
sive studies depends on several factors including the
type of equipment used, the patient population and the
experience of the operators (especially for ultrasound).
Our results are based on fairly early IV-DSA instru-
mentation and technique as well as a less than optimal
ultrasound system in which Doppler and B-mode tech-
niques were used independently. It is possible that
superior results might be obtained with more advanced
instrumentation.
TABLE 8 Correlation ofArteriography and IV-DSA for Occlusive
Lesions
Arteriography
s 50 Occlu-
DSA None < 50% < 70% > 707c sion
None 76 25 1 0 0
< 50% occlusion 7 23 6 0 0
> 50 < 70% occlusion 1 3 12 1 0
1 11 1
2: 70% occlusion 2 4
occlusion 0 0 0 1 5
Total 85 53 23 13 6
(n = 180 vessels).
642 STROKE VOL 16, No 4, JULY-AUGUST 1985

TABLE 9 Analysis of Ultrasound* and IV-DSA Results for Occlusive Lesions

All vessels (n = 180 vessels)
a 50% occlusion a 70% occlusion Occlusion
Diameter — reduction US DSA US DSA US DSA
Sensitivity 79% 89% 957c 95% 100% 837c
Specificity 99% 95% 99% 967c 98% 997c
Positive predictive value 97% 83% 90% 72% 60% 837c
Negative predictive value 94% 95% 99% 99% 100% 997c
Number of lesions 42 19 6
Prevalence 23% 11% 37o
*High resolution sonography and continuous wave Doppler.

TABLE 10 Analysis of Ultrasound* and IV-DSA for Occlusive Lesions. CC and IC

50% occlusion 70% occlusion Occlusion
Diameter — reduction us DSA US DSA US DSA
Sensitivity 82% 857c 1007c 94% 1007c 837c
Specificity 99% 957c. 997c 95% 987c 997c
Positive predictive value 97% 92% 89% 767c 677o 837c
Negative predictive value 92%> 937c 100% 997o 1007c 997c
Number of lesions 22 11 6
Prevalence 187o 97< 5%
*High resolution sonography and continuous wave Doppler.
n = 120 vessels.

Conclusions 2. Wolverson MK, Heibe rg E, Sundaram M, Tantanasiv

Direct comparison of ultrasound and IV-DSA for
cervical carotid evaluation indicates relatively little
difference in accuracy or potentially clinical efficacy.
Although the quality of examination for IV-DSA was
superior to ultrasound overall, both procedures dem-
onstrated adequate examination quality for the assess-
ment of the common and internal carotids. HRS is
adequately sensitive for the presence of common and
internal carotid plaque. IV-DSA also is quite sensitive,
but displays a greater tendency to overlook minimal
plaque involvement as compared with HRS. Both pro-
cedures are highly sensitive for occlusive lesions that
reduce lumen diameter by 70% or more and i.e. either
study may be relied upon for selection of potential
surgical candidates at this level of atheromatous in-
volvement. For occlusive lesions at the > 50%, ultra-
sound and IV-DSA are only about 85% sensitive and
some caution is advised in clinical application of re-
sults at this level of disease due to the greater potential
for false negative diagnoses.
Considering that neither test exhibits clear superior-
ity over the other in terms of accuracy or potential
efficacy, other factors must be taken into account when
one is deciding which study is appropriate for a par-
ticular patient. For instance, if the likelihood of sur-
gery is high based on signs, symptoms and patient
condition, and the surgeon favors operating solely on
the basis of IV-DSA results, then IV-DSA would be a
preferable study in comparison to ultrasound.
References
I. Wetzner SM, Riser LC, Bezreh JS: Duplex ultrasound imaging:
vascular applications. Radiology ISO: 507-514, 1984
Shields JB: Carotid atherosclerosis: high-resolution real-time so-
nography correlated with angiography. AJR 140: 355-361, 1983
Wood GW, Lukin RR. Tomsick TA, Chambers AA: Digital sub-
traction angiography with intravenous injection: Assessment of
1,000 carotid bifurcations. AJR 140: 855-859, 1983
Chilcote WA, Modic MT, Pavlicek WA, Little JR, Furlan AJ,
Duchesneau PM, Weinstein MA: Digital subtraction angiography
of the carotid arteries: A comparative study in 100 patients. Radi-
ology 139: 287-295, 1981
Glover JL. Bendick PJ. Jackson VP, Becker GJ, Dilley RS, Holden
RW: Duplex ultrasonography, digital subtraction angiography, and
conventional angiography in assessing carotid atherosclerosis.
Arch Surg 119: 664-669, 1984
Harrison MJG, Marshall J: Angiographic appearance of carotid
bifurcation in patients with completed stroke, transient ischaemic
attacks, and cerebral tumour. Brit Med J 1: 205-207, 1976
Thiele BL, Young JV, Chikos PM, Hirsch JH, Strandness DE:
Correlation of arteriographic findings and symptoms in cerebrovas-
cular disease. Neurology 30: 1041-1046, 1980
Pessin MS, Duncan GW, Mohr JP, Poskanzer DC: Clinical and
angiographic features of carotid transient ischemic attacks. New
Eng J Med 246: 358-362, 1977
Harrison MJG, Marshall J: Prognostic significance of severity of
carotid atheroma in early manifestations of cerebrovascular dis-
ease. Stroke 13: 567-569, 1982
10 Lusby RJ, Ferrell LD, Ehrenfeld WK, Stoney RJ, Wylie EJ: Carot-
id plaque hemorrhage. Aurch Surg 117: 1479-1487, 1982
Eikelboom BC, Riles TR, Mintzer R, Baumann FG, DeFillip G,
Lin J, Imparato AM: Inaccuracy of angiography in the diagnosis of
carotid ulceration. Stroke 14: 882-892, 1983
12 Spencer MP: Technique of Doppler examination. [In] Spencer MP,
Reed JM (eds). Cerebrovascular evaluation with Doppler ultra-
sound. The Hague, Martinus Nijhoff, 1981, pp 77-80
13 Planilo T, Pourcelot L, Pottier JM, et al: Etude de La Circulations
Carotidienne par les Methodes Ultrasoniques et La Thermographie.
Rev Neural (Paris) 126: 127-141, 1972
14. Brokenbough EC: Periorbital velocity evaluation of carotid ob-
struction. [In] Bernstein EF (ed). Noninvasive diagnostic tech-
niques in vascular disease. St. Louis, C.V. Mosby, 1978, pp
212-220

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013

 CAROTID EVALUATION/Zit'/efoe/ et al 643

15. Zwiebel WJ. Austin CW. Sackett JF, Strother CM: Correlation of
high-resolution. B-mode and continuous-wave Doppler sonog-
raphy with arteriography in the diagnosis of carotid stenosis. Radi-
ology 149: 523-532. 1983
16. Zwiebel WJ: Doppler carotid imaging and evaluation of flow ab-
normalities. [In] Zwiebel WJ (ed). Introduction to vascular ultra-
sonography. New York, Grune & Stratton. 1982. pp 77-102
17. Turski PA, Zwiebel WJ. Strother CM, Crummy AB. Celesia GG.
Sackett JF: Limitations of intravenous digital subtraction angiog-
raphy. AJNR 4: 271-273. 1983
18. Phillips WC. Scott JA. Blasczcynski G: Statistics for diagnostic
procedures: How sensitive is "sensitivity"; how specific is "speci-
ficity"? AJR 140: 1265-1270. 1983
19. Comerota AJ. Cranley JJ. Cook SE: Real-time B-mode carotid
imaging in diagnosis of cerebrovascular disease. Surgery 89:
718-729. 1981

Enhanced In Vivo Platelet Activation In Subtypes Of

Ischemic Stroke
ARUN B. SHAH, M.D., NANCY BEAMER, M.S., AND BRUCE M. COULL, M.D.

SUMMARY It remains uncertain whether platelet activation in ischemic stroke is contributory or secon-
dary to brain ischemia. The efficacy of aspirin (ASA) in stroke prevention suggests that platelet activation
contributes to the occurrence of stroke. On the other hand, platelet activation may be simply a generalized
consequence of cerebral ischemic damage. To examine this issue, plasma levels of the platelet specific
proteins /3-thromboglobulin (/3-TG) and platelet factor 4 (PF4) were measured in fifty-eight patients with
various defined types of acute ischemic strokes.
/3-TG was a broader indicator of platelet activation than PF4. Compared with an age-matched control
group, thromboembolic and cardioembolic stroke patients had significantly elevated /3-TG levels (p <
0.001). Also, /3-TG levels in these stroke categories were significantly higher in samples drawn within the
first week after the event than in those drawn later (p < 0.001). In contrast, /3-TG levels in lacunar stroke
patients and in most TIA patients were normal.
/8-TG levels did not correlate with the volume of cerebral infarction as measured by planimetry from CT
scans. Moreover, /3-TG levels in patients on chronic ASA therapy at the time of stroke did not differ from
those in patients of the same diagnostic categories not taking aspirin. These data indicate that platelet
activation may be important in some, but not all, subtypes of ischemic stroke and that platelet activation can
occur in stroke even though the platelet cyclooxygenase pathway is suppressed.
Stroke Vol 16, No 4, 1985

CLINICAL AND EXPERIMENTAL STUDIES treatment and earlier identification of individuals at

which examine peripheral vascular, cardiovascular
and cerebrovascular disease indicate that tissue ische-
mia can induce platelet activation.1"3 This has raised
concern that the elevated parameters of platelet func-
tion seen in stroke patients may be a secondary phe-
nomenon unrelated to the primary events causing in-
farction.3 On the other hand, convincing clinical
evidence supports the view that platelet activation is a
primary event in the occurrence of stroke.4"6 Thus, the
benefit of aspirin in preventing stroke in persons with
transient ischemic attacks (TIA) of the brain is thought
to occur via suppression of platelet activation. The
resolution of the role of platelet activation in cerebral
ischemia is pivotal, since it may provide insight into
differing mechanisms of stroke. This may lead to new
From the Oregon Health Sciences University, Department of Neurol-
ogy, 3181 S.W. Sam Jackson Park Road, Portland, Oregon 97201.
This research was supported by NINCDS Grant #1 P50 17493-03
and was supported in part by USPHS Stroke Clinical Center Grant
#NS-17493-02.
Presented in preliminary form at the Ninth International Conference
on Stroke and Cerebral Circulation, Tampa, Florida, February 1984.
Address correspondence to: Bruce M. Coull, M.D., Oregon Health
Sciences University, Department of Neurology, 3181 S.W. Sam Jack-
son Park Road, Portland, Oregon 97201.
Received June 15, 1984; revision #1 accepted December 19, 1984.
high risk for stroke.
When platelets become activated the platelet specif-
ic proteins beta thromboglobulin (/3-TG) and platelet
factor 4 (PF4) are released into the circulation.7 Plasma
levels of these proteins in persons with normal renal
function serve as markers for enhanced in vivo platelet
activation.8 Since ischemic infarction may accompany
multiple disease processes, platelet activation could be
expected to distribute uniformly among subtypes of
stroke if the activation is simply a nonspecific conse-
quence of cerebral ischemic damage. Similarly, the
degree of non-specific platelet activation might be cor-
related with the volume of infarcted tissue. To test
these hypotheses, we measured circulating levels of/8-
TG and PF4 and estimated the volume of infarcted
brain tissue in patients with various types of ischemic
strokes.
Methods
Fifty-eight patients with either TIAs or acute cere-
bral infarction of all types and 35 persons comprising
two control groups were examined. The first control
group consisted of 20. young, healthy individuals, the
second consisted of 15 elderly persons who were age
matched with the stroke patients. None of the elderly

Downloaded from http://stroke.ahajournals.org/ by guest on November 17, 2013