From the Midwestern Vascular Surgical Society

B-mode ultrasonographic characterization ofcarotid atherosclerotic plaques in symptomaticand asymptomatic patientsJennifer K. Grogan, MD, Wael E. Shaalan, MD, Hongwei Cheng, MD, Bruce Gewertz, MD,Tina Desai, MD, Gretchen Schwarze, MD, Seymour Glagov, MD, Laurie Lozanski, RVT,Andrea Griffin, RVT, Maria Castilla, RVT, and Hisham S. Bassiouny, MD, Chicago, Ill

Objective: To identify features on B-mode ultrasonography (US) prevalent in symptomatic plaques and correlate thesefindings with histopathologic markers of plaque instability.Methods: Carotid endarterectomy (CEA) plaques from symptomatic and asymptomatic patients with critical stenoses(>70%) were qualitatively assessed using preoperative B-mode US for echolucency and calcific acoustic shadowing. USecholucency was quantitated ex vivo using computerized techniques for gray-scale median (GSM) analysis. Histopatho-logic correlates for US plaque echolucency (percentage of necrotic core area) and acoustic shadowing (percentage ofcalcification area) were determined.Results: Fifty CEA plaques were collected from 48 patients (46 unilateral and two bilateral); 26 of these plaques were fromsymptomatic patients. Age, degree of stenosis, and atherosclerotic risk factors were similar for the symptomatic andasymptomatic patients. Using preoperative B-mode US, 58%, 35%, and 7% of symptomatic plaques and 18%, 41%, and41% of asymptomatic plaques were found to be echolucent, echogenic, and calcific, respectively (P < .05). Using ex-vivoB-mode US and GSM analysis, symptomatic plaques were more echolucent (41 � 19) than asymptomatic plaques (60 �13), P < .03. A strong inverse correlation was found between the percent plaque necrotic area core and GSM (R � �0.9,P < .001). Percentage of calcification area in plaques with acoustic shadowing was 66% and only 27% in those withoutacoustic shadowing (P < .05).Conclusions: Using B-mode US, symptomatic plaques are more echolucent and less calcified than asymptomatic plaquesand are associated with a greater degree of histopathologic plaque necrosis. Such features are indicative of plaqueinstability and should be considered in the decision-making algorithm when selecting patients with high-grade

asymptomatic carotid stenosis for intervention. (J Vasc Surg 2005;42:435-41.)

Currently, operative or endovascular intervention forcarotid bifurcation disease is largely predicated on theseverity of internal carotid artery (ICA) arteriographic ste-nosis in both symptomatic and asymptomatic patients.1-3

However, most patients with high-grade asymptomatic ICAstensoses will remain stroke free on long-term follow-up.3,4 Inthis patient cohort with �60% ICA stenosis, the 5-year riskof disabling stroke without any premonitory symptoms is11% to 12%.5-7 These findings indicate that for a givencritical stenosis, other variables such as plaque structuralmorphology and instability play an important role in thenatural history of carotid occlusive disease. Previously, wehave shown that unstable and disrupted plaques exhibitincreased lipid concentration and plaque necrotic core for-mation in association with macrophage/foam cell infiltra-tion in the ablumenal fibrous cap.8,9 We have also demon-

From the Vascular Section, Department of Surgery, University of Chicago.Competition of interest: none.Presented at the Twenty-eighth Annual Meeting of the Midwestern Vascular

Surgical Society, Omaha, Neb, September 10, 2004.Reprint requests: Hisham S. Bassiouny, MD, Vascular Section, Department

of Surgery, University of Chicago, 5841 S. Maryland Ave, MC 5028,Chicago, IL 60637 (e-mail: hbassiou@surgery.bsd.uchicago.edu).

0741-5214/$30.00Copyright © 2005 by The Society for Vascular Surgery.

doi:10.1016/j.jvs.2005.05.033

strated that the degree of plaque calcification is associatedwith plaque stability and a lower risk of clinical ischemicevents.10

Studies correlating such carotid plaque histopathologiccharacteristics with B-mode ultrasonography (US) suggestthat the degree of plaque echolucency is attributed toplaque necrosis or hemorrhage.11-15 Steffen et al16 andothers17-21 have demonstrated that for all degrees of he-modynamically significant stenoses, patients with echolu-cent carotid plaques are at increased risk of symptomaticthromboembolic ischemic cerebrovascular events and cere-bral infarction on computed tomography (CT). Computer-assisted plaque characterization using standardized B-modeUS imaging and digital post-processing now allows thequantitative evaluation of differences in plaque echolu-cency between symptomatic and asymptomatic plaquesusing a gray-scale median (GSM) score.18,20,22,23 To fur-ther define the role of B-mode US in characterizing unsta-ble carotid bifurcation plaques, we studied symptomaticand asymptomatic plaques with similar degrees of steno-sis to determine differences in the prevalence of echolu-cency, echodensity, and acoustic shadowing. After carotidendarterectomy (CEA), GSM scores were calculated fromcross-sectional images of the excised plaques and correlatedwith histomorphometric measures of plaque necrosis. Ad-

ditionally, the presence of preoperative B-mode acoustic

435

JOURNAL OF VASCULAR SURGERYSeptember 2005436 Grogan et al

shadowing was related to the degree of carotid plaquecalcification.

METHODS

Patient population and CEA plaques. Fifty carotidbifurcation plaques were collected from 48 consecutivepatients who underwent CEA for high-grade ICA stenosis.These plaques were also studied in a previous study thatreported the degree of calcification and plaque inflamma-tion in relation to symptomatic outcome.10 Risk factors foratherosclerotic disease, including tobacco use, hyperten-sion, diabetes mellitus, coronary artery disease, peripheralartery disease, and hypercholesterolemia, were recorded forall patients. Indications for CEA were transient ischemicattack (16 plaques), stroke (six plaques), and amaurosisfugax (four plaques) within 6 months of CEA, and high-grade asymptomatic carotid stenosis (24 plaques). Plaqueswere endarterectomized using a standard semi-eversiontechnique that allowed removal of the intact specimen withpreservation of plaque structural integrity and minimizedthe possible disruption of the plaque luminal surface.9

Preoperative color duplex US (CDU). Preoperativenoninvasive evaluation of bilateral carotid arteries was per-formed for every patient using an ATL HDI 3000, ATLHDI 5000 (Advanced Technology Laboratories, Bothell,Wash) or Acuson Sequoia 512 (Acuson Corp., MountainView, Calif) with a linear array 4-7 MHz transducer. Allscans were performed by one experienced registered vascu-lar technologist (R.V.T.). The study included longitudinaland transverse views. Also included were pulsed Dopplerwaveform samples of the subclavian, vertebrals, and exter-nal carotid arteries; two samples of the common carotidartery (CCA) (mid and distal), and three samples of theICA (proximal, mid and distal). All Doppler waveformsamples were recorded from the midstream of the vesselbeing interrogated. The angle of insonation was kept at 60degrees. Color flow mapping was additionally employed toidentify any suspect regions of hemodynamic disturbance.Specific flow parameters calculated included peak systolicvelocity (PSV), end diastolic velocity (EDV), the ratio of

Fig 1. Preoperative high-resolution B-mode ultrasondegrees of echolucency. Echolucency was graded subjecholucent (B) if �50% of its area was involved. A plaqushadowing interfering with pulsed Doppler sampling. I(distal); ECA, external carotid artery.

PSV in the ICA and CCA (ICA/CCA ratio), and the

presence of spectral broadening or turbulence due to ste-nosis. Gray-scale (B-mode) imaging was used as an adjuncttool to detect the presence, location, extent of disease, andthe degree of stenosis when present. Plaque morphologiccharacteristics were subjectively evaluated, and each plaquewas graded as predominantly (�50% area of plaque images)as echolucent � hypoechoic or echodense � hyperechoic(Fig 1A and B). The presence of acoustic shadowing andinterference with pulsed Doppler sampling was indicativeof a calcific plaque (Fig 1C) intraobserver reproducibility ofplaque morphology by B-mode US was assessed, withacceptable results (� � 0.83). Strandness velocity criteriawere used for the diagnosis and grading of carotid disease inthis study.24,25 Additionally, the degree of stenosis basedon B-mode image measurements was recorded.

Ex vivo B-mode US. Ex vivo B-mode US imaging ofintact endarterectomy plaques was performed to quantitateplaque echodensity. Carotid bifurcation plaques were em-bedded in a gelatin basin with a 5-mm interval echogenicfloor grid (Fig 2A). B-mode US scans were performedusing an Acuson Sequoia 512 ultrasound scanner with a10-MHz foot probe. B-mode scan settings were standard-ized using a dynamic range to a linear post-processing curveand a high frame rate and persistence. The probe wasadjusted so that the ultrasonic beam was at a right angle tothe surface of the plaque. Images were magnified using thehigh-definition zoom feature and the time gain compensa-tion (TGC) curve, and overall gain was adjusted to achievean optimum image of the near and far walls. All settingswere saved as a preset in the US machine for use with eachexamination. All B-mode US scans were performed byR.V.T. Each plaque was subjected to cross-sectional B-mode imaging at 2- to 3-mm intervals. Images at eachinterval were captured and saved to a digital file on amagneto-optical disk (MOD). An average of 12 images(range 8-20) was obtained from each plaque for furtherquantitative echodensity analysis (Fig 2B).

Computer-assisted analysis of plaque echodensity.Plaque echodensity was measured in each plaque image andthen averaged to calculate the mean plaque echodensity.

hic longitudinal view of carotid plaque with varyingly as predominantly echogenic (A) or predominantly

s classified as calcific (C) if there was significant acousticinternal carotid artery; CCAD, common carotid artery

ograpectivee waCA,

Using commercially available software (Adobe Photoshop

hen s

JOURNAL OF VASCULAR SURGERYVolume 42, Number 3 Grogan et al 437

3.0), the GSM in the noncalcific regions of the plaque wasused as an objective measure of echodensity, with lowervalues corresponding to echolucency and higher valuescorresponding to echodensity. The following protocol forimage standardization was used26: (1) Color informationwas discarded. (2) GSM measurement of the plaque lumenand echodense area of plaque was performed. (3) GSMmeasurement of the entire plaque using a temporary loga-rithmic scaling was done. (4) Image standardization usingalgebraic scaling of the whole image using the “curve”facility of the software was performed. The gray scale wasadjusted so that the image of the plaque lumen would be inthe 0-5 range and the echodense region would be in the185-195 range. Thus, the gray values of all pixels wouldchange according to the new linear scale defined by thereference values for plaque lumen and echodense region.(5) Measurement of GSM of the noncalcific regions of the

Fig 2. Intact endarterectomized plaques were embeddedwere then subjected to B-mode ultrasonography. Serial croaverage of 12 sections/plaque (range 8-20). Images were t

plaque using the standardized scale (Fig 3).

Histopathologic/histomorphometric analysis. Plaqueswere fixed and sectioned transversely into 3-mm blocks(4-7 blocks/plaque). The blocks were embedded in paraf-fin and sectioned at 5-�m intervals. Sections from eachblock were stained with hematoxylin and eosin (H & E),and the presence of atheroma, necrotic lipid core, calcifica-tion, fibrous cap integrity, and ulceration was determinedin each section. Computer-assisted morphometric tech-niques were used to quantify the percentage of area ofcalcification, percentage of area of necrotic core, and short-est distance of the necrotic core to the lumen for eachsection (Fig 4).

Statistical analysis. For continuous demographicvariables, results were expressed as mean � SD, and t testwere used for group comparison. For categorical demo-graphic variables, �2 tests were used for group comparison.When multiple measurements were made of a particular

atin with a 5-mm echogenic scale (A). The ex vivo plaquestional views were obtained at 2- to 3-mm intervals, with antored digitally on a magneto-optical disk (B).

in gelss-sec

parameter in a given plaque, the mean value of those

JOURNAL OF VASCULAR SURGERYSeptember 2005438 Grogan et al

measurements was used in the statistical analysis. Differ-ences were considered significant at the P � .05 level.

To classify asymptomatic and symptomatic patients onthe basis of the percentage of area necrotic core and GSM,the sensitivity and specificity were calculated for a range of

Fig 3. Ex vivo gray-scale median analysis of cross-sectional viewsof endarterectomized plaques. A, Predominantly echolucentplaque (gray-scale median [GSM] � 28) with histogram demon-strating gray-scale frequency distribution of pixels in the selectedarea (plaque) using a standardized linear scale. B, Predominantlyechogenic plaque (GSM � 98) with corresponding histogram.U/S, ultrasonography.

Fig 4. Histopathologic section of an asymptomatic plaque withthe necrotic core (NC) delineated and quantified with histomor-phometry (hematoxylin and eosin, �40). L, lumen; FC, fibrouscap.

cutoff points. Receiver operating characteristic (ROC)

curves were constructed to determine the optimal thresh-old values. Linear regression was used to correlate GSMwith the percentage of area of necrotic core.

RESULTS

Demographic characteristics. The patients’ meanage (� SD) and mean carotid percentage of stenosis were66 � 8 years and 76% � 16% in the symptomatic and 72 �7 years and 82% � 11% in the asymptomatic (P � .05).Forty percent patients in the symptomatic group (10/25)and 43% of patients in the asymptomatic (10/23) werefemales (�2 � 0.06, P � .05). Clinical risk factors foratherosclerosis, including coronary artery disease, diabetesmellitus, hypertension, and smoking were similar for allpatients. There was a higher incidence of hypercholesterol-emia in symptomatic patients (38%, 10/26) than in asymp-tomatic patients (25%, 6/24), P � .05.

Preoperative CDU estimation of ICA stenosis usingCDU spectral waveform analysis and velocity profileshowed that 54% (14/26) of symptomatic patients had50%-79% stenosis and 46% (12/26) had 80%-99% stenosis.In asymptomatic patients 42% (10/24) had 50%-79% ste-nosis and 58% (14/24) had 80%-99% stenosis. Differenceswere not significant. The mean percentage of stenosis mea-sured by B-mode US was 76% for symptomatic patients and82% for asymptomatic patients (Table I).

58%, 35%, and 7% of symptomatic and of 18%, 41%,and 41% of asymptomatic plaques were found to be echolu-cent, echogenic, and calcific, respectively. These findingswere statistically significant (P � .05) (Table II). Significantacoustic shadowing was demonstrated in ten asymptomaticand in only two symptomatic plaques, whereas 38 plaqueswere devoid of acoustic shadowing.

Ex vivo B-mode US and computer-assisted analysisof plaque echodensity. Ex vivo B-mode US images (mean� 12) for each plaque were captured and digitized forquantitative computer-assisted echodensity measurements.Plaques from symptomatic patients (n � 26) were moreecholucent, with a mean GSM of 41 � 19, when comparedto plaques from asymptomatic patients (n � 24), whoseGSM mean was 60 � 13 (P � .03). The mean GSM ofplaques that were identified by preoperative CDU asecholucent was 33.8 � 11.5, and the mean GSM of

Table I. Degree of carotid stenosis using preoperativeB-mode ultrasonography and Doppler spectral analysiswith velocity profile

Doppler spectral analysisvelocity profile

B-modeultrasonography50%-79% 80%-99%

Symptomatic(n � 26) 54% (14/26) 46% (12/26) 76%

Asymptomatic(n � 24) 42% (10/24) 58% (14/24) 82%

Strandness criteria were used to classify stenosis based on velocity criteria.

echodense plaques was 53.6 � 10 (P � .001).

sensitivity and specificity.

JOURNAL OF VASCULAR SURGERYVolume 42, Number 3 Grogan et al 439

Histopathologic/histomorphometric analysis. Thepresence of necrotic core (foam cells, cholesterol clefts,myxomatous debris) was more marked in symptomaticplaques than asymptomatic plaques. The mean percentageof plaque necrotic lipid core was 51% � 18% for symptom-atic plaques and 23% � 20% for asymptomatic plaques. Thenecrotic core was twice as close to the lumen in symptom-atic plaques compared to asymptomatic plaques (0.29 �0.2 vs 0.53 � 0.4, P � .05). Figure 5 summarizes his-topathologic characteristics of symptomatic and asymp-tomatic plaques.

The mean percentage of calcification area was 66% and27%, respectively, in plaques with and without acousticshadowing (P � .05).

Correlation between symptomatology, percentageof necrotic core area, and GSM with ROC analysis. Astrong inverse correlation (R � 0.9, P � .001) was foundbetween the GSM scores and histomorphometric measuresof necrotic core burden. The lower the GSM was, ie,greater plaque echolucency, the larger the percentage ofnecrotic core area/plaque (Fig 6). ROC analysis was per-formed that demonstrated that a GSM 47 could predictsymptomatic plaques with a sensitivity of 73% and a speci-

Fig 5. Bar graph of the histomorphometric quantitative charac-terization of symptomatic and asymptomatic carotid plaques.

Fig 6. Correlation between gray-scale median (GSM) analysisand percentage of area of necrotic core was –0.9 (P � .001),indicative of the strong association between echolucency (lowGSM score) and percentage of necrotic core area.

ficity of 83% (Fig 7). ROC analysis of the percent area

Fig 7. Receiver operating characteristic curve for differentiatingbetween asymptomatic and symptomatic groups based on gray-scale median (GSM). A cutoff point of 47% yielded the highest

Fig 8. Receiver operating characteristic curve for differentiatingbetween asymptomatic and symptomatic groups based on thepercentage of necrotic core area. A cutoff point of 29% yielded the

highest sensitivity and specificity.

JOURNAL OF VASCULAR SURGERYSeptember 2005440 Grogan et al

necrotic core (histology) and symptomatic plaques wasperformed which showed that a percent area necrotic coreof 29% could predict symptomatic plaques with a sensitivityof 100% and a specificity of 67% (Fig 8).

DISCUSSION

In this study, preoperative subjective evaluation of ca-rotid plaque echolucency revealed that 58% of symptomaticplaques were predominantly echolucent compared to only18% of asymptomatic plaques. These findings are supportedby several studies relating plaque echolucency to neuro-logic events.27-32 Langsfield et al29 and Belcaro et al30

demonstrated that echolucent or heterogeneous plaquesare at a significantly increased risk of becoming symptom-atic when compared to dense, echo-rich plaques. A study byO’Holleran et al31 followed 293 patients over a 4-yearperiod, and all 42 patients with a soft lesion (lipid andintraplaque hemorrhage) with �75% stenosis becamesymptomatic. Confirming these results, Bock et al19

showed that echolucent plaques were associated with a5.7% annual incidence of transient ischemic attack andstroke compared to only 2.4% for echogenic plaques. In across-sectional study, Geroulakos et al32 found echolucentplaques to be associated with a 37% incidence of braininfarcts on CT scans compared to a combined incidence of18% in plaques that were predominantly echolucent (typeII), predominantly echogenic (type III), and echogenic(type IV). In contrast to these findings, Holdsworth et al33

in their study of 4258 patients only found amaurosis fugaxto be associated with echolucent and heterogeneousplaques, whereas the degree of stenosis overall seemedmore predictive of events.

To confirm the preoperative subjective B-mode evalu-ation of carotid plaque echolucency, we quantitated theecholucency of the ex vivo plaques following CEA usingGSM analysis of multiple cross-sectional views. When thepreoperative subjective assessments of echogenicity werecompared to the ex vivo GSM values, subjectively echolu-cent plaques had significantly lower GSM scores (meanGSM � 34) compared to subjectively echodense plaques(mean GSM � 53). This observation agrees with a previousstudy by Mayor et al,34 who found a correlation betweenvisual evaluation of plaque echogenicity and computerized

Table II. B-mode ultrasonographic characterization ofcarotid atherosclerotic plaques, including preoperativesubjective evaluation and ex vivo gray-scale mediananalysis

Group ISymptomatic

(n � 26)

Group IIAsymptomatic

(n � 24) P

Echolucent plaque 58% (15/26) 18% (4/24) �.03Calcified plaque 7% (2/26) 41% (10/24) �.03GSM (Ex-vivo B-mode

ultrasound) 41 � 19 60 � 13 �.03

measurements of GSM. Our results indicate that symptom-

atic plaques were more echolucent than asymptomaticplaques having a significantly lower GSM score (GSMscores 41 � 19 vs 60 � 13, respectively). The thresholdGSM value that would best separate symptomatic andasymptomatic plaques was a GSM of 47 (Fig 7). This trendhas been observed in other studies.23,35-37 Elatrozy et al26

studied 91 plaques from 77 patients and demonstrated thatthe GSM score of symptomatic plaques (mean � 21.5 �16) was significantly lower than that of asymptomaticplaques (mean � 37.6 � 26). El Barghouty et al38 notedthat plaques with a GSM �32 (echogenic) were associatedwith an incidence of 11% CT infarction compared to a 55%incidence of CT infarction for echolucent (GSM �32)plaques. More recently, in a prospective study, Gronhodtet al18 found that plaque echolucency could predict futurestroke in symptomatic patients with 50%-99% stenosis butnot in asymptomatic patients with comparable degrees ofstenosis. In that study, echolucency was defined by themean of the standardized GSM values, with GSM �74equaling an echolucent plaque and a GSM �74 equaling anecho-rich plaque. These studies involved the in vivo deter-mination of the GSM score in contrast to this report inwhich the GSM was calculated ex vivo. Plaque ex vivofixation may have resulted in an overall increase in plaqueechogenicity and an upward drift of the GSM score range.

Most previous studies have relied on a single longitu-dinal image for data acquisition and GSM score. In thisstudy, however, the GSM score was determined from trans-verse as well as longitudinal images offering a more detailedrepresentation of echogenicity of the entire plaque.Wijeyaratne et al39 found that multiple cross-sectionalviews in vivo (five to eight per plaque) increased GSManalysis sensitivity in detecting differences between symp-tomatic and asymptomatic patient populations. They alsoemphasized that a single longitudinal view was limited inaccurately representing plaque echolucency.

Finally, histomorphometric analysis of the scannedplaques was undertaken to develop a correlative analysisbetween the GSM scores and the percentage of necroticcore area (Fig 6). The area of necrotic core in symptomaticplaques was twofold greater than in asymptomatic plaques.Additionally, the distance of the necrotic core to the lumenwas doubled in asymptomatic patients compared to symp-tomatic patients. We found a strikingly significant inverserelationship between the GSM scores and the percentage ofnecrotic core area (R � –0.9, P � .001). This is inagreement with previous studies which have shown thatpixel evaluation can be used to differentiate between tissuetypes.40,41

The presence of carotid plaque calcification as evi-denced by significant acoustic shadowing was more com-monly found in asymptomatic plaques on preoperative USimaging (41% vs 7%, P � .01) than symptomatic plaques.This concurs with our previous report using ex vivo spiralCT to assess plaque calcification area in symptomatic andasymptomatic plaques.10 In the previous study, asymptom-atic plaques exhibited more extensive calcification in con-

trast to symptomatic plaques.

JOURNAL OF VASCULAR SURGERYVolume 42, Number 3 Grogan et al 441

From the data presented herein and the above discus-sion, it is evident that the B-mode US findings of ICAplaque echolucency and minimal calcification are associatedwith plaque necrosis, potential structural instability, andpotentially a higher risk of ischemic thromboembolicevents. Future refinements in the in vivo three-dimensionalquantitative assessment of plaque echolucency and burdenby B-mode US will further advance our understanding ofthe natural history of significant asymptomatic ICA steno-ses and assist in better selection of patients for carotidintervention.

REFERENCES

1. North American Symptomatic Carotid Endarterectomy Trial Collabo-rators: Beneficial effect of carotid endarterectomy in symptomatic pa-tients with high-grade carotid stenosis. N Engl J Med 1991;325:445.

2. European Carotid Surgery Trialists’ Collaborative Group: MRC Euro-pean Surgery Trial: interim results for symptomatic patients with severe(70-99%) carotid stenosis. Lancet 1991;337:1235.

3. The Executive Committee for the Asymptomatic Carotid Atherosclero-sis (ACAS) Study: endarterectomy for asymptomatic carotid arterystenosis. JAMA 1995;273:1421.

4. Haliday A, Mansfield A, Marro J, Peto C, Peto R, Potter J, Thomas D.MRC Asymptomatic Carotid Surgery Trial (ACST) CollaborativeGroup. Prevention of disabling and fatal strokes by successful carotidendarterectomy in patients without recent neurological symptoms: ran-domized controlled trial. Lancet 2004;363:1491-502.

5. Shanik GD, Moore JD, Leahy A, Grouden MC, Colgan MP. Asymp-tomatic carotid stenosis: a benign lesion? Eur J Vasc Surg 1992;6:10-5.

6. Chambers BR, Norris JW. Outcome in patients with asymptomatic neckbruits. N Engl J Med 1998;315:860-5.

7. Johnson JH, Kennelly MM, Decesare D, Morgan S, Sparrow A. Naturalhistory of asymptomatic carotid plaque. Arch Surg 1985;120:1010-2.

8. Feeley TM, Leen EJ, Colgan MP, et al. Histological characteristics ofcarotid artery plaque. J Vasc Surg 1191;12:719-24.

9. Bassiouny HS, Sakaguchi Y, Mikucki SA, et al. Juxtalumenal location ofplaque necrosis and neoformation in symptomatic carotid stenosis. JVasc Surg 1997;26:585-94.

10. Shaalan WE, Cheng H, Gewertz B, et al. Degree of carotid plaquecalcification in relation to symptomatic outcome and plaque inflamma-tion. J Vasc Surg 2004;40:262-9.

11. Gronholdt M-LM. Ultrasound and lipoproteins as predictors of lipid-rich, rupture-prone plaques in the carotid artery. Arterioscler ThrombVasc Biol 1999;19:2-13.

12. O’Donnell TF Jr, Erodes L, Mackey WC, et al. Correlation of B-modeultrasound imaging and arteriography with pathologic findings at ca-rotid endarterectomy. Arch Surg 1985;120:443-9.

13. Gray-Weale AC, Graham GC, Burnett JR, et al. Carotid artery ather-oma: comparison of preoperative B-mode ultrasound appearance withcarotid endarterectomy specimen pathology. J Cardiovasc Surg 1998;29:676-81.

14. Wolverson MK, Bashiti HM, Peterson GJ. Ultrasonic tissue character-ization of atheromatous plaques using a high resolution real timescanner. Ultrasound Med Biol 1983;9:599-609.

15. Goes E, Janssens W, Maillet B, et al. Tissue characterization of athero-matous plaques: correlation between ultrasound image and histologicalfindings. J Clin Ultrasound 1990;18:611-7.

16. Steffen CM, Gray-Weale AC, Byrne KE, et al. Carotid artery atheroma:ultrasound appearance in symptomatic and asymptomatic vessels. AustN Z J Surg 1989;59:529-34.

17. Geroulakos G, Ramaswami G, Nicolaides A, et al. Characterization ofsymptomatic and asymptomatic carotid plaques using high-resolutionreal-time ultrasonography. Br J Surg 1993;80:1274-7.

18. Gronholdt MLM, Nordestgaard BG, Schroeder TV, Vostrup S, SillesenH. Ultrasonic echolucent carotid plaques predict future strokes. Circu-

lation 2001;104:68-73.

19. Bock RW, Gray-Weale AC, Mock PA, et al. The natural history ofasymptomatic carotid artery disease. J Vasc Surg 1993;17:160-9.

20. El Barghouty N, Geroulakos G, Nicolaides A, et al. Computer-assistedcarotid plaque characterization. Eur J Vasc Surg 1995;9:389-93.

21. Sterpetti AV, Scultz RD, Feldhaus RJ, et al. Ultrasonographic featuresof carotid plaque and the risk of subsequent neurologic deficits. Surgery1988;104:652-60.

22. Mercuri M, Bond MG. B-mode ultrasound characterization of athero-sclerosis. J Cardiovasc Tech 1992;10:277-91.

23. Biasi GM, Sampaolo A, Mingazzini P, et al. Computer analysis ofultrasonic plaque echolucency in identifying high risk carotid bifurca-tion lesions. Eur J Vasc Endovasc Surg 1999;17:476-9.

24. Strandness DE Jr. Extracranial arterial disease. In: Strandness DE Jr,editor. Duplex scanning in vascular disorders, 2nd ed. New York, RavenPress, 1993, pp. 113-58.

25. Zweibel WJ. Doppler evaluation of carotid stenosis. In: Zweibel WJ,editor. Introduction to vascular ultrasonography, 3rd ed. Philadelphia,WB Saunders, 1992, pp. 123-32.

26. Elatrozy T, Nicolaides A, Tegos TH, et al. The effect of B-modeultrasonic image standardization on the echodensity of symptomaticand asymptomatic carotid bifurcation plaques. Int Angiol 1998;17:179-86.

27. Mathieson EB, Bonaa KH, Joakimsen O. Echolucent plaques areassociated with high risk of ischemic cerebrovascular events in carotidstenosis. The Tromso study. Circulation 2001;2171-5.

28. AbuRahma AF, Wulu JT, Crotty B. Carotid plaque ultrasonic hetero-geneity and severity of stenosis. Stroke 2002;33:1772-5.

29. Langsfield M, Gray-Weale AC, Lusby RJ. The role of plaque morphol-ogy and diameter reduction in the development of new symptoms inasymptomatic carotid arteries. J Vasc Surg 1989;9:548-57.

30. Belcaro G, Laurora G, Cesarone MR, et al. Ultrasonic classification ofcarotid plaques causing less than 60% stenosis according to ultrasoundmorphology and events. J Cardiovasc Surg 1993;34:287-94.

31. O’Holleran LW, Kennelly MM, McClurken M, et al. Natural history ofasymptomatic carotid plaque. Am J Surg 1987;154:659-61.

32. Geroulakos G, Domjan J, Nicolaides A, et al. Ultrasonic carotid arteryplaque structure and the risk of cerebral infarction on computed tomog-raphy. J Vasc Surg 1994;20:263-6.

33. Holdsworth RJ, Bryce JS, Harrison DK, et al. Gender, age and carotidplaque morphology. Minerva Cardioangiol 1997;45:363-7.

34. Mayor I, Momijan S, Lalive P, Sztajzel R. Carotid plaque: comparisonbetween visual and grey-scale median analysis. Ultrasound Med Biol2003;29:961-6.

35. Tegos TJ, Stavropoulos P, Sabetai MM, Khodabakhsh P, Sassano A,Nicolaides AN. Determinants of carotid instability: echoicity versusheterogeneity. Eur J Vasc Endovasc Surg 2001;22:22-30.

36. Sabetai MM, Tegos TJ, Nicolaides AN, El-Atrozy TS, Dhanjil S, GriffinM, et al. Hemispheric symptoms and carotid plaque echomorphology.J Vasc Surg 2000;31:39-49.

37. Pedro LM, Pedro MM, Goncalves I, et al. Computer-assisted carotidplaque analysis: Characteristics of plaques associated with cerebrovas-cular symptoms and cerebral infarction. Eur J Vasc Endovasc Surg2000;19:118-23.

38. El Barghouty N, Nicolaides A, Bahal V, Geroulakos G, Androulakis A.The identification of the high risk carotid plaque. Eur J Vasc EndovascSurg 1996;11:470-8.

39. Wijeyaratne SM, Jarvis S, Stead LA, Kibria SG, Evans JA, Gough MJ. Anew method for characterizing carotid plaque: multiple cross-sectionalview echomorphology. J Vasc Surg 2003;37:778-84.

40. Lal BK, Hobson RW, Pappas PJ, Kubicka R, Hameed M, ChakhturaEY, et al. Pixel distribution analysis of B-mode ultrasound scan imagespredicts histologic features of atherosclerotic plaques. J Vasc Surg2002;35:1210-7.

41. Aly S, Bishop C. An objective characterization of atherosclerotic lesion;an alternative method to identify unstable plaque. Stroke 2000;31:1921-4.

Submitted Oct 1, 2004; accepted May 8, 2005.