Does color-flow imaging improve the accuracy of duplex carotid evaluation?

Transcription

1 Does color-flow imaging improve the accuracy of duplex carotid evaluation? Gregg L. Londrey, MD, Donald P. Spadone, MD, Kim J. Hodgson, MD, Don E. Ramsey, MD, Lynne D. Barkmeier, MD, and David S. Sumner, MD, Springfield, IlL To determine whether color-flow imaging enhances the accuracy of noninvasive carotid evaluation, the results of carotid duplex examinations t~om two laboratories, one with color-flow and the other with standard duplex imaging were compared. The techniques used by both laboratories were identical. All studies were interpreted by one of the authors, using the same criteria. From October 1988 through December 1989, 307 internal carotid arteries were evaluated with both color-flow imaging and standard angiography; and 206 underwent routine duplex scanning and angiography. Perfect agreement between test and angiographic results was significantly better with color-flow (86.6%) than with conventional duplex scanning (79.6%),p = (t test for independent samples). Significantly fewer vessels were over classified by one category with color-flow (8.5%) than with routine duplex scanning (16.5%), p = However, no difference was found in the number underclassified by one category (4.5% vs 3.4%),p = 0.5. Although these data support the accuracy of both modalities, there appears to be a trend toward improved results with the newer method. We attribute this to more precise placement of the pulsed Doppler sample volume afforded by the color-flow image. (J VAsc Suim 1991;13: ) The combination of B-mode ultrasonic imaging and pulsed Doppler, commonly referred to as duplex scanning, has become the diagnostic procedure of choice for noninvasive evaluation of the extracranial carotid system. Excellent results have been reported from several centers documenting its accuracy and confirming its widespread use. 1-4 The addition of color-flow technology to duplex scanning allows visualization of blood flow within the vessel. This modality offers some potential advantages over conventional duplex evaluation, mainly because of color-coding of flow velocity and direction. The increased velocity and turbulence associated with a stenotic area are readily appreciated, as are tortuous vessels, allowing precise placement of the pulsed Doppler sample volume-which is critical in the accurate determination of disease severity. We have performed more than 3600 color-flow carotid examinations since 1987 and have been impressed by its advantages, especially for evaluating From the Department of Surgery, Section of Peripheral Vascular Surgery, Southern Illinois University School of Medicine Springfield. Presented at the Fourteenth Annual Meeting of the Midwestern Vascular Surgical Society Toledo, Ohio, Sept , Reprint requests: David S. Sumner, MD, Southern Illinois University School of Medicine, Department of Surgery, P.O. Box 19230, Springfield, IL /6/27421 Table I. Criteria for grading internal carotid artery stenosis Diameter stenosis 0%-15% 16%-49% 50%-79% 80%-99% 100% Criteria PSV < 120 cm/sec minimal spectral broadening PSV < 120 cm/sec spectral broadening and/or PSV approaching 120 cm/sec PSV > 120 cm/sec EDV < 120 cm/sec PSV > 120 cm/sec EDV > 120 cm/sec Absence of detectable flow PSV, Peak systolic velocity; EDV, end diastolic velocity. difficult vessels with either high-grade lesions, significant tormosity, or possible occlusion. To determine whether this new technology results in any real improvement in accuracy, the results of color and standard duplex examinations from our institution were compared with arteriography as the gold standard. MATERIAL AND METHODS The accuracy of carotid duplex examinations from two laboratories over a recent 15-month period (October 1988 through December 1989) were compared. One laboratory used a standard duplex instrument (Diasonics 400; Diasonics, Inc., Milpitas, Calif.) and the other a color-flow scanner (Quad 1 Quantum Angiodynograph; Quantum Medical Sys- 659

2 Journal of 660 Londrey et al. VASCULAR SURGERY Table II. Comparison of standard duplex scanning and angiography in 206 vessels Duplex Arteriography (% stenosis) (% stenosi~) 0%-15% 16%-49% 50%-79% 80%-99% 100% 0%-15% %-49% %-79% %-99% % Kappa = , Perfect agreement = 79.6%. Over-classified by one category = 16.5%. Under-classified by one category = 3.4%. Table III. Accuracy of standard duplex examination Predictive value % Stenosis Sensitivity Specificity Positive Negative Accuracy -< 14 vs -> 15 98% 81% 89% 97% 92% -<49 vs > % 90% 87% 100% 94% -< 79 vs > 80 87% 96% 82% 97% 94% -< 99 vs % 99% 82% 100% 99% tems, Issaquah, Wash.). Both laboratories are supervised by the Section of Peripheral Vascular Surgery and are located in the two hospitals affiliated with our institution. The examination techniques used were identical, and all studies were interpreted by one of the authors using the same criteria (Table I). Doppler velocity spectra were recorded with the angle between the sound beam and the long axis of the vessel adjusted as close to 60 degrees as possible. During the study period 1294 carotid studies were performed with the standard duplex scanner. One hundred eight patients underwent carotid arteriography within 1 month of the noninvasive study. After inadequate and unilateral arteriograms had been excluded, 206 carotid arteries were available for comparison. Over the same period color-flow imaging was used in 1502 studies. Arteriograms were obtained in 168 patients, and 307 vessels were left for comparison after exclusions. Carotid arteriograms were either cut-film or digital arterial studies obtained by use of selective catheterization techniques in two planes. Occasionally venous digital studies ( < 5%) were accepted for comparison if the degree of internal carotid artery (ICA) stenosis could be clearly determined. All arteriograms were reviewed by one of the authors without knowledge of the results of the noninvasive examination. The degree of stenosis was determined by measuring the diameter of the vessel lumen at the narrowest area of the diseased internal carotid artery (x) and the diameter of the normal vessel distal to the area of disease (y). The percent stenosis was calculated by the following formula: (1 - x/y)100. Sensitivity, specificity, positive and negative predictive values, and accuracy were calculated for each category of disease severity, s The kappa statistic was determined as a measure of correlation between predicted ICA stenosis and angiographic measurement. A kappa value of 1 indicates perfect agreement? Because the patient populations were not identical, neither the accuracy statistics nor the kappa statistic for the 2 methods could be compared with standard tests of significance. Thus to more confidently determine whether any real difference in accuracy existed between color-flow and duplex imaging, the munber of vessels with perfect correlation, one category overclassified, and one category underclassified were compared by use of the t test for independent samples. It was thought that if significant differences occurred in at least two of these areas, it could be postulated that a true difference in accuracy existed between the two tests. RESULTS Standard duplex test results are compared to arteriographic findings in Table II, and accuracy statistics for each disease category are listed in Table III. Tables IV and V provide the same data for color-flow imaging. Perfect agreement between test

3 ",7o!vme 73 Numbcr 5, Ma; 199t Color-flow imaging of carotid arteries 661 Table IV. Comparison of color-flow scanning and angiography in 307 vessels Color-flow Arteeiography (% Stenosis) (% Stenosis) 0%-15% 16%-49% 50%-79% 80%-99% 100% 0%-15% %-49% %- 79% %-99% % Kappa = ; Perfect agreement = 86.6%. Over-classified by one category = 8.5%. Under-classified by one category = 4.5%. Table V. Accuracy of color-flow examination Predictive value % Stenosis Semitivity Specificity Positive Negative Accuracy _< 14 vs -> 15 96% 93% 96% 93% 95% -<49 vs _> 50 99% 94% 91% 99% 96% -< 79 vs -> 80 92% 97% 89% 98% 96% -< 99 vs % 99% 92% 100% 99% and arteriographic results was significantly better with color-flow (86.6%) than standard duplex scanning (79.6%), p = There were also significantly fewer vessels over-classified by one category with color-flow (8.5%) than with conventional duplex scanning (16.5%), p = However, no difference was found in the number underclassified by one category (4.5% vs 3.4%),p = 0.5. The kappa statistic was also higher with colorflow than with standard duplex imaging (0.820 vs 0.727). The tendency to over-classify stenoses with the standard duplex examination is reflected by lower specificity and positive predictive values for all disease categories in comparison to color-flow imaging. Sensitivity and negative predictive values were generally comparable for the two tests (Tables III and V). DISCUSSION Over the last decade, duplex scanning has become the most commonly used method of noninvasive carotid evaluation. Confidence in its accuracy has improved to the point that some groups now recommend the performance of carotid endarterectomy based on duplex results without preoperative arteriography. 7,s Perfect agreement between test and arteriographic results of 85% and 87% have been reported, 1'2 as have sensitivities and specificities exceeding 90% for differentiating stenoses of greater than 50% from those of less severity. 3'4'8 Less data exist documenting the accuracy of color-flow imaging in comparison to a gold standard. Polack et al.9 reported a sensitivity of 90%, specificity of 79%, and accuracy of 87% in detecting stenoses of greater than 50% in 60 ICAs. Steinke et al.10 found an accuracy of 80% in 60 carotid vessels compared with angiography. Although these papers documented good results with color-flow imaging, the number of studies were limited, and they included no data to recommend its use over standard duplex imaging. Some groups have focused on the color-flow image itself as a means of determining the degree of diameter reduction-just as one would measure the residual lumen on an arteriogram. 11"12 Erickson et al.12 found this method particularly useful in differentiating degrees of stenosis of less than 50%. We have not attempted to assign a stenosis category based on the image and continue to rely on the Doppler spectrum as the primary method of disease classification. In our experience the color-flow image is helpful in four ways: (1) it clearly shows the course of the vessel to be sampled, which is important for the proper evaluation of tortuous and kinked carotid arteries (Fig. 1); (2) external and internal carotid vessels are more easily differentiated based on visualization of external carotid branches (Fig. 2); (3) the absence of flow in occluded vessels is much more confidently determined; and (4) in vessels with significant stenoses the precise location and direction of flow in the area of highest velocity is readily

4 Journal of VASCULAR SURGERY 662 Londrey et al Fig. 1. The course of a tortuous internal carotid artery is clearly demonstrated, allowing the Doppler sample volume to be aligned accurately. Interrogation by pulsed Doppler of the portion of the vessel coursing directly away from the probe would result in a falsely elevated velocity measurement.,~ Fig. 2. Normal carotid bifurcation with the superior thyroid artery (white arrow) arising from the proximal external carotid artery, differentiating it from the internal carotid artery. visualized, permitting precise Doppler sampling (Fig. 3). Although disease severity is graded primarily by use of velocity criteria, the degree of stenosis suggested by the color-flow image can be quite helpful in Fig. 3. Color desaturation to white within the area of greatest stenosis targets the highest flow velocities fbr Doppler spectrum analysis. confirming the category of disease suggested by the Doppler spectrum. The image is also useful in grading lesions that fall on the margin of two categories by velocity criteria and in cases where a clean, measurable spectrum cannot be obtained. We believe these factors allow greater confidence in interpreting the examination, thus decreasing the tendency to over-classify lesions. Our data must be viewed with the realization that several variables existed in addition to the type of duplex scanner, the most significant of which was the different vascular technologists in the two laboratories. There is no question that the level of skill and experience of the technologist can significantly affect the reliability of the examination. Both laboratories are staffed by experienced personnel, and all examinations are reviewed on videotape and interpreted by one of the authors. In spite of this, the technologist factor remains an uncontrolled variable. That the patient group studied with conventional duplex scanning was different from that studied with color-flow imaging constitutes a second important variable. However, the two vascular laboratories drew from the same community, and the patient populations have consistently had similar disease patterns. Our experience with both conventional duplex and color-flow imaging supports the accuracy of both modalities with a trend toward improved results with the newer method. We believe that the additional

5 Volume 13 Number 5 May 1991 Color-flow imaging of carotid arteries 663 information provided by the color-flow image contributes significantly to the accuracy and reliability of carotid duplex evaluation. The authors thank Paul Kolm, PhD, Southern Illinois University, Division of Statistics and Research Consulting, for evaluating the data and performing the statistical analyses and to Mary Garfield, BS, MT(ASCP), for data storage and preparation. REFERENCES 1. Roederer GO, Langlois YE, Chan AW, et al. Ultrasonic duplex scanning of extracranial carotid arteries: improved accuracy using new features from the common carotid artery. J Cardiovasc Ultrasonography 1982;1: Jacobs NM, Grant EG, Schellinger D, Byrd MC, Richardson JD, Cohan SL. Duplex carotid sonography: Criteria for stenosis, accuracy and pitfalls. Radiology 1985;154: Kenagy JW. Comparison of duplex scanning and contrast arteriography: A community hospital experience. J VASC SURG 1985;2: Martin KD, Patterson RB, Fowl RJ, Kempczinski RF. Is the continued use of ocular pneumoplethysmography necessary for the diagnosis of cerebrovascular disease? J VASC SUWG 1990; 11: Sumner DS. Evaluation of noninvasive testing procedures: data analysis and interpretation. In: Bemstein EF, ed. Noninvasive diagnostic techniques in vascular disease. St. Louis: CV Mosby, 1985: Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960;20:37-46, 7. Goodson SF, Flanigan DP, Bishara RA, Schuler JJ, Kikta MJ, Meyer JP. Can carotid duplex scanning supplant arteriography in patients with focal carotid territory symptoms? J VASe Sung 1987;5: Moore WS, Ziomek S, Quifiones-Baldrich WJ, Machleder HI, Busuttil RW, Baker JD. Can clinical evaluation and noninvasive testing substitute for arteriography in the evaluation of carotid artery disease? Ann Surg 1988;208: Polak JF, Dobkin GR, O'Leary DH, Wang A, Cutler SS. Internal carotid artery stenosis: accuracy and reproducibility of color-doppler assisted duplex imaging. Radiology 1989; 173: Steinke W, Kloetzsch C, Hennerici M. Carotid artery disease assessed by color Doppler flow imaging: correlation with standard Doppler sonography and angiography. AJNR 1990; 11: Middleton WD, Foley WD, Lawson TL. Color-flow Doppler imaging of carotid artery abnormalities. AJR 1988;150: Erickson SJ, Mewissen MW, Foley WD, et al. Stenosis of the internal carotid artery: assessment using color Doppler imaging compared with angiography. AJR 1989; 152: Submitted Oct. 9, 1990; accepted Dec. 13, 1990.