| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ultrasonography |
1 From the Departments of Radiology (E.G.G., A.J.D., S.M.E.S., M.L.M., G.M.H., P.T.Z., A.K.M.), Neurology (S.N.C.), and Surgery (J.D.B.), West Los Angeles Veterans Affairs Medical Center, 11301 Wilshire Blvd, Los Angeles, CA 90073. Received November 25, 1998; revision requested January 18, 1999; revision received April 8; accepted June 28. Address reprint requests to E.G.G. (e-mail: egrant@ucla.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: ICA peak systolic velocity (PSV) and the ratio of the PSV in the ICA to that in the ipsilateral common carotid artery (VICA/VCCA) were compared with the degree of arteriographically measured stenosis. ICAs were arteriographically subgrouped at 10% incremental levels of stenosis and broader ranges. Mean PSV, VICA/VCCA, and SDs were calculated for each category. Histograms showing the numbers of stenotic ICAs in subgroups and for vessels with stenoses of greater than or equal to or less than 70% narrowing were constructed. The number of vessels correctly subgrouped with typical Doppler US thresholds was calculated.
RESULTS: Mean PSV and VICA/VCCA increased with stenosis level (P < .01); SDs were wide. Histograms showed Doppler US values in the central groups across all disease levels. Histograms differentiating at least or less than 70% stenosis showed minimal overlap. PSV and VICA/VCCA helped classify, respectively, 185 and 181 of 204 vessels with stenoses of less than 50%, 15 and 21 of 46 vessels with stenoses of 50%–69%, and 73 and 67 of 84 vessels with stenoses of 70% or greater. When classifying stenoses as 69% or less or 70% or more, PSV and VICA/VCCA were correct in 90.6% and 90.3% of vessels.
CONCLUSION: Doppler US is excellent for classifying stenoses as above or below a single degree of severity but does not function well in stenosis subclassification.
Index terms: Angiography, comparative studies, 172.1245, 172.1247, 904.1223 • Carotid arteries, angiography, 172.1245, 172.1247, 904.1223 • Carotid arteries, stenosis or obstruction, 172.721 • Carotid arteries, US, 172.12983, 904.12983
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Numerous articles have attested to the accuracy of color duplex US in the diagnosis of carotid arterial disease (1–16). Most report a peak accuracy of 90% or greater, despite the fact that a wide range of Doppler US thresholds have been published in these articles. While modern US combines gray-scale technology with the visual display of color and the velocity information of spectral Doppler US, the spectral examination remains the primary method through which stenoses in the internal carotid artery (ICA) are quantified.
Staff at vascular laboratories usually examine a mixed population of patients, who are clinically grouped into symptomatic or asymptomatic cohorts in accordance with two large studies performed in North America—the North American Symptomatic Carotid Endarterectomy Trial (NASCET) (17) and the Asymptomatic Carotid Atherosclerosis Study (ACAS) (18). The researchers in these studies chose different surgical thresholds (stenoses of >70%- and >60%-diameter narrowing, respectively). Recently, the results of carotid endarterectomy in symptomatic patients with stenoses of 50%–69%-diameter narrowing have become available from the NASCET and show a statistically significant improvement in outcome in this population, as well (19).
Practitioners in most noninvasive testing facilities approach the US diagnosis of ICA stenosis by stratifying the degree of vessel narrowing into several categories and by predicting the severity of stenosis in accordance with the amount of elevation of a given Doppler US parameter or parameters. These divisions vary between laboratories. In our own institution, we stratify stenoses into ranges of 0%–49%, 50%–69%, and 70% or greater narrowing, but this is by no means the only scheme in use today.
The literature and the practice patterns of most vascular laboratories support the notion that there is a direct correlation between the degree of elevation of a given Doppler US parameter and the degree of stenosis in the ICA. However, we believe that the ability to use duplex US to determine the degree of stenosis present has not been adequately investigated and that it is taken largely for granted. We undertook this study to determine the limits within which color duplex US can be used to predict the degree of stenosis in the ICA.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patients were referred for carotid arterial US on the basis of having clinically accepted indications such as stroke, transient ischemic attack, or amaurosis fugax in symptomatic patients or carotid bruit, plans to undergo coronary arterial bypass, or severe peripheral vascular disease in asymptomatic patients. Of patients who underwent arteriography, 171 of 202 (84.7%) were referred after US showed a hemodynamically significant stenosis. Among the remaining patients, 14 of 202 (6.9%) had stroke with no apparent cause, eight of 202 (4.0%) were suspected of having vasculitis, seven of 202 (3.5%) had severe vertebrobasilar symptoms, and two of 202 (less than 1.0%) had an arteriovenous malformation.
Imaging
All carotid arterial US was performed by experienced technologists in a single accredited laboratory in a Veterans Hospital. Equipment included commercially available, state-of-the-art scanners (HDI, Mark 9, Mark 3000, Mark 5000, Advanced Technology Laboratories, Bothell, Wash; XP-128, Acuson, Mountain View, Calif). Five-megahertz or 7.5-MHz linear-array transducers were used as dictated by patient body habitus.
All scans were obtained according to a set laboratory protocol. Angle adjustment was based on flow direction depicted by color Doppler US; the Doppler US angle was maintained below 60°. US images were reviewed by a single radiologist (E.G.G.); the highest angle-adjusted peak systolic velocity (PSV) was recorded from within each ICA. The ratio of the PSV in the ICA to that in the ipsilateral common carotid artery (VICA/VCCA) was also noted (2).
Carotid arteriography was performed from a femoral arterial approach, and two or more orthogonal views of each bifurcation were obtained. The arteriographic measurement of the degree of stenosis was determined in accordance with NASCET methods (17). The diameter of the minimal residual lumen identified on a single view was measured on the image by using a jeweler's magnifying glass with an embedded measuring scale marked in 0.1-mm increments. This was compared with the diameter of the "normal" part of the ICA distal to the stenosis. The formula for computation of the degree of stenosis was [1 - (DMRL/DNL)] x 100, where DMRL is the diameter of the minimal residual lumen and DNL is the diameter of the normal lumen.
Two experienced neuroradiologists (S.M.E.S., G.M.H.) reviewed all arteriograms independently, without knowledge of the US results. Readings from the two neuroradiologists were averaged. This value was used as the standard against which Doppler US velocity readings were compared.
The identification with arteriography of a common carotid artery with a narrowing of 50% or greater anywhere along its length resulted in the exclusion of the ipsilateral ICA from our study because of the potential to alter distal flow dynamics. Arteries with occlusion also were excluded from statistical data, as they were diagnosed on the basis of color power Doppler US appearance and not Doppler US velocity.
ICAs having an atrophied or "string-sign" residual lumen over much or all of their cervical segment were classified in two ways, depending on the US findings. Such lesions were said to be present on arteriograms when the diameter of the distal part of the ICA was less than the diameter of the ipsilateral internal maxillary artery (20).
Vessels without high-velocity flow, which were diagnosed as subtotal occlusions on the basis of color power Doppler US appearance, were excluded, as their velocity values at spectral analysis were either unmeasurable or well within the range of normal for any Doppler US threshold. Arteries with high-velocity flow were included in our calculations and were considered to have 99% stenosis.
Statistical Analysis
Maximum PSV and VICA/VCCA within a given vessel were compared with the calculated percentage of arteriographically determined stenosis. We then divided ICAs into subgroups with 10% incremental degrees of stenosis (0%–9%, 10%–19%, 20%–29%, etc) on the basis of arteriographic results. We calculated the mean PSV and VICA/VCCA and their SDs for each subgroup. The Student t test was used for individual comparisons of the subgroups. Similar calculations were performed by using broader ranges of stenoses, as well (ie, 0%–49%, 50%–69%, and 
70%). An analysis of variance was performed to compare the mean PSV and VICA/VCCA for each of the groups at the .01 significance level.
We then plotted the distribution of PSV and VICA/VCCA in terms of the numbers of carotid arteries with stenoses in the aforementioned categories (10% increments and broad ranges). Histograms of PSV and VICA/VCCA for vessels with 69% stenosis or less or with greater than or equal to 70% stenosis also were constructed.
Finally, we calculated the number of vessels correctly classified into the broader subgroups by using a spectral Doppler US scheme employed in our laboratory (Table 1). Similar calculations were performed to assess the ability to use spectral Doppler US to classify lesions as being less than or greater than or equal to a single threshold. Threshold values for PSV and VICA/VCCA were derived by using receiver operating characteristic curve data from our own laboratory but were not unlike those used in other institutions or in the literature.
|
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
PSV and VICA/VCCA as a Function of Lesion Severity
Mean PSV and VICA/VCCAs with SDs for the 10% incremental categories of stenosis are plotted in Figure 1a and 1b. These graphs in Figure 1 clearly demonstrate that the mean PSV and VICA/VCCA increased in proportion to the degree of stenosis being considered. Analysis of variance showed statistically significant differences in the mean PSV and VICA/VCCA for a comparison of all 10% incremental subgroups (P < .01).
|
|
SDs, however, were extremely wide about all means, and considerable overlap was present at most adjacent 10% incremental categories for both PSV and VICA/VCCA. Similar findings were present when broad categories of stenosis were considered, as well (graphs not shown).
Distribution of PSV and VICA/VCCA within Lesion Subgroups
Figure 2a shows the distribution of PSV for 10% increments of stenosis; Figure 2b shows the distribution of PSV for lesions grouped in broader categories: 0%–49%, 50%–69%, and 70% or greater stenosis. Reviewing Figure 2a, we find that reasonable differentiation at the 10% degree of stenosis was possible between the extremes of the various categories, for example, between lesions with 50%–59% stenosis and those with greater than 80% stenosis. However, PSVs in the pivotal 60%–69% and 70%–79% ranges were spread across all degrees of stenosis. When broader categories of stenosis were evaluated (Fig 2b), again, considerable overlap in velocities remained between the central group (50%–69% stenosis) and the categories on either side (0%–49% and 70% stenosis).
|
|
|
By using the PSV threshold shown in Table 1 to classify lesions with less than 50% narrowing, 185 of 204 carotid arteries (90.7%) were classified correctly. For VICA/VCCA, 181 of 204 carotid arteries (88.7%) were classified correctly. For lesions at the 50%–69% degree, with PSV, only 15 of 46 carotid arteries (33%) were classified correctly. If VICA/VCCA was used, then 21 of 46 carotid arteries (46%) were classified correctly. For lesions with narrowing of at least 70%, with PSV, only 73 of 84 (87%) carotid arteries were classified correctly. If VICA/VCCA was used, then 67 of 84 carotid arteries (80%) were classified correctly (Table 2).
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Different surgical thresholds apply for symptomatic and asymptomatic patients in accordance with the ACAS and NASCET studies. The need to differentiate lesions with 50%–69% narrowing from those with 70% or greater narrowing in symptomatic patients may be necessary, per the recent results of the NASCET that show improved outcome (although far less positive than the outcome in patients with 70% stenoses) after carotid endarterectomy in the former group (19).
Other surgical thresholds may be used as well; in our own institution, vascular surgeons tend to reserve carotid endarterectomy in asymptomatic patients for those with greater than 80% stenoses. While many laboratories claim to accurately diagnose increasingly focused categories of stenosis with duplex US, this ability has not been systematically evaluated.
A review of the older literature (4–8), most current textbooks (21–24), didactic lectures, and general practice patterns lends support to the notion that Doppler US is capable of being used to differentiate various degrees of stenosis successfully. Findings of a recent survey of noninvasive vascular testing facilities accredited by the Intersocietal Commission for Accreditation of Vascular Laboratories showed that, although there was considerable variation in exactly how diagnostic categories were broken down, more than 90% of responding institutions stratified their ICA diagnoses into three or more degrees (25).
The results of our study show that the mean PSV and VICA/VCCA increase in proportion to the degree of arteriographically determined stenosis. Our results are almost identical to those published in 1988 after a large multicenter study (5), and differences in mean Doppler US values at the 10% degree of stratification are statistically significant. However, the application of confidence levels about the mean PSVs or VICA/VCCAs in our plots reveals that, in the diagnosis of categories of greater or less than a certain degree of stenosis, there is almost complete overlap due to the wide range of Doppler US values for a given degree of arteriographically determined stenosis.
Widening the categories into broader ranges of stenoses (
49%, 50%–69%, and 70% narrowing) still leaves a great deal of overlap in the moderate range (50%–69% narrowing) from both categories above (70% narrowing) and those below (49% narrowing). This problem is further substantiated by our histograms, which graphically illustrate the degree of variation of spectral values that are found in association with a specific degree or range of stenosis.
Assessment of the actual numbers of stenoses able to be classified into the three broad categories produces quantitative evidence that Doppler US does not function well when compared with arteriography in helping to subclassify stenoses.
Of 46 patients having arteriographically proved lesions of between 50% and 69% narrowing, with PSV, only 15 (33%) had stenoses that were correctly classified by using thresholds of 150 cm/sec and 225 cm/sec for the lower and upper limits of the range, respectively. Widening the threshold ranges more would improve results in this category, but at the expense of accuracy in the adjacent groups. While results obtained by using VICA/VCCA were better (46%), more than half the patients had stenoses that were still misclassified by using this diagnostic parameter.
That our results and those published in the literature show Doppler US to be an excellent examination for the differentiation of lesions with narrowing of greater than or equal to or of less than a single degree is not contradictory to the poor results we found in differentiating multiple degrees of stenosis in the same population. The reason for this apparent discrepancy is well illustrated at comparison of Figure 2b and 2c. In Figure 2b, again, there is a great deal of overlap of velocities among arteries with 50%–69% stenoses and among those in the adjacent categories. Eliminating the central or moderate stenosis category, however, produces far better results, as the majority of normal arteries (in this example, those with <70% stenoses) have stenoses below surgical thresholds. In fact, specificity remains at 100% until the PSV exceeds 125 cm/sec. Conversely, arteries with 70% or greater stenosis are scattered throughout the remainder of the Doppler US velocity range.
When identifying patients with vessels with stenosis greater or less than a single degree, significantly little overlap exists between velocities in normal and those in abnormal vessels when all the vessels in a population are considered. It should be pointed out that, statistically, use of broader categories tends to result in improved classification of stenoses because of the larger absolute numbers relative to the numbers in narrower categories.
The fact that most recent studies have dealt with the evaluation of a single degree of stenosis has allowed investigators to avoid discussing the potential difficulties of diagnosing different degrees of stenosis in a single population. A close inspection of the older literature, however, demonstrates that authors have described difficulties with spectral Doppler US in the quantification of carotid arterial disease. In Zwiebel and colleagues' series (3), Doppler US was used to correctly identify vessels with lesions between 50% and 69% narrowing in only 43% of cases, a result remarkably similar to our own.
Arteriography, the standard with which US is compared, has been noted to have difficulty in helping to classify lesions into categories as tight as 10% (26), even with an allowance for the minimal interobserver error usually ascribed to the NASCET method of measurement (27). It would, therefore, be unreasonable to expect Doppler US to perform better.
Furthermore, exact correlation between arteriography and US would be unlikely, as the former quantifies stenoses on the basis of anatomic data and the latter quantifies stenoses on the basis of physiologic information. The NASCET method of measurement itself introduces a further chance for mismatch between arteriography and US, as the degree of stenosis is not determined by measuring the actual degree of vessel narrowing but by comparing the residual lumen with the distal lumen.
How then does one laboratory handle the need to estimate various degrees of stenosis? One solution might be to segregate the population into symptomatic and asymptomatic cohorts and to assign patients to surgical or nonsurgical categories in each on the basis of appropriate thresholds. This method takes advantage of the fact that Doppler US is excellent at helping to distinguish between patients with arteriographically determined stenoses above or below a single given degree yet avoids the problem of estimating specific degrees of vessel narrowing. In a recent investigation (28), we demonstrated that higher levels of accuracy, specificity, and sensitivity can be achieved by segregating the population into symptomatic and asymptomatic cohorts and by applying appropriate thresholds that can be achieved by evaluating a mixed population.
One could also continue to provide diagnoses that are stratified into several categories but be aware that there are inaccuracies inherent in this form of analysis.
Finally, if knowing the exact degree of stenosis in a given patient is of great clinical import, arteriography may be necessary. While magnetic resonance (MR) arteriography is being used in carotid arterial imaging with increasing frequency and a combination of US and MR arteriography may provide sufficient information to determine if a lesion is operable, MR arteriography does not currently provide images of sufficient resolution to measure the exact degree of stenosis (29). Therefore, arteriography remains the only alternative, and cost plus potential complications of this invasive procedure must be factored into the decision.
In summary, our study shows that duplex US is an excellent examination for classifying vessels as having stenosis above or below a single degree of severity in a given population. Conversely, duplex US is limited in its ability to help quantify the degree of stenosis in any given ICA. While results are better when broader categories are considered, the correct identification of vessels in the moderate range of stenosis (50%–69% narrowing) was less than 50% in this study. When characterization of specific degrees of stenosis in individual patients is needed, arteriography remains the only alternative.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, E.G.G.; study concepts and design, E.G.G., A.J.D.; definition of intellectual content, E.G.G., A.J.D.; literature research, E.G.G., A.K.M.; clinical studies, E.G.G., A.J.D., S.M.E.S., M.L.M., G.M.H., P.T.Z.; data acquisition, E.G.G.; data analysis, E.G.G., A.J.D., S.M.E.S., G.M.H., A.K.M.; statistical analysis, E.G.G., A.J.D.; manuscript preparation, E.G.G.; manuscript editing, E.G.G., A.J.D., M.L.M., G.M.H., S.N.C., J.D.B.; manuscript review, G.M.H., S.N.C., J.D.B.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. R. Sprouse II, G. H. Meier, F. N. Parent, R. J. DeMasi, C. J. LeSar, C. Nelms, K. Carter, M. J. Marcinczyk, R. G. Gayle, and B. Mendoza Are We Undertreating Carotid Stenoses Diagnosed by Ultrasound Alone? Vascular and Endovascular Surgery, March 1, 2005; 39(2): 143 - 151. [Abstract] [PDF]
|
|
|
|
 |
|
 G. M. Hathout, J. R. Fink, S. M. El-Saden, and E. G. Grant Sonographic NASCET Index: A New Doppler Parameter for Assessment of Internal Carotid Artery Stenosis AJNR Am. J. Neuroradiol., January 1, 2005; 26(1): 68 - 75. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Kono, S. P. Pinnell, C. B. Sirlin, S. R. Sparks, B. Georgy, W. Wong, and R. F. Mattrey Carotid Arteries: Contrast-enhanced US Angiography--Preliminary Clinical Experience Radiology, February 1, 2004; 230(2): 561 - 568. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. G. Grant, C. B. Benson, G. L. Moneta, A. V. Alexandrov, J. D. Baker, E. I. Bluth, B. A. Carroll, M. Eliasziw, J. Gocke, B. S. Hertzberg, et al. Carotid Artery Stenosis: Gray-Scale and Doppler US Diagnosis--Society of Radiologists in Ultrasound Consensus Conference Radiology, November 1, 2003; 229(2): 340 - 346. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. M Hathout, M. J. Duh, and S. M. El-Saden Accuracy of Contrast-Enhanced MR Angiography in Predicting Angiographic Stenosis of the Internal Carotid Artery: Linear Regression Analysis AJNR Am. J. Neuroradiol., October 1, 2003; 24(9): 1747 - 1756. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Schulte-Altedorneburg, D. W. Droste, S. Felszeghy, L. Csiba, V. Popa, K. Hegedus, J. Kollar, L. Modis, and E. B. Ringelstein Detection of Carotid Artery Stenosis by In Vivo Duplex Ultrasound: Correlation With Planimetric Measurements of the Corresponding Postmortem Specimens Stroke, October 1, 2002; 33(10): 2402 - 2407. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Long, J. A. Kern, S. M. Fiser, A. K. Kaza, D. C. Cassada, B. T. Miller, J. A. Claridge, I. L. Kron, and C. G. Tribble Carotid Arteriography Impacts Carotid Stenosis Management Vascular and Endovascular Surgery, July 1, 2001; 35(4): 251 - 256. [Abstract] [PDF]
|
|
|
|
|
|
M. Vergara, E. G. Grant, S. M. El Saden, and G. M. Hathout Duplex US for the Estimation of Internal Carotid Stenosis Drs Grant and colleagues respond: Radiology, May 1, 2001; 219(2): 575 - 577. [Full Text]
|
|
|
|
 |
|
 How Accurate Is Carotid Ultrasound? Journal Watch (General), January 18, 2000; 2000(118): 3 - 3. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) and the percentage of stenosis and (b) the mean VICA/VCCA (
) and the percentage of stenosis measured arteriographically. The PSV and VICA/VCCA increase with the increasing severity of stenosis. Error bars show 1 SD about the mean. Note the overlap in adjacent categories of stenosis.
