2435
Equivalence of Measurements of Carotid Stenosis
A Comparison of Three Methods on 1001 Angiograms
Peter M. Rothwell, MB, ChB, MRCP; Rod J. Gibson, MRCP, FRCR; Jim Slattery, MSc;
Robin J. Sellar, FRCP, FRCR; Charles P. Warlow, MD, FRCP, for the European Carotid
Surgery Trialists' Collaborative Group
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Background and Purpose There is confusion about how
carotid stenosis should be measured on angiograms. If the
results of research based on different methods of measurement of stenosis are to be discussed and the results of clinical
trials properly applied to routine clinical practice, measurements made by the different methods must be formally
compared.
Methods The method of measurement of stenosis used in
the European Carotid Surgery Trial (ECST), that used in the
North American Symptomatic Carotid Endarterectomy Trial
(NASCET), and a method based on measurement of the
common carotid (CC) artery lumen diameter were compared.
Carotid stenosis was measured by two observers, working
independently and using the three different methods of measurement, on the angiographic view of the symptomatic carotid
stenosis that showed the most severe disease in 1001 patients
from the ECST.
Results The results of using the ECST and CC methods
differed from those of using the NASCET method in the
classification of stenoses as mild (0% to 29%), moderate (30%
to 69%), or severe (70% to 99%) in 51% of measurements.
The ECST and CC methods indicated that twice as many
stenoses were severe as did the NASCET method, and classified less than a third of the number of stenoses as mild. The
results of the ECST and CC methods differed from each other
in 15% of measurements. The relations between measurements made by each method to those made by the others were
approximately linear, so a simple equation could be derived to
convert measurements made by one method to measurements
made by the others.
Conclusions There were major and clinically important
disparities between measurements of stenosis made using
different methods of measurement on the same angiograms.
However, it is possible to convert measurements made by one
method to those of another using a simple arithmetic equation.
(Stroke. 1994;25:2435-2439.)
Key Words • angiography • carotid arteries • diagnosis
S
third method using the diameter of the visible diseasefree distal common carotid (CC) artery has also been
advocated.7
Twenty randomly selected studies involving measurement of stenosis on angiograms were reviewed; six used
the ECST method, six used the NASCET method, and in
eight no details of the method of measurement were
given.5 Use of more than one method has led to confusion in clinical practice and undermined the generalizability of the results of research. Because of the lack of
agreement about which method is the most appropriate,
the extent to which measurements made by each of the
methods actually differ must be defined. A preliminary
collaborative study between the ECST and NASCET
groups indicated that the ECST and NASCET methods
produced significantly different results,8 and the CC
method has not been systematically studied. If there are
major differences between the three methods, the relation between the measurements must be defined so
measurements made using one method can be compared
with those made using the other two. The aim of this
study is to compare the measurements made by the same
observers using the three methods on the same
angiograms.
ince the publication of the preliminary results of
the European Carotid Surgery Trial (ECST)1
and the North American Symptomatic Carotid
Endarterectomy Trial (NASCET),2 there has been increasing interest in how stenosis of the internal carotid
artery (ICA) should be measured. 35 The results of both
trials indicate that the degree of stenosis, expressed as a
percentage reduction in vessel diameter, is a major
factor in determining whether a patient is likely to
benefit from endarterectomy. Indeed, patient management frequently hinges on this single measurement.
However, stenosis was measured differently in the two
trials (Fig 1). In both the lumen diameter was measured
at the point of maximum stenosis (Dmin), but the
denominator in the equation used to calculate stenosis
was different. In the ECST the estimated normal lumen
diameter at the site of the lesion, based on a visual
impression of where the normal arterial wall was before
development of the stenosis, was used. In the NASCET
the diameter of a visible portion of disease-free ICA
distal to the stenosis was used, or the stenosis was
classified as 95% if the distal ICA had collapsed.6 A
Received June 3, 1994; final revision received July 8, 1994;
accepted July 29, 1994.
From the Department of Clinical Neurosciences, Western General Hospital, Edinburgh, Scotland.
Correspondence to Dr P.M. Rothwell, Department of Clinical
Neurosciences, Western General Hospital, Crewe Road, Edinburgh, Scotland EH4 2XU.
© 1994 American Heart Association, Inc.
Methods
One thousand and one consecutively selected carotid angiograms from patients randomly assigned to the "no surgery"
group in the ECST were studied. No angiograms were excluded. The sample included angiograms performed at approx-
2436
Stroke
Vol 25, No 12 December 1994
ECST method:
%*•
x 100% stenosis
NASCET method:
—
B
x 100% stenosis
c
TABLE 2. Comparisons of the Categorization
of Stenoses as Mild, Moderate, or Severe
by Each of Three Methods of Measurement
Stenosis by the ECST Method
CC method:
x 100% stenosis
70-99%
0-29%
30-69%
276
615
5
30-69%
2
385
389
70-90%
0
2
328
Stenosis by
0-29%
the NASCET
Method
Stenosis by the CC Method
FIG 1. Diagram of three methods of measuring carotid stenosis
on an angiogram when the stenosis is within the bulb. A,B, and
D are measurements made on a visible column of x-ray contrast;
C is a visual estimate of the likely normal lumen diameter before
development of the stenosis.
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
imately 100 centers in 14 European countries, and comprised
789 selective arterial angiograms, of which 307 were digitally
subtracted; 174 aortic arch angiograms, of which 92 were
digitally subtracted; 29 intravenous digital subtraction angiograms; and 9 angiograms for which the technique used was not
clear. The mean age of the patients studied was 62.1 years (SD,
7.8), and 71% were male. Two independent observers (P.M.R.
and R.J.G.) measured the degree of stenosis of the symptomatic carotid artery by using each of the methods detailed in Fig
1. In patients with bilateral symptoms, the most stenosed
artery was measured. Measurements were made using a
jeweler's eyepiece graduated in tenths of millimeters on the
single angiographic view that showed the greatest stenosis. The
same measurement of Dmin was used for each of the methods.
Each observer was blind to the measurements made by the
other and to the clinical details. No marks indicating the points
of measurement were placed on the angiogram films by either
observer.
Results
The two observers made a total of 6006 stenosis
measurements using each of the three methods. The
numbers of stenoses classified as mild (0% to 29%),
moderate (30% to 69%), or severe (70% to 99%) by
each of the three methods are shown in Table 1. The
numbers of stenoses within each group as measured by
the ECST and CC methods were similar, but the
NASCET method classified three times as many stenoses as mild than did the ECST method, and with the
ECST and CC methods more than twice as many
stenoses were classified as severe than with the
NASCET method. On 94 angiograms (9%) Dmin was
proximal or distal to the bulb. However, exclusion of
these angiograms altered the results very little, and they
have therefore been included in all analyses.
TABLE 1. Number of Stenoses Categorized as Mild,
Moderate, or Severe by Three Methods of Measurement
Degree of Stenosis
Method
0-29%
30-69%
70-99%
ECST (% of total)
278(14)
1002 (50)
722 (36)
NASCET (% of total)
896 (45)
776 (39)
330(16)
CC (% of total)
294 (15)
953 (48)
755 (37)
ECST indicates European Carotid Surgery Trial; NASCET,
North American Symptomatic Carotid Endarterectomy Trial; and
CC, common carotid.
Two observers made a total of 2002 measurements on 1001
angiograms.
Stenosis by
0-29%
294
598
4
the NASCET
30-69%
0
354
422
Method
70-90%
0
1
329
Stenosis by the CC Method
Stenosis by
0
0-29%
200
78
the ECST
30-69%
94
825
83
Method
70-90%
0
50
672
ECST indicates European Carotid Surgery Trial; NASCET,
North American Symptomatic Carotid Endarterectomy Trial; and
CC, common carotid.
Two observers made a total of 2002 measurements on 1001
angiograms.
Comparisons between each of the methods in the
categorization of stenoses as mild, moderate, or severe
are shown in Table 2. The ECST and NASCET methods
differed in the classification of stenoses in 51% of
measurements; in 1009 (99.6%) of the 1013 cases in
which the results of the two methods differed, the
NASCET method underestimated the severity of stenosis compared with the ECST method. The NASCET
and CC methods differed for 1025 (51%) measurements, with relative underestimation by the NASCET
method on 1024 (99.9%). The CC method and the
ECST method differed for 305 (15%) measurements,
with no significant bias in either direction.
The relations between measurements made by each
of the methods are shown in Fig 2. The means of the
stenosis measurements made by the two observers were
used in compiling the graphs to minimize errors that
may have been due to observer variability. In each case
there was some spread of measurements, particularly
with mild stenosis, but the relations were approximately
linear. The plot of ECST measurement versus CC
measurement approximately fit a line with a slope of
1.00 and an intercept on the CC axis of 0%. The lines
fitting the plots of ECST versus NASCET and CC
versus NASCET are identical, with a slope of 0.6 and an
intercept on the ECST and CC axes at -40%. The
following equation therefore describes the relation between measurements made by the NASCET method
and those made by either the ECST or CC methods:
ECST % stenosis or CC % stenosis=0.6(NASCET %
stenosis)+40%. By use of this equation, measurements
made by one method can be converted to those expected to be made by the other methods. The equation
can then be internally validated by assessing the agreement between the predicted measurements and the
actual measurements in categorization of stenosis as
mild, moderate, or severe (Table 3). For example,
Rothwell et al Three Methods of Measuring Carotid Stenosis
Fra 2. Scatterplots of the mean measurements of carotid stenosis made by two observers on 1001 angiograms. Top, results
of the European Carotid Surgery Trial (ECST) method versus
those of the North American Symptomatic Carotid Endarterectomy Trial (NASCET) method; middle, results of the common
carotid method versus those of the NASCET method; and
bottom, results of the ECST method versus those of the CC
method. Each point represents measurements from one or more
angiograms.
80 80
70(0
o
c
60 -fj
e>
SO -
ECST'
;*
40 30 2010-
-100
-80
-80
-40
-20
0
20
40
60
80
100
NASCET % stenosis
80 -
teno
90 SIS
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
100
70 -
• ••'-jtif
'JgF
• • ;^w
a
60
2
a
o
c
60-
•
• •
30-
o
20-
' •
••'.
"i ">Jlul
"- "• 1". 1
•" • T »ii
•
40-
o
E
E
o
2437
«
\
"Ul \ '
""T
'
•
•.••
1
-
10•
.
»—•
. -
"• F"
/ -.' 1
i
- .1
*
;
>
. . .
"T •'
.
•
0 -100
^. "-
'1
" •"
- S O - S O - 4 0 - 2 0
0
(M..llllljlllllllll|lltlllll(|IIIUIIII
20
40
SO
80 100
The differences between the results of the methods of
measurement were accounted for by the use of different
denominators. The 95% confidence intervals of the
ratio of each denominator to the others are given in
Table 4. The ECST denominator was, on average, the
same as the measurement of the CCA (common carotid
artery) lumen, although the 95% confidence interval of
the range of the proportion was 0.71 to 1.41. The
distribution of the ECST denominatonCCA ratio was
similar for both observers (Fig 3). The ICA:CCA ratio
varied with the severity of stenosis, remaining constant
up to 70% stenosis as measured by the ECST method
(50% by the NASCET method), then falling steadily as
stenosis increased (Fig 4).
Discussion
The first aim of this study was to determine the
frequency with which the three methods of measuring
stenosis placed the same patient into different categories of stenosis severity. The differences between the
results of the ECST and NASCET methods are considerable. Moreover, these differences are of major clinical
importance. Both the ECST and NASCET trials demonstrated that surgery is beneficial in patients with
severe stenosis (of 70% to 99%). However, classification
using the ECST method results in twice as many stenoses' being classified as severe (Table 1). Conversely, the
ECST trial showed that in patients with mild stenosis
(less than 30%) surgery is certainly not beneficial and
NASCET % stenosis
A E3 Ob—nwB
100
o
o
I
o
a
a
a
300
E
a
O
40
o
n
E
o
E «>
E
o 20
o
10
20
30
40
SO
SO
70
SO
90
100
ECST % stenosis
agreement between the ECST and NASCET methods
increased from 49% to 83% when NASCET measurements were transposed to an ECST scale.
<0.7
0.7-0.9
ii
0.H.1
1.1-1.3
>1.3
ECST denominator / CCA lumen
FIG 3. Graph of the number of bifurcations measured by two
observers (P.M.R. and R.J.G.) and classified according to the
ratio of the denominator of the European Carotid Surgery Trial
(ECST) equation to the measurement of the common carotid
artery (CCA) lumen.
2438
Stroke
Vol 25, No 12 December 1994
TABLE 4. Ratios and Confidence Intervals of Each of the
Denominators of Three Methods of Measuring Stenosis
Compared With the Others
Denominator
O
o
95% Cl
ECST denominator and CCA lumen
1.00
0.71-1.41
ECST denominator and ICA lumen
1.68
1.11-2.55
CCA lumen and ICA lumen
1.68
1.15-2.44
ECST indicates European Carotid Surgery Trial; CCA, common carotid artery; ICA, internal carotid artery.
See Fig 1 for description of ECST denominator.
.. 0.6
0)
<
O 0.5
<10 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-89 90-99
ECST % Stenosis
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
FIG 4. Plot of the mean ratio, by decile of stenosis calculated
according to the European Carotid Surgery Trial (ECST)
method, of the distal internal carotid artery (ICA) lumen to the
common carotid artery (CCA) lumen.
may be harmful. Using the limit of 30% stenosis with the
NASCET method of measurement, three times as many
stenoses are classified as mild as with the ECST method.
The second aim of the study was to define the
equivalence of measurements made by the three methods. The disparities between the results of the ECST
and NASCET methods can be illustrated using the
conversion equation given above. Stenoses of 30% and
70% as measured by the ECST method are approximately - 1 7 % and 50%, respectively, as measured by the
NACSET method. A stenosis of 70% as measured by
the NASCET method is equivalent to a measurement
with the ECST method of 82%. Indeed, it is likely that
this difference in the definition of severe stenosis accounts for the differences between the preliminary
ECST and NASCET results. Reanalysis of the ECST
results, confined to patients with 82% to 99% stenosis
(ie, severe stenosis by the NASCET method of measurement), produces results almost identical to those reported for the NASCET severe stenosis group.4
The approximately linear relations between measurements made by each of the three methods allows simple
conversion equations to be derived. These conversions
TABLE 3. Actual Agreement in the Categorization of
Stenosis as Mild, Moderate, or Severe Between Each of
Three Methods and Agreement After Conversion of
Measurements Made by the First Method to Those
Made by Second
Agreement (%)
Methods
Ratio
Actual
Measurements
Converted
Measurements
NASCET and ECST
49
CO
CC and ESCT
85
00
NASCET and CC
49
00
NASCET indicates North American Symptomatic Carotid Endarterectomy Trial; ECST, European Carotid Surgery Trial; and
CC, common carotid.
are not exact, because there is a spread of measurements around the line to which they are fitted, but the
internal validation suggests that they work reasonably
well. Moreover, the spread of measurements decreases
as the clinical importance of accurate measurement
increases: ie, with moderate and severe stenosis. The
disagreement in categorization of stenosis according to
the ECST and NASCET methods is decreased threefold after conversion of the NASCET measurements to
ECST measurements. Conversion from NASCET measurements to CC measurements works equally well.
That an error of categorization remains for about 15%
of stenoses must be interpreted in light of the 16% to
20% interobserver variability of each of the methods in
assigning stenoses to these categories. Even if the
conversion equations were perfect, observer variation in
the individual measurements would limit the agreement
after conversion to less than 90%. Ideally, if the results
of studies based on different measurements of stenosis
were compared, the angiograms would be remeasured
using the same method. However, this is often not
practical, and conversion equations will be required.
The close approximation of the results of the ECST
and CC methods indicates that the ECST denominator
is, on average, the same as the measurement of the CCA
lumen. Two previous studies using angiograms reported
a mean ratio of the bulb lumen to the CCA lumen of
1.19,7,10 Tm s has recently been misinterpreted as a
"fixed anatomical relationship."11 In fact, in both studies the ratio of 1.19 was far from fixed, having a standard
deviation of 0.09 in one7 and 0.19 in the other.10 In other
words, the ratio varied from less than 1.0 to more than
1.4. These studies used only normal angiograms and, of
particular importance, involved measuring the diameter
of the bulb at its widest point. The ECST method of
measurement of stenosis uses the lumen diameter at the
point of maximum stenosis, which is not infrequently
outside the bulb. Moreover, even when the maximum
stenosis is within the bulb it seldom coincides exactly
with the widest point. The ratio of the ECST denominator to the measurement of the CCA lumen diameter
will clearly vary with the site of the stenosis and will
usually be less than 1.2. In the present study the ratio of
the ECST denominator to the CCA lumen, in 1001
bifurcations, varied from 0.7 to 1.4, and the average
ratio was 1.00, both observers' results having similar
distributions.
Collapse of the ICA distal to tight stenoses has been
noted previously.12 With the NASCET method, any
stenosis distal to which the ICA is collapsed is classified
as a stenosis of 95%. 13 Despite attempts to define
criteria for determining exactly when the ICA should be
Rothwell et al
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
regarded as collapsed,6 the judgment must remain
somewhat subjective, and there is likely to be observer
variability. In the present study a steady reduction of
the ICA lumen, compared with the CCA lumen, begins
at about 70% stenosis as measured by the ECST method
(50% stenosis by the NASCET method). Moreover,
collapse of the ICA is marked in some patients but not
in others, so groups of patients with similar stenosis
measurements according to the ECST or CC methods
might have a different measurement according to the
NASCET method. This inconsistency of measurement
of severe stenoses is a problem of using the NASCET
method. Given the difficulty some observers have with
estimation of the normal bulb lumen in the ECST
method, the CC method would appear to have the
fewest drawbacks. The CCA is usually well visualized on
angiography, is without overlapping vessels, and is
rarely severely affected by atheroma. Most importantly,
the CC method is the most reproducible of the three.9
In the present study major differences between the
results of three methods of measurement of carotid
stenosis were demonstrated on the same angiograms.
These differences have important implications for clinical practice. However, the relations between measurements made by the three methods are approximately
linear, and conversions can be made using the equation
given above. These data provide the basis for informed
debate about which method of measurement of stenosis
should be adopted as the standard. The choice of a
standard method should depend on the ability to use it
to predict ipsilateral ischemic stroke and on its reproducibility. It is essential that a single method eventually
be adopted by all. An enormous amount of effort has
been put into determining which patients benefit from
carotid endarterectomy. It would be a great pity if,
research having come this far, the results of these trials
were obscured by argument and confusion about how
stenosis should be measured in routine clinical practice.
Three Methods of Measuring Carotid Stenosis
2439
Acknowledgments
Dr Rothwell's and Mr Slattery's research and the European
Carotid Surgery Trial are funded by the United Kingdom
Medical Research Council.
References
1. European Carotid Surgery Trialists' Collaborative Group. MRC
European Carotid Surgery Trial: interim results for symptomatic
patients with severe (70-99%) or with mild (0-29%) carotid
stenosis. Lancet. 1991;337:1235-1243.
2. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J Med.
1991;325:445-453.
3. Alexandrov AV, Bladin CF, Maggisano R, Norris JW. Measuring
carotid stenosis: time for a reappraisal. Stroke. 1993;24:1292-1296.
4. Barnett HJM, Warlow CP. Carotid endarterectomy and the measurement of stenosis. Stroke. 1993;24:1281-1284.
5. Rothwell PM, Warlow CP. Making sense of measurement of
carotid stenosis. Cerebrovasc Dis. In press.
6. Fox AJ. How to measure carotid stenosis. Radiology. 1993;186:
316-318.
7. Williams MA, Nicoliades AN. Predicting the normal dimensions of
the internal and external carotid arteries from the diameter of the
common carotid. EurJ Vase Surg. 1987;l:91-96.
8. Rothwell PM, Warlow CP. The European Carotid Surgery Trial
(ECST). In: Greenhalgh RM, Hollier LH, eds. Surgery for Stroke.
London, England: WB Saunders; 1993.
9. Rothwell PM, Gibson RJ, Slattery J, Warlow CP. Prognostic value
and reproducibility of measurements of carotid stenosis: a comparison of three methods on 1001 angiograms. Stroke. 1994;25:
2440-2444.
10. Schneidau A, Harrison MJG, Hurst C. Predicting the normal
dimensions of the internal carotid artery. EurJ Vase Surg. 1988;2:
273-275.
11. Bladin CF, Alexandrov A, Norris JW. A new method of measuring
carotid stenosis: the Toronto Stenosis Index. Stroke. 1994;25:250.
Abstract.
12. Brown PM, Johnston KW. The difficulty of quantifying the severity
of carotid stenosis. Surgery. 1982;92:468-473.
13. North American Symptomatic Carotid Endarterectomy Trial
(NASCET) Steering Committee. North American Symptomatic
Carotid Endarterectomy Trial: methods, patient characteristics,
and progress. Stroke. 1991;22:711-720.
Equivalence of measurements of carotid stenosis. A comparison of three methods on 1001
angiograms. European Carotid Surgery Trialists' Collaborative Group.
P M Rothwell, R J Gibson, J Slattery, R J Sellar and C P Warlow
Stroke. 1994;25:2435-2439
doi: 10.1161/01.STR.25.12.2435
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1994 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/25/12/2435
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/