Duplex Carotid Sonography Peak Systolic Velocity in Quantifying Internal Carotid Artery Stenosis

Duplex Carotid Sonography Peak Systolic Velocity in Quantifying Internal Carotid Artery Stenosis Cynthia E Withers, MD", Barbara B Gosink, MD", Alison M Keightley, MD", Giovanna Casola, MD", Arthur A Lee, MD, Eric vansonnenberg, MD"t, John F Rothrock, MD:j:, Patrick D Lyden, MD:J: Duplex ultrasonography combining highresolution imaging and Doppler spectrum analysis was performed in 92 consecutive patients (total, 80 vessels) and com pared with the findings of conventional arteriography All duplex studies were categorized into four groups based upon the maximum internal carotid artery (ICA) velocity: group : < 25 cmjsec; group 2: 25 to 224 cmjsec; group 3: >225 cmfsec; and group 4: no flow Sensitivities and specificities were highest when peak ICA velocity was used as one of several criteria in quantifying the degree of ICA stenosis These additional cri teria were: () the presence of extensive sonographically visible plaque within the ICA; (2) an abnormal spectral waveform with elevated diastolic velocity(> 00 cmfsec); (3) resistive pattern ("externalization") of the common carotid artery (CCA) waveform; and (4) the ratio of the right CCA peak velocity to the left of <07 or >3 The overall accuracy for the combined groups using all criteria was 94% KEY WORDS: carotid arteries, stenosis or obstruction; carotid arteries, ultrasound studies; ultrasound, Doppler studies (/ Ultrasound Med 9:345, 990) D uplex sonography of the carotid arteries is widely used in the screening of patients with atherosclerotic disease Pulsed or continuouswave Doppler is combined with highresolution realtime sonography to estimate the degree of carotid artery stenosis Reliable criteria are required to predict accurately the severity of carotid occlusive disease Several studies have proposed criteria for an abnormal duplex examination Blackshear and colleagues report that the ratio of mean velocity at peak systole in the internal carotid artery (ICA) to that in the common carotid artery (CCA) accurately indicates the degree of ICA stenosis 2 A velocity ratio (VICA/VCCA) below 08 in Received August, 989, from the Departments of Radiolog~ fmedicine, and ;Neurosciences, University of California Medical Center, San Diego, California Revised manuscript accepted for publication December 7, 989 Address correspondence and reprint requests to Dr Withers: Department of Radiolog~ Sunnybrook Medical Centre, 2075 Bayview Avenue, Toronto, Canada M4N 3MS dicates no stenosis while a velocity ratio above 5 determines a highgrade stenosis (60% or greater diameter reduction) Garth et al propose five criteria for a flowlimiting stenosis (SO% or greater diameter reduction): a ratio of the maximum ICA flow velocity to the maximum CCA flow velocity of 5 or more; turbulence, as indicated by an ICA velocity spectral width of 40 cmfsec or more; visible plaque that produces an ICA stenosis of SO% or more; a maximum ICA velocity of 00 cmfsec or more; and an inability to detect ICA flow 3 In the study by Vaisman and Wojciechowski, two additional criteria in the prediction of significant disease (greater then 50% diameter reduction) are proposed: peak systolic average velocity in the CCA of less than 25 cmfsec and the ratio of right to left peak systolic average velocities in the CCA of less than 07 or greater than 3 4 In a recent study, Robinson et al distinguished four categories of ICA cross sectional area (rather than diameter) stenosis: < 70%, 70% to 89o/o, 90% to 99%, and occlusion 5 This study found that peak systolic velocity, systolic velocity ratios, and enddiastolic velocity were equivalent pre 990 by the American Institute of Ultrasound in Medicine J Ultrasound Med 9:345349, 990 02784297/90/$350

346 DUPLEX CAROTID SONOGRAPHY J Ultrasound Med 9:345 349, 990 dictors of significant stenosis Combined parameters did not significantly improve diagnostic accuracy As the carotid lumen narrows, flow velocity rises Absolute flow velocity in the ICA is influenced by the patient's blood pressure, cardiac output, vessel size and elasticity, peripheral resistance, and the presence of dis ease in the ipsilateral CCA or the contralateral CCA or ICA 6 The arterial stenosis may become so severe that systolic velocity decreases to normal or below normal values due to the markedly elevated resistance to flow Because of these factors, reliance on systolic ICA velocity measurements alone may result in inaccurate prediction of the severity of disease There is considerable variety in technique of measurement (peak velocity, average velocity, frequency) Like Robinson, we feel criteria are required to diagnose highergrade (>75% diameter) stenosis that may be considered for surgical intervention There is a need for a practical technique of evaluating JCA stenosis that can be used with current duplex instrumentation, is readily reproducible, and can distinguish <75% from >75% stenosis In this prospective study, we wish to describe our experience using maximum ICA flow velocity as a mea sure of ICA stenosis MATERIALS AND METHODS Between December 985 and December 988, 2,890 patients underwent duplex carotid sonography Of these, 92 patients had conventional carotid arteriography within one month of their ultrasound study One hundred eighty carotid arteriograms were performed; in four patients, only one side was studied The angiegraphic percent stenosis of the ICA was measured in two projections based upon the reduction in diameter of the residual lumen when compared with the diameter of the more distal JCA Findings of arteriography and/or surgery were used as the standard by which duplex sonography was evaluated Most duplex exams were performed on an Acuson 28 machine (5MHz transducer for both Doppler insonation and imaging) Some of the earlier scans were done on a Technicare Autosector 5 scanner (5MHz pulseddoppler transducer and a 75 MHz imaging transducer) Realtime imaging was performed in multiple planes from the proximal CCA to the maximal distal point of visualization of the ICA, usually approximately 3 em beyond the carotid bifurcation When possible, the degree of diameter stenosis was measured directly from the transverse images of the JCA The proximal external carotid artery (ECA) was also evaluated The audible signal of Doppler shift frequency was used to document and localize flow disturbance and to confirm the data provided by quantitative Doppler spectrum analysis Fast Fourier transformation of the Doppler signal provided calculation of the peak systolic velocity The flow velocities and velocity spectral widths were sampled using a 2 or 4 mm pulse gate in the proximal, mid, and distal CCA and ICA Flow velocities were also recorded in the ECA; however, because of their limited clinical relevance, they will not be discussed Confirmation of occlusion was made with continuouswave Doppler or widepulse gate All studies for each vessel were categorized into four empirically chosen groups of diameter stenosis depend ing upon the maximum velocity in the JCA (Table ) Receiver operating characteristic (ROC) curves were generated to determine the cut off velocities that resulted in an optimal balance between high sensitivity and specificity The falsepositive rate was kept below 0% In our series, the sensitivity in predicting stenoses less than 50% was higher when using peak systolic ICA velocities < 25 cmjsec than when using <00 cmjsec as in Garth's study The ROC curve for various velocities at the 50% diameter stenosis level is shown in Figure Because of the relatively few vessels in groups 2 and 3, there was a range of velocities that provided equal re suits Therefore, we chose velocities for groups 2 and 3 that are easy to remember Four comparable angiegraphic groups were selected These groups reflect the major clinically significant categories of ICA stenosis (Table 2) Group consists of those vessels with less than 50% diameter stenosis Group 2 (50% to 74% diameter stenosis) is traditionally considered "hemodynamically significant" At our institution, patients in group 2 are treated medically and followed with duplex studies With group 3 (75o/o to 99% diameter stenosis), many surgeons will consider carotid endarterectomy in symptomatic patients Group 4, documented occlusion, generally precludes surgery These categories are comparable to the previously mentioned crosssectional area steno ~ sis measurements described by Robinson Comparison between the ultrasound and the angiog raphy group was then made to determine if these broad categories could be used for clinically useful decisions as Table : Categories of Degree of Arterial Narrowing by Sonography Group 2 3 4 ICA internal carotid artery Peak ICA Velocity (cmjsec) < 25 25224 >225 NoOow

0 N 0 J Ultrasound Med 9:345 349, 990 WITHERS ET AL 347 Cll : = 08 06 R 0 C curve I or <50% S I en o sis I o o N ~~ ~ ~~~ ~, ~ ~ ~ ~ ' ~ r r ~ r ~~ ~ I~ I ~ : c ~ ; : ; : ~~ ~ B ~, ~ "" _ :: ~, : M, ~ r r r r ~ ~,, r, r r a ~ I I I I I p I p j A ~ : ~ : ~ ' ' : r = _ ~ r! > = 0 N 0 N Ill 04 ~ ~ r + ~ ~ ~? r + ~ ~ ~~ ~~ ~ I I I ~ ' ' Gl ' ' : ' ' I ~ ~ I~ I I ~ I I I I I I ~ I I r I I I I I ra I I I I~ r, r I r r ~ I r ~I~ ~ I~~ I ~ I I~ I PI I~ I~ r I~ r r r r ~I i 02 I I I e 0 I I I I r ' ' I I ~ "I"" N I ~ I ~ I ~ I ~ I I~ I I r ~ I A I I p ~ I r r I I I~ I I~ r I IN I ~ ~ ~ I~~ I ~ I ~ I~~~ I I ~ r I I~ ~ I I r 0 02 04 OG 08 False Positive Fraction Figure Receiver operating characteristic (ROC) curve for peak systolic velocity in inter nal carotid artery (ICA) at <SO% diameter stenosis, Point A represents 00 cmjsec as the peak ICA velocity threshold Points B and C represent 25 cmjsec and 75 cmjsec, respec tively to degree of ICA stenosis based on duplex findings alone In groups 2 and 3, we identified a subset of six patients (total, six vessels) with significant stenoses in whom the peak systolic velocities were artifactually low In five of the patients, all with stenoses > 90%, this was felt to be due to the greatly narrowed channel available for blood flow In the sixth patient, whose angiogram showed a 60% stenosis, the low velocity was most likely due to sampling error, ie, failure to insonate the point of maximum stenosis In this subset of patients with artifactually low velocities, one or more additional duplex criteria were used to correctly diagnose the stenosis These criteria included: Extensive plaque visible within the ICA on the transverse image (one vessel) 2 An abnormal ICA spectral waveform with elevated diastolic velocity (>00 cmfsec) (three vessels) 3 Resistive pattern ("externalization") of the CCA waveform This is represented by very low diastolic ve locity which is normally characteristic of the ECA (two vessels)' 4 Decreased velocity of the CCA ("damped" wave form) on the affected side versus the unaffected side (three vessels) This is identified by the ratio of the right CCA peak velocity to the left of <07 or >3 4 These four ancillary findings were also seen alone or in combination in many other patients whose velocity criteria alone were consistent with the correct diagnosis The sensitivity specificity and overall accuracy of these ancillary criteria used individually to diagnose significant stenosis were not addressed in this paper RESULTS Initially we compared estimated stenosis by duplex exam with the results of arteriography and surgery using peak ICA velocity as the sole criterion Table 3 summarizes our findings Table 2: Categories of Degree of Arterial Narrowing by Angiography Group Percent arteriogram diameter stenosis 0 49 2 5074 3 75 99 4 Occlusion Table 3: Angiographically Determined Stenosis Versus Peak ICA Velocity Group <'b stenosis) Sensitivity of Specificity of angiogram ultrasound ultrasound (0 49) 94% (73f78) 93% (98/02) 2 (50 74) 75% (5/20) 94% (5/60) 3 (75 99) 88% (5f58) 97% (8/22) 4 (Occlusion) 96% (23/24) 99% (55/56)

348 J Ultrasound Med 9:345 349, 990 DUPLEX CAROTID SONOGRAPHY Table 4: Comparative Accuracy Using Peak ICA Velocity and Additional Criteria Group (% stenosis) angiogram (0 49) 2 (50 74) 3 (75 99) 4 (Occlusion) Sensitivity of ultrasound Specificity of ultrasound 94% (73/78) 80% (6/20) 98% (57J58) 96% (23/24) 97% (55/60) 97% (8/22) 99% (0/02) 99% (55/56) Overall accuracy = 94% We then reexamined our data invoking the additional criteria of extensive visible plaque, elevated diastolic velocity (> 00 cmjsec), resistive pattern ("externalization") of the CCA wave form, and abnormal right CCA/ left CCA velocity ratios as outlined in Materials and Methods (see also Table 4) Sensitivities and specificities are all > 90% except in category 2 where the sensitivity is only 80o/o (Table 4)In four patients in group 2, Doppler velocities overestimated the degree of stenosis, due to either a contralateral ICA occlusion or contralateral tight ICA stenosis When the contralateralica is markedly narrowed, the redistribution of blood flow may yield elevated velocities in the presence of a normal caliber vessel The positive and negative predictive values for each category are summarized in Table 5 Overall accuracy for the combined groups using the ancillary criteria was 94% DISCUSSION The role of endarterectomy in the therapy of carotid atherosclerotic disease is controversiaj u At our institution, carotid endarterectomy may be recommended in symptomatic patients with an angiographically demon strated ICA stenosis of 75% or more For this reason, we have attempted to distinguish stenoses of less than 75% from those that are more severe In no case did duplex sonography misjudge the degree of stenosis by more than one category Because carotid endarterectomy is performed only in the patent vessel, a Table S: Predictive Values of Ultrasound for Detection of ICA Stenosis Group (tt'o stenosis) angiogram (049) 2 (50 74) 3 (75 99) 4 (Occlusion) Positive predictive value of ultrasound Negative predictive value of ultrasound 99% (73/74) 76% (6/2) 93% (57/6) 96% (23/24) 95% (0/06) 97% (55 /59) 97% (8/9) 99% (55/56) potentially serious error is failing to detect occlusion of the ICA or predicting occlusion when in fact a high~rade stenosis is present Thus arteriographic confirmation of ICA stenosis or occlusion is critical prior to surgery In our study, there was one falsepositive and one falsenegative diagnosis of ICA occlusion The falsepositive error occurred early in our experience and was due to failure to confirm the apparent occlusion by reexamination of the vessel with either continuous wave or a wide pulse Angiography demonstrated a 98% stenosis The small jet of flow was not detected by pulsed Doppler Colorflow Doppler will likely reduce such falsepositive diagnoses of ocdusion In the falsenegative error of ICA occlusion, duplex sonography estimated a 7So/o or greater stenosis (peak ICA velocity 52 cmjsec) The angiogram performed 4 weeks later demonstrated totallca occlusion It is possible that the vessel ocduded in the interim This patient had an ipsilateral 98% stenosis of the ECA and underw went an external carotid endarterectomy Surgery confirmed the ICA occlusion Of particular interest is one patient in our series with a falsepositive arteriographic diagnosis of occlusion In this patient, duplex examination showed markedly elevated velocities in the ICA compatible with a highgrade stenosis Arteriography, however, demonstrated total ICA occlusion Th~ ~ltrasound was repeated and again showed hagh velocataes On the basis of the duplex studies, ~he patient was scheduled for an endarterectomy, a~d mdeed ~t surgery the ICA was patent This example raases questions regarding the use of arteriography as the gold standard When peak ICA velocities are used as the sole measure to predict ICA stenosis, the degree of luminal narrowing may be over or underestimated This is because absolute velocity is influenced by multiple physiologic factors such as hypertension or hypotension, cardiac output, and the presence of contralateral disease The group most influenced by these factors is group 3 (75% to 99% stenosis) Beyond a 90% diameter stenosis velocities may decrease to normal or subnormallev:ls7 We found this to be true in a minority of patients (six); all but one had an extremely tight stenosis (> 90%) This is an important subset, however, and use of the following additional criteria will increase diagnostic accuracy in the presence of artifactually low peak systolic velocity measurements: extensive ICA plaque; 2 elevated diastolic velocity (>00 cmjsec); 3 resistive pattern ("externalization") of the CCA waveform; 4 abnormal peak velocity of the right compared to the left CCA

J Ultrasound Med 9:345349, 990 WITHERS ET AL 349 An additional useful criterion, not addressed in this study, is the use of ratios of peak systolic velocity in the ICA to peak systolic velocity in the CCA 3 s,u Elevated ratios may be useful in identifying stenosis in patients with decreased myocardial contractility and consequent reduced velocity throughout the carotid system Duplex sonography is a valuable screening method for carotid artery disease Analysis of the peak ICA velocity is an important criterion in the quantification of the degree of stenosis Care must be taken to use ancillary criteria, particularly when diagnosing stenoses greater than 90% REFERENCES Blackshear WM Jr, Phillips OJ, Thiele BL, et al: Detection of carotid occlusive disease by ultrasonic imaging and pulsed Doppler spectrum analysis Surgery 86:698, 979 2 Blackshear WM Jr, Phillips OJ, Chikos PM, et al: Carotid artery velocity patterns In normal and stenotic vessels Stroke :67, 980 3 Garth KE, Carroll BA, Sommer FG, et al: Duplex ultrasound scanning of the carotid arteries with velocity spectrum analysis Radiology 47:823, 983 4 Vaisman U, Wojciechowski M: Carotid artery disease: New criteria for evaluation by sonographic duplex scanning Radiology 58:253, 986 5 Robinson ML, Sacks 0, Perlmutter GS, et al: Diagnostic criteria for carotid duplex sonography AJR 5:045, 988 6 Zwiebel WJ: Analysis of carotid Doppler signals In: Zwiebel WJ (ed): Vascular Ultrasonography, ed 2 Or lando, Grune & Stratton, 986, p 80 7 jacobs NM, Grant EG, Schellinger 0, et al: Duplex carotid sonography: Criteria for stenosis, accuracy and pitfalls Radiology 54:385, 985 8 Levin SM, Sondheimer FK, Levin JM: The contralateral diseased but asymptomatic carotid artery: To operate or not? Am J Surg 40:203, 980 9 Turnipseed WD, Berkoff HA, Belzer FO: Postoperative stroke in cardiac and peripheral vascular disease Ann Surg 92:365, 980 0 Breslau PJ, Fell G, Ivey TO, et al: Carotid arterial disease in patients undergoing coronary artery bypass operations J Thorac Cardiovasc Surg 82:765, 98 Barnes RW, Leibman PR, Marszalek PB, et al: The natural history of asymptomatic carotid disease in patients undergoing cardiovascular surgery Surgery 90:075, 98 2 Baker JD, Marich KW, Bluth El, et al: Standardized imaging and Doppler criteria for cerebrovascular diagnosis using duplex sonography Presented at the AlUM Annual Convention, Las Vegas, September, 986