Duplex Criteria for Determination of 50% or Greater Carotid Stenosis

Transcription

1 Article Duplex Criteria for Determination of 50% or Greater Carotid Stenosis David G. Neschis, MD, Frank J. Lexa, MD, Julia T. Davis, RN, RVT, Jeffrey P. Carpenter, MD, RVT Recently the North American Symptomatic Carotid Endarterectomy Trial investigators reported a benefit of carotid endarterectomy compared with medical therapy for symptomatic patients with 50% or greater carotid stenosis. This has necessitated the development of screening parameters for diagnosis of 50% or greater carotid stenosis on the basis of the reference standards used in the study by the North American Symptomatic Carotid Endarterectomy Trial. The duplex scans and arteriograms of 110 patients (210 carotid arteries) were reviewed by blinded readers. Duplex measurements of peak systolic velocity and end diastolic velocity were recorded, and the ratio of these velocities in the internal and common carotid arteries was calculated. The criteria determined for detection of 50% or greater stenosis were as follows: peak systolic velocity of the internal carotid artery greater than 170 cm/s (sensitivity, 92%; specificity, 90%; positive predictive value, 92%; negative predictive value, 90%; and accuracy, 91%); end diastolic velocity of the internal carotid artery greater than 60 cm/s (sensitivity, 92%; specificity, 86%; positive predictive value, 95%; negative predictive value, 79%; and accuracy, 91%); ratio of peak systolic velocity of the internal carotid artery to peak systolic velocity of the common carotid artery greater than 2 (sensitivity, 93%; specificity, 75%; positive predictive value, 83%; negative predictive value, 89%; and accuracy, 85%); and ratio of end diastolic velocity of the internal carotid artery to end diastolic velocity of the common carotid artery greater than 2.4 (sensitivity, 96%; specificity, 79%; positive predictive value, 88%; negative predictive value, 92%; and accuracy, 89%). It is concluded that 50% or greater carotid artery stenosis can be reliably determined by duplex criteria. The use of receiver operating characteristic curves allows the individualization of duplex criteria to the clinical situation. Key words: duplex; carotid ultrasonography; noninvasive testing; peripheral vascular disease. Abbreviations ACAS, Asymptomatic Carotid Atherosclerosis Study; CCA, common carotid artery; CEA, carotid endarterectomy; DL, diameter of normal-appearing internal carotid artery distal to bifurcation; EDV, end diastolic velocity; ICA, internal carotid artery; MRL, minimal residual lumen diameter; NASCET, North American Symptomatic Carotid Endarterectomy Trial; NPV, negative predictive value; PPV, positive predictive value; PSV, peak systolic velocity; ROC, receiver operating characteristic Received May 8, 2000, from the Departments of Surgery (J.T.D., J.P.C.) and Radiology (F.J.L.), University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and Division of Vascular Surgery, University of Maryland Medical System, Baltimore, Maryland (D.G.N.). Revised manuscript accepted for publication November 14, Address correspondence and reprint requests to Jeffrey P. Carpenter, MD, Hospital of the University of Pennsylvania, 4 Silverstein, 3400 Spruce St, Philadelphia, PA In November 1998 the North American Symptomatic Carotid Endarterectomy Trial (NASCET) investigators reported a significant benefit of carotid endarterectomy (CEA) compared with medical therapy alone for patents who had a transient ischemic attack or nondisabling stroke and had moderate (50% 69%) stenosis of the carotid artery. 1 Because traditional duplex criteria were based on parameters other than those used by the NASCET investigators, new criteria for the evaluation of moderate stenoses are required. 2 We therefore sought to determine duplex criteria for the diagnosis of 50% or greater carotid stenosis based on the reference standards used in the NASCET. Duplex ultrasonography is widely regarded as the preferred noninvasive modality for measuring carotid stenosis because it is safe, reliable, and inexpensive. Some reports have suggested that duplex 2001 by the American Institute of Ultrasound in Medicine J Ultrasound Med 20: , /01/$3.50

2 Duplex Criteria for Carotid Stenosis ultrasonographic measurements may be even more reflective of the actual stenotic lesion as measured directly from explanted specimens than arteriography. 3,4 Because the stroke rate of diagnostic cerebral angiography can be as high as 1.2%, it is advantageous that the study to screen patients for carotid lesions be noninvasive. 5 Duplex ultrasonography can be used as a screening test to select patients for angiography; however, recent reports have supported the use of duplex ultrasonography as the sole modality for characterizing carotid lesions before endarterectomy. 6 9 The goal of this study was to develop duplex criteria that would help identify a 50% or greater stenosis and that would be particularly useful in helping other vascular specialists interested in performing receiver operating characteristic (ROC) curve analysis and validation of their own criteria. Patients and Methods Patients Patients being evaluated for carotid artery stenosis at the Hospital of the University of Pennsylvania during a 24-month interval were identified. Patients who underwent both duplex scanning and cerebral angiography within 1 month of each other were used in the study. The data set comes from 110 patients (210 carotid arteries) and represents all patients during the study interval who had complete arteriographic examinations and duplex scanning data available for review. Arteriography Percutaneous catheter carotid arteriograms were obtained in all patients with at least 2-view or, in most cases, 4-view biplane selective arteriography. These studies were performed with either standard cut film techniques or the use of highresolution digital subtraction imaging ( matrix). The percent stenosis, determined by the arteriogram, was calculated from direct measurements of the maximal stenosis minimal residual lumen diameter (MRL) in the carotid bifurcation region (distal common carotid artery [CCA] and proximal internal carotid artery [ICA]) made with the use of a handheld magnifier marked in 1-mm increments. This was compared with the diameter of the normal-appearing ICA distal to the bifurcation (DL) with the technique described for the NASCET study. We calculated stenosis diameter using the MRL and DL in the equation [1 (MRL/DL)] 100 (Fig. 1). Observers were blinded to both the results of the duplex study and the other observers readings. The first 70 vessels were evaluated by 3 blinded readers, and an interim calculation of interobserver agreement was made. As a result of the near-perfect agreement of the 3 observers (see Results), a single observer completed the remaining 140 carotid arteries, providing a total of 210 carotid arteries for evaluation with complete duplex and arteriographic data. Duplex Doppler Ultrasonography Duplex Doppler ultrasonographic studies were performed on a Hewlett-Packard (Andover, MA) Sonos 1000 color duplex system with the use of a 7.5-MHz linear array transducer with a 5.6-MHz Doppler frequency. The cervical ICA, CCA, and external carotid artery were examined. Velocity waveforms were obtained routinely from the CCA at the base of the neck just proximal to the carotid bifurcation; the proximal, mid, and distal Figure 1. Diagram of carotid bifurcation demonstrating sites of interrogation for duplex velocities, indicated by solid circles, as well as locations of angiographic diameter measurements. A indicates diameter of maximal stenosis; B, estimated diameter of the carotid bulb; and C, diameter of the ICA without stenosis just distal to the carotid bulb. Technique used in the NASCET and ACAS and recommended by the Committee on Standards for Noninvasive Testing by the Joint Council of the Society for Vascular Surgery: [1 (A/C)] 100%. Technique used by the European Cooperative Trial of Carotid Endarterectomy and University of Washington criteria: [1 (A/B)] 100%. ECA indicates external carotid artery; and Prox., proximal. 208 J Ultrasound Med 20: , 2001

3 Neschis et al ICA; and the external carotid artery. In addition, velocity waveforms were obtained from any location where stenosis was suspected by either B-mode appearance or color flow mapping. The highest peak systolic velocity (PSV) and end diastolic velocity (EDV) were recorded from each location. Analysis Maximal PSV and EDV in the carotid bifurcation region (distal CCA or ICA [ and ]) were used for comparison with maximal angiographic stenosis. The maximal carotid bifurcation and were compared with the maximal PSV or EDV in the proximal CCA low in the neck (PSV CCA and EDV CCA ), and their ratios ( and ) were calculated. Sensitivity was defined as the number of truepositive results divided by the sum of truepositive and false-negative results; specificity was defined as the number of true-negative results divided by the sum of true-negative and false-positive results; positive predictive value (PPV) was defined as the number of true-positive results divided by the sum of true-positive and false-positive results; negative predictive value (NPV) was defined as the number of truenegative results divided by the sum of truenegative and false-negative results; and accuracy was defined as the sum of true-positive and true-negative results divided by the total number of studies performed. Receiver operating characteristic curves were generated to predict 50% or greater angiographic stenosis. These curves describe sensitivity, specificity, PPV, NPV, and accuracy of each criterion (,,, and / EDV CCA ). Interobserver variability for interpretation of arteriographic stenosis was assessed with the κ statistic, in which the degree of agreement between different readers was defined by the scale of Landis and Koch 10 : less than 0.00, poor; 0.00 to 0.20, slight; 0.21 to 0.40, fair; 0.41 to 0.60, moderate; 0.61 to 0.80, substantial; and 0.81 to 1.0, almost perfect. evaluated. Interobserver variability for the first 70 carotid arteriograms selected at random and evaluated by 3 observers was almost perfect (κ = 0.86). Plots of the degree of arteriographic stenosis versus the cardinal duplex parameters (PSV, EDV,, and ) are shown in Figures 2 5. From these data, the sensitivity, specificity, PPV, NPV, and accuracy for various values and combinations of the,,, and were determined and are shown in Figures 6 9 and discussed in more detail below. Suggested criteria for determination of 50% or greater stenosis are shown in Table 1. Combined criteria for determination of 50% or greater stenosis are shown in Table 2. The highest accuracy was at a PSV of greater than 170 cm/s. At this velocity high accuracy was present without significantly sacrificing sensitivity (92%). At lower velocities, sensitivity was higher but associated with a precipitous drop in specificity (90% at >170 cm/s). Additionally, PPV was high at 92%, and NPV was high at 90%. At higher velocities, NPV declined with only a slight increase in specificity and PPV. Data are shown in Figure 6. Figure 2. PSV versus arteriographic stenosis. Results A 50% or greater arteriographic stenosis was present in 99 cases (47%), not including 17 occluded ICAs (8%), of the 210 carotid arteries J Ultrasound Med 20: ,

4 Duplex Criteria for Carotid Stenosis The highest accuracy was at an EDV of greater than 60 cm/s (91%). Similarly, sensitivity was high at 93%, and specificity was high at 86%. Higher velocities were associated with a sudden fall in sensitivity, and lower velocities were associated with an equally sudden fall in specificity. At an EDV of greater than 60 cm/s, PPV was 96% and NPV was 79%. Data are shown in Figure 7. Figure 3. EDV versus arteriographic stenosis. Although the greatest accuracy was at a ratio of 2.6, this would be at the expense of sensitivity (84%). At a ratio of greater than 2.0, sensitivity was 93%, with specificity of 75% and accuracy of 85%. At this ratio, PPV was acceptable at 83%, and NPV was acceptable at 89%. Data are shown in Figure 8. At a ratio of greater than 2.4, high accuracy was maintained (89.5%), and high sensitivity (96%) and NPV (92%) were preserved as well. At higher ratios, accuracy increased only slightly, but there was a concomitant loss of sensitivity and NPV. At a ratio of greater than 2.4, PPV was 86% and specificity was 79%. Data are shown in Figure 9. Figure 4. PSV ratio ( ) versus arteriographic stenosis. Combined Criteria When all 4 criteria were used together ( >170cm/sec, >60cm/sec, >2.0, and >2.4), sensitivity was 95% and specificity was 100%, with accuracy of 97% (Table 2). PPV was 100% as well. Discussion Figure 5. EDV ratio ( ) versus arteriographic stenosis. It is important that criteria be developed that incorporate the findings of and methods used in the large multicenter studies that help direct clinical decision making. The NASCET investigators initially reported a beneficial effect of CEA when compared with medical therapy for symptomatic patients with 70% or greater carotid stenosis. 11 Accordingly, we developed duplex criteria appropriate to the NASCET findings for use in the vascular laboratory. 12 After the recent NASCET report of a beneficial effect of CEA in symptomatic patients with lesions of 50% or greater, we sought to determine duplex criteria for 50% or greater lesions that were validated by the angiographic methods used in the NASCET J Ultrasound Med 20: , 2001

5 Neschis et al Specifically, angiographic stenosis was measured by comparison of the diameter of maximal stenosis in the ICA or distal CCA with that of the normal ICA just distal to the carotid bulb. This was also the method used in the Asymptomatic Carotid Atherosclerosis Study (ACAS) 13 but not in the European trial that compared the diameter of maximal stenosis with the estimated diameter of the carotid bulb. 14 This is an important distinction, because 80% stenosis in the European trial is approximately equivalent to 60% stenosis in the NASCET and ACAS trials (50% stenosis by the European trial method could be read as 0% stenosis by the NASCET method). Traditional duplex criteria were developed by comparison with arteriographically measured lesions by the European trial method and are therefore not applicable to the more recent multicenter trials. 2 The Committee on Standards for Noninvasive Testing by the Joint Council of the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascular Surgery recommend using the method adopted by the NASCET and ACAS trials. 15 Various duplex criteria for 50% or greater stenosis based on the NASCET method for evaluating angiograms have been published previously (Table 3) Our criteria are appropriate for screening; however, it is important to note that such criteria need to be individualized to the clinical situation. For instance, if a screening test were desired to decide who should undergo arteriography, criteria with high sensitivity and a high NPV would be necessary, making it unlikely that a lesion would be missed. On the other hand, if the decision to perform endarterectomy were based solely on the duplex evaluation, high specificity and a high PPV would be required to avoid unnecessary surgery. We sought to provide criteria suitable for a screening tool as well as criteria that would minimize false-positive results. Of the criteria examined, greater than 170 cm/s and greater than 60 cm/s gave the highest accuracy (91% each). At these velocities sensitivity was not significantly sacrificed (92% each). When used in combination, accuracy rose to 95% and PPV rose to 97%. Peak systolic velocities greater than 210 cm/s are associated with a PPV of 95% and might be an appropriate criterion for use as a sole preoperative imaging study. EDV is useful for severe Figure 6.. The highest accuracy (91%) was found at 170 cm/s. This value allowed for high sensitivity (92%) as well. Lower velocities would give higher sensitivity but with a precipitous drop in specificity and PPV. Although 170 cm/s is an acceptable value for use as a screening test, a velocity of 210 cm/s may be considered for criteria if the test is to be used as the sole preoperative modality because of the higher PPV (95%). ACCUR indicates accuracy; SENS, sensitivity; and SPEC, specificity. stenoses in which aliasing is occasionally problematic for the determination of PSV. As discussed, although an EVD ICA of 60 cm/s may be a good screening threshold, velocities greater than 70 cm/s provide a 98% PPV and sensitivity of 89%. Figure 7.. At a velocity of 60 cm/s, the highest accuracy was achieved (91%). At this velocity, sensitivity (93%) and specificity (86%) were relatively high. Lower velocities were associated with a dramatic fall in specificity. Higher velocities were associated with a marked drop in NPV. If used for planning surgery, an EDV of 70 cm/s would provide a 98% PPV and sensitivity of 89%. ACCUR indicates accuracy; SENS, sensitivity; and SPEC, specificity. J Ultrasound Med 20: ,

6 Duplex Criteria for Carotid Stenosis Figure 8.. Although the highest accuracy was at a ratio of 2.6 (87%), this sacrificed too much sensitivity (84%). The best balance is found at a ratio of 2.0, giving sensitivity of 93% and accuracy of 85%. ACCUR indicates accuracy; SENS, sensitivity; and SPEC, specificity. Ratios of PSV and EDV are useful for overcoming variability in isolated PSV and EDV measurements as a result of changes in blood pressure and hemodynamic effects of contralateral stenoses or occlusions, as well as ipsilateral tandem lesions. Although the greatest accuracy was found at a ratio of 2.6, the best combination of factors was at a ratio of 2.0, which provided sensitivity of 93% and accuracy of 85%. At this ratio PPV was 83% and NPV was Figure 9.. Here the best balance of accuracy and sensitivity was at an EDV ratio of 2.4. At this velocity sensitivity was 96% and accuracy was 89.5%. Even at considerably higher velocities, the PPV did not appreciably improve (92% at 2.4). ACCUR indicates accuracy; SENS, sensitivity; and SPEC, specificity. 89%. In evaluating, a ratio of 2.4 maintained high accuracy (89.5) and preserved sensitivity (96%) and NPV (92%). Even at considerably higher velocities, the EDV did not appreciably improve (92% at 2.4). When all 4 criteria were met, 97% accuracy and a 100% PPV were achieved. A simpler combination of and does not significantly reduce accuracy (95%), sensitivity (97%), and PPV (97%). Our conclusions are similar to those of other authors who have reported criteria for evaluating carotid arteries with 50% or greater stenosis (Table 3). Winkelaar et al 16 determined that the most accurate criterion for 50% or greater stenosis was a ratio of 2.0 or greater. Although we found the single most accurate criterion to be a PSV of greater than 170 cm/s, the threshold level we determined for the ratio was also 2.0. The data reported by Winkelaar and colleagues 16 revealed that a PSV of 170 was associated with high accuracy (91%) and sensitivity (92.5%), which were very similar to our findings. Additionally, these authors reported high accuracy (93.8%) using an EDV of 60 cm/s. AbuRahma et al 18 determined that the optimum criterion for a 50% or greater lesion was a PSV of 140 cm/s or greater. This value was associated with accuracy of 93%, sensitivity of 92%, specificity of 95%, PPV of 97%, and NPV of 89%. This was similar to our threshold PSV of 170 cm/s. Interestingly, the criteria AbuRahma and colleagues 18 chose for EDV and the ratio were very similar to ours, namely, EDV of 60 cm/s and ratio of As with our experience, the distinction between 50% and 60% was very slight. Faught et al 17 reported that the most useful criteria to determine a lesion with 50% to 69% stenosis was a combination of the PSV of greater than 130 cm/s and EDV of 100 cm/s or less. As in our study, Faught and colleagues 17 demonstrated a threshold ratio of 2.0 or greater for lesions greater than 50%. It is interesting that the velocity criteria that we believe are the best for assessing a lesion of 50% or greater are similar to the criteria we reported in 1995 for evaluation of lesions of 60% or greater. 19 This demonstrates some limitation in the ability of duplex ultrasonography to distinguish subtle differences between stenoses of moderate severity. 212 J Ultrasound Med 20: , 2001

7 Neschis et al Table 1. Suggested Criteria for Determination of 50% or Greater Carotid Artery Stenosis Criterion Sensitivity, % Specificity, % PPV, % NPV, % Accuracy, % >170 cm/s >60 cm/s > > Table 2. Combined Criteria for Determination of 50% or Greater Carotid Artery Stenosis Combined Criteria Sensitivity, % Specificity, % PPV, % NPV, % Accuracy, % PSV + EDV PSV + PSVR PSV + EDVR EDV + PSVR EDV + EDVR PSVR + EDVR PSV + EDV + PSVR PSV + EDV + EDVR EDV + PSVR + EDVR PSV + PSVR + EDVR PSV + EDV + PSVR + EDVR PSVR indicates ratio; EDVR, ratio. Table 3. Reports Describing Criteria for 50% or Greater Carotid Stenosis Report Criterion Faught et al 17 AbuRahma et al 18 Winkelaar et al 16 This Study PSV, cm/s * 170 Accuracy, % Sensitivity, % Specificity, % PPV, % NPV, % NR 90 EDV, cm/s NR 50* 50* 60 Accuracy, % Sensitivity, % Specificity, % PPV, % NPV, % 76 NR 79 NR 2.12* Accuracy, % Sensitivity, % Specificity, % PPV, % NPV, % 77 NR 89 NR NR NR 2.4 Accuracy, % 89 Sensitivity, % 96 Specificity, % 79 PPV, % 88 NPV, % 92 NR indicates not reported. *Velocities or ratios with the highest accuracy per category but not specifically recommended by the authors. J Ultrasound Med 20: ,

8 Duplex Criteria for Carotid Stenosis We believe that the generation of comprehensive ROC curves incorporating sensitivity, specificity, PPV, NPV, and accuracy greatly enhances development of duplex criteria. ROC curves allow one to tailor selected criteria to the needs of the particular institution and clinical situation. For instance, if one had available excellent angiography with a low complication rate, criteria could be selected for screening patients who would undergo angiography. Such criteria would focus on high sensitivity and NPV to avoid missing any potentially correctable lesions. However, if one were to determine criteria to select patients for CEA, one would focus on criteria with a high PPV and specificity, so that unnecessary operations would be avoided even at the expense of missing a few moderate lesions. It is important that individual institutions develop their own ROC curves and establish criteria for determination of carotid artery stenosis. References 1. Barnett HJ, Taylor DW, Eliasziw M, et al. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators [see comments]. N Engl J Med 1998; 339: Roederer GO, Langlois YE, Chan ATW, et al. Ultrasonic duplex scanning of the external carotid arteries: improved accuracy using new features from the common carotid artery. J Cardiovasc Ultrasonography 1982; 1: Pan XM, Saloner D, Reilly LM, et al. Assessment of carotid artery stenosis by ultrasonography, conventional angiography, and magnetic resonance angiography: correlation with ex vivo measurement of plaque stenosis. J Vasc Surg 1995; 21:82 88, discussion Alexandrov AV, Bladin CF, Maggisano R, Norris JW. Measuring carotid stenosis. Time for a reappraisal [see comments]. Stroke 1993; 24: Clinical advisory: carotid endarterectomy for patients with asymptomatic internal carotid artery stenosis. Stroke 1994; 25: Gertler JP, Cambria RP, Kistler JP, et al. Carotid surgery without arteriography: noninvasive selection of patients. Ann Vasc Surg 1991; 5: Gelabert HA, Moore WS. Carotid endarterectomy without angiography. [review]. Surg Clin North Am 1990; 70: Moore WS, Ziomek S, Quinones-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: Su LT, Carpenter JP. Decreasing carotid endarterectomy length of stay at a university hospital. Cardiovasc Surg 1999; 7: Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33: Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators [see comments]. N Engl J Med 1991; 325: Carpenter JP, Lexa FJ, Davis JT. Determination of duplex Doppler ultrasound criteria appropriate to the North American Symptomatic Carotid Endarterectomy Trial. Stroke 1996; 27: Study design for randomized prospective trial of carotid endarterectomy for asymptomatic atherosclerosis. The Asymptomatic Carotid Atherosclerosis Study Group. Stroke 1989; 20: MRC European Carotid Surgery Trial: interim results for symptomatic patients with severe (70 99%) or with mild (0 29%) carotid stenosis. European Carotid Surgery Trialists Collaborative Group [see comments]. Lancet 1991; 337: Thiele BL, Jones AM, Hobson RW, et al. Standards in noninvasive cerebrovascular testing. Report from the Committee on Standards for Noninvasive Vascular Testing of the Joint Council of the Society for Vascular Surgery and the North American Chapter of the International Society for Cardiovascular Surgery. J Vasc Surg 1992; 15: Winkelaar GB, Chen JC, Salvian AJ, Taylor DC, Teal PA, Hsiang YN. New duplex ultrasound scan criteria for managing symptomatic 50% or greater carotid stenosis. J Vasc Surg 1999; 29: Faught WE, Mattos MA, van Bemmelen PS, et al. Color-flow duplex scanning of carotid arteries: new 214 J Ultrasound Med 20: , 2001

9 Neschis et al velocity criteria based on receiver operator characteristic analysis for threshold stenoses used in the symptomatic and asymptomatic carotid trials. J Vasc Surg 1994; 19: , discussion AbuRahma AF, Robinson PA, Strickler DL, Alberts S, Young L. Proposed new duplex classification for threshold stenoses used in various symptomatic and asymptomatic carotid endarterectomy trials. Ann Vasc Surg 1998; 12: Carpenter JP, Lexa FJ, Davis JT. Determination of sixty percent or greater carotid artery stenosis by duplex Doppler ultrasonography [see comments]. J Vasc Surg 1995; 22: , discussion J Ultrasound Med 20: ,