High-Grade Stenosis of the Internal CarotidArtery Assessed by Color and PowerDoppler Imaging

Peter Schmidt, MD, Ulrich Sliwka, MD, Stefan-Georg Simon, MD, Johannes Noth, PhD

Department of Neurology, Technical University of Aachen, Pauwelsstrasse 30, D-52074 Aachen, Germany

Received 8 July 1996; accepted 6 August 1997

ABSTRACT: Purpose. Visualization of carotid arterystenosis is important to improve our understanding ofthe etiology of ischemic stroke. Two sonographicmethods are available for visualizing carotid stenosis:power Doppler imaging (PDI) and color Doppler im-aging (CDI). The purpose of this study was to comparethe advantages and limitations of these methods inthe evaluation of internal carotid artery (ICA) stenosis.

Methods. Thirty-two patients with at least 70% ex-tracranial ICA stenosis were included in this prospec-tive study. All subjects were examined with CDI andPDI. Four criteria were used to evaluate diagnostic ac-curacy and overall efficacy: identifying the target ves-sel, determining the causal pathology, evaluatingblood flow, and the examiner’s confidence in the di-agnosis.

Results. All patients had high-grade stenosis (17patients had 70% stenosis, 11 had 80%, 2 had 90%,and 2 had pseudo-occlusions of the ICA). PDI was sig-nificantly (p < 0.05) superior to CDI in identifying thetarget vessel and highly significantly (p < 0.01) supe-rior to CDI in evaluating blood flow. There were nosignificant differences between CDI and PDI in deter-mining the causal pathology or in the examiner’s con-fidence in the diagnosis.

Conclusions. This study demonstrates that PDI is agood additional method to CDI and should be usedwith it to evaluate ICA stenosis. © 1998 John Wiley &Sons, Inc. J Clin Ultrasound 26:85–89, 1998.

Keywords: carotid stenosis; color Doppler ultrasonog-raphy; power Doppler ultrasonography

Characteristics of carotid artery stenosis suchas plaque morphology, surface structure, and

vessel lumen diameter are important pathogenicmechanisms for cerebral ischemia.1 Understand-ing the role played by carotid artery stenosis inthe etiology of cerebral ischemia requires clearvisualization and quantification of the stenosis.In most neurovascular laboratories, color-codedimaging with duplex scanning is an essential partof routine sonographic examinations of the extra-cranial arteries.2–6 Such examination is noninva-sive and cost-effective and provides reliable re-sults. However, duplex scanning is highlyexaminer-dependent, and areas of stenosis maybe completely obscured by dense, calcifiedplaques.7,8

At present, 2 Doppler methods are used to im-age blood flow, color Doppler imaging (CDI) andpower Doppler imaging (PDI). The 2 methodshave different sensitivities and specificities for in-vestigating blood flow velocity, vessel wall struc-ture, and plaque morphology. In CDI, the signal isbased on frequency-shift analysis. In PDI, the in-travascular signal depends on the density of redblood cells within the sample volume.9–15 Select-ing the most appropriate of these techniques willincrease confidence in the diagnosis and facilitatetherapeutic management (eg, anticoagulation orendarterectomy). Use of CDI is already well es-tablished in routine clinical practice,6,16 but thevalue of PDI for routine use is still being assessed.For PDI to be accepted into routine practice, itmust be shown to improve diagnostic accuracyand not to extend the time required to performimaging. The technique should also have a rea-

Correspondence to: P. Schmidt

© 1998 John Wiley & Sons, Inc. CCC 0091-2751/98/020085-05

85VOL. 26, NO. 2, FEBRUARY 1998

sonable learning curve for examiners. The pur-pose of this study was to compare the diagnosticaccuracy and overall efficacy of PDI and CDI inroutine examinations of carotid artery stenosisperformed in a neurovascular laboratory.

PATIENTS AND METHODS

Patients

We examined 32 patients (19 men and 13 women)whose ages ranged from 44 to 72 years (mean ±standard deviation, 58.3 ± 7.7 years) in the neu-rovascular laboratory of the Department of Neu-rology at the Technical University of Aachen. Allpatients included in the study had high-grade ste-nosis of the extracranial internal carotid artery(ICA) as determined by criteria for continuouswave (CW) Doppler examination,17,18 which wasdone for initial screening. Thirteen patients hadsymptomatic and 19 had asymptomatic ICA ste-nosis. Eleven of the symptomatic patients had ex-perienced a cerebral transient ischemic attack,and 2 had experienced an acute ischemic stroke.The asymptomatic patients with ICA stenosis hadbeen referred by their primary care physicians forroutine sonographic examination to evaluate thedegree of stenosis. After the screening examina-tion, patients who agreed to participate were in-cluded prospectively in this study. Written in-formed consent was obta ined from al lparticipants. All patients were then re-examinedwith PDI and CDI in a single session.

CW Doppler

The CW Doppler examination was initially per-formed to determine the degree of ICA stenosis.All patients were examined with a 4-MHz probe(Multi-DopX, DWL Inc., Sipplingen, Germany).The stenoses were classified by analysis of thefrequency spectrum as described elsewhere.19

Middle-grade stenosis was defined as a 50–69%reduction of vessel lumen diameter, and high-grade stenosis was defined as a reduction of 70%or greater. CW Doppler criteria of diameter re-duction were as follows: 40–69% reduction, mod-erate local increase of peak blood flow velocitywith poststenotic turbulence; 70–89% reduction,marked increase of peak and mean blood flow ve-locity with poststenotic systolic deceleration; and90–99% reduction, more severe increase of peakand mean blood flow velocities with reduction ofipsilateral pre- and poststenotic blood flow veloc-ities.20

PDI and CDI

All patients were examined using an Acuson sys-tem (Acuson Inc., Mountain View, CA) with a lin-ear-array pulsed-wave transducer operating at 5or 7 MHz that enabled both PDI and CDI. Longi-tudinal and transverse views of the ICA in thearea of stenosis were obtained. PDI was per-formed indifferently before or after CDI. After thestenosis was visualized with CDI or PDI, imagesof the stenosis were stored digitally on magneto-optical disks for subsequent evaluation, and theexamination was repeated with the other method.

Examination Evaluation

To evaluate diagnostic accuracy and overall effi-cacy, the examiner was required to assess theusefulness of CDI and PDI for 4 criteria: (1) theeffort required to identify the ICA, including thetime required to do so and the diagnostic accuracyof the findings; (2) the ability to determine thecausal pathology with clear imaging of vessel wallstructures and a distinct flow signal getting ahigh rating and incomplete visualization of vesselwall structures and flow getting a low rating; (3)the effort required to evaluate blood flow in theregion of interest with clear visualization and astrong signal rating high and difficulty in imagingflow and poor signal quality rating low; and (4)the examiner’s subjective impression regardingthe reliability of the diagnosis.

Criteria 1 and 4 were considered indicators ofefficacy, whereas criteria 2 and 3 were consideredindicators of the sensitivity and specificity of thediagnostic method. To semiquantify the subjec-tive impression, the examiner was instructed tograde each criterion by awarding 0–3 points asfollows: 3, excellent; 2, satisfactory; 1, less thansatisfactory; and 0, poor.

Data Analysis

The scores were analyzed with the nonparametricWilcoxon test for paired samples. A p value of 0.05was considered significant, and a p value of 0.01was considered highly significant. Linear rela-tionships were demonstrated with the use ofPearson’s-Bravais product-moment correlation(r). The data are expressed as means ± standarddeviations.

RESULTS

On the basis of CW Doppler criteria, 17 patientshad ICA stenosis of 70%, 11 patients had 80% ICA

SCHMIDT ET AL

86 JOURNAL OF CLINICAL ULTRASOUND

stenosis, 2 patients had 90% ICA stenosis, and 2had so-called pseudo-occlusion of the ICA.

The scores for PDI and CDI for each evaluatedcriterion are shown in Table 1. Statistical analy-sis showed that PDI was significantly (p < 0.05)superior to CDI in identifying the target vessel(criterion 1) and highly significantly (p < 0.01)superior in imaging blood flow in the region ofinterest (criterion 3). There was no significant dif-ference between the 2 methods in determining thecausal pathology or the examiner’s confidence inthe diagnosis (criteria 2 and 4).

Good correlation was found between the 2methods in determining the causal pathology (r4 0.98) and the examiner’s confidence in the di-agnosis (r 4 0.99). A lower correlation was foundin the evaluation of blood flow (r 4 0.59) and theidentification of the target vessel (r 4 0.82).

DISCUSSION

PDI is nearly independent of the insonation anglecompared with CDI. In PDI, the signal is based onthe area under the power spectral density curve,which increases with the number of reflectors(red blood cells). The area under the curve, butnot its shape, is independent of the insonationangle as long as the frequency shift depicted ex-ceeds the soft-tissue-motion (clutter) canceller of

the imaging unit. Thus, the color appears to beessentially the same from all angles. Alteration ofthe insonation angle in CDI leads to distinct sig-nal changes that can easily lead to a reduction orloss of signal in difficult examination situations(eg, a winding vessel course or inadequate probeplacement) and demands skilled examiners. Com-pared with the pronounced angle dependence ofCDI, PDI with its minor angle dependence leadsto a stronger signal and facilitates difficult exami-nations. Table 2 summarizes the characteristicsof each method.

In the present study, PDI and CDI were com-pared using evaluation criteria for ICA stenosis.PDI was found to be superior to CDI in identify-ing the target vessel, and all examiners found itmuch easier to detect the target vessels using PDI(Figure 1). PDI was less time-consuming and en-abled a higher amplification level without thepresence of overwhelming noise artifacts, whichare easily generated with CDI (eg, in low-flowsituations such as dilatative arteriopathy orpseudo-occlusion or recanalization of the ICA).The reduced angle dependence of PDI also helpedin identifying the target vessel in cases of devia-tion of the vessel course (eg, vessel elongation incases of long-standing hypertension). The abilityof CDI to determine the direction of blood flowand thus to enable easy discrimination betweenarteries and veins was found to be less importantthan presumed.

In determining causal pathology, CDI and PDIwere of equal value as diagnostic tools. Examin-ers preferred CDI for quick identification of ste-nosis of the carotid artery because the color alter-ation due to increased blood flow velocities wasseen at once. In PDI, the region of abnormalitywas more difficult to identify, with only narrowedvessel walls or suddenly thinned out color provid-

TABLE 1Assessment of Examination Criteria by PDI and CDI

Criterion/Score

Frequency ofScore Awarded

PDI(n = 32)

CDI(n = 32)

Identification of target vesselExcellent (3) 25 14Satisfactory (2) 5 11Less than satisfactory (1) 2 6Poor (0) 0 1Mean score ± SD 2.72 ± 0.58 2.19 ± 0.86

Determination of causal pathologyExcellent (3) 12 14Satisfactory (2) 14 13Less than satisfactory (1) 4 3Poor (0) 2 2Mean score ± SD 2.12 ± 0.87 2.22 ± 0.87

Evaluation of blood flowExcellent (3) 21 9Satisfactory (2) 9 12Less than satisfactory (1) 2 6Poor (0) 0 5Mean score ± SD 2.59 ± 0.61 1.78 ± 1.04

Examiner’s confidence in diagnosisExcellent (3) 14 12Satisfactory (2) 10 10Less than satisfactory (1) 5 6Poor (0) 3 4Mean score ± SD 2.09 ± 1.0 1.94 ± 1.05

Abbreviations: PDI, power Doppler imaging; CDI, color Doppler im-aging; SD, standard deviation.

TABLE 2Comparison of PDI and CDI

Characteristic PDI CDI

Color-codedparameter

Amplitude Frequency shift

Correspondingblood flowparameter

Density of bloodcell aggregation

Blood flow velocity

Angle dependence Relativelyindependent

Dependent

Directionalinformation

Absent Present

Noise Uniformly low Totally randomAliasing Absent Above Nyquist

limitSignal pulsatility Low High

Abbreviations: PDI, power Doppler imaging; CDI, color Doppler im-aging.

INTERNAL CAROTID ARTERY STENOSIS

87VOL. 26, NO. 2, FEBRUARY 1998

ing evidence. The hemodynamic changes causedby the stenosis itself was much more obvious withCDI. With PDI, it was often necessary to use anadditional pulsed-wave Doppler technique in thearea of stenosis to confirm hemodynamic changes.For judgments concerning plaque morphology,PDI was preferred because the entire image wassmoother and the vessel wall/lumen interface wasless disturbed by color dots representing the noisecomponent. We suggest, however, that for evalu-ation of plaque morphology, the safest approach is

to turn color Doppler off and to use the black-and-white B-mode alone.

In the evaluation of blood flow, PDI was clearlysuperior to CDI, especially in low-flow situations(eg, in cases of dilatative arteriopathy or afterendarterectomy and in cases of recanalization ofan older ICA occlusion, particularly pseudo-occlusions) (Figure 2). Pseudo-occlusion is anextreme type of stenosing lesion that is easilymisdiagnosed as complete occlusion during con-ventional intra-arterial angiography. The slow re-

FIGURE 1. Differences in the abilities of PDI and CDI to visualize the target vessel. (A) PDI scan (at 7 MHz) of carotid bifurcation with a stenosis ofthe internal carotid artery. (B) CDI scan (at 5 MHz) of the same area shows missing color at the point where the ultrasound beam is perpendicularto the direction of blood flow. ACC, common carotid artery; ACE, external carotid artery; ACI, internal carotid artery; R, right.

FIGURE 2. Differences in the abilities of PDI and CDI to image the blood flow at carotid bifurcation with a pseudo-occlusion of the internal carotidartery. (A) With PDI, the flow signal in the internal carotid artery is clearly visible. (B) With CDI, there is no flow signal. (The PDI scan was obtainedat 7 MHz and the CDI scan at 5 MHz). ACC, common carotid artery; ACE, external carotid artery; ACI, internal carotid artery; LI, left.

SCHMIDT ET AL

88 JOURNAL OF CLINICAL ULTRASOUND

sidual arterial flow is detectable only by usingspecially altered angiographic techniques.21,22

Only sonography enables detection of this condi-tion noninvasively. The closing vessel lumencauses a substantial reduction of the blood flow.For all these abnormalities, the sensitivity of PDIwas substantially higher than that of CDI. Insome cases, CDI was not capable of imaging theabnormalities at all.

With regard to the subjective criterion of diag-nostic confidence, neither method was superior.The visual impression of the imaged vessel was,subjectively, more complete with PDI. However,the ability of CDI to display the various blood flowvelocities and flow direction was appreciated bythe examiners. On the other hand, the superiorability of PDI to visualize pathologic low-flowsituations was praised as well. A significant dis-advantage of PDI is its high sensitivity to tissue-motion artifacts. Disturbing artifacts were easilygenerated, sometimes even by a patient’s breath-ing. On the other hand, CDI examinations werelimited by noise artifacts and aliasing.

PDI is a good adjunct method that providesvaluable information for certain clearly definedabnormalities. However, PDI should not replaceCDI. We believe that combined use of the 2 meth-ods will lead to a diagnostic accuracy that cannotbe achieved by either method alone.

ACKNOWLEDGMENTS

We thank Drs. Stuart Fellows and Garry Brook ofthe Department of Neurology at Rheinisch-West-falische Technische Hochschule Aachen, Aachen,Germany, for their helpful suggestions.

REFERENCES

1. Toole JF, Chambless LE, Heiss G, et al: Prevalenceof stroke and transient ischemic attacks in the Ath-erosclerosis Risk in Communities (ARIC) study.Ann Epidemiol 1993;3:500.

2. Grant EG: Duplex ultrasonography: its expandingrole in noninvasive vascular diagnosis. Radiol ClinNorth Am 1985;23:563.

3. Johnston KW, Baker WH, Burnham SJ, et al:Quantitative analysis of continuous-wave Dopplerspectral broadening for the diagnosis of carotid dis-ease: results of a multicenter study. J Vasc Surg1986;4:493.

4. Kessler C, von Maravic C, von Maravic M, et al:Colour Doppler flow imaging of the carotid arter-ies. Neuroradiology 1991;33:114.

5. Polak JF, Dobkin GR, O’Leary DH, et al: Internalcarotid artery stenosis: accuracy and reproducibil-ity of color-Doppler-assisted duplex imaging. Radi-ology 1989;173:793.

6. Steinke W, Hennerici M, Rautenberg W, et al:Symptomatic and asymptomatic high-grade ca-rotid stenosis in Doppler color-flow imaging. Neu-rology 1992;42:131.

7. Jacobs NM, Grant EG, Schellinger D, et al: Duplexcarotid sonography: criteria for stenosis, accuracy,and pitfalls. Radiology 1985;154:385.

8. Mitchell DG: Color Doppler imaging: principles,limitations, and artifacts. Radiology 1990;7:1.

9. Bude RO, Rubin JM, Adler RS: Power versus con-ventional color Doppler sonography: comparison inthe depiction of normal intrarenal vasculature. Ra-diology 1994;192:777.

10. Cloutier G, Shung KK: Study of red cell aggrega-tion in pulsatile flow from ultrasonic Dopplerpower measurements. Biorheology 1993;30:443.

11. Cloutier G, Shung KK: Cyclic variation of the powerof ultrasonic Doppler signals backscattered by poly-styrene microspheres and porcine erythrocyte sus-pensions. IEEE Trans Biomed Eng 1993;40:953.

12. Jain SP, Fan PH, Philpot EF, et al: Influence ofvarious instrument settings on the flow informa-tion derived from the power mode. Ultrasound MedBiol 1991;17:49.

13. Minich LL, Snider AR, Meliones JN, et al: In vitroevaluation of volumetric flow from Doppler power-weighted and amplitude-weighted mean velocities.J Am Soc Echocardiogr 1993;6:227.

14. Rubin JM, Adler RS, Fowlkes JB, et al: Fractionalmoving blood volume: estimation with power Dopp-ler US. Radiology 1995;197:183.

15. Rubin JM, Bude RO, Carson PL, et al: Power Dopp-ler US: a potentially useful alternative to meanfrequency-based color Doppler US. Radiology 1994;190:853.

16. Riley WA, Barnes RW, Applegate WB, et al: Repro-ducibility of noninvasive ultrasonic measurementof carotid atherosclerosis. The Asymptomatic Ca-rotid Artery Plaque Study. Stroke 1992;23:1062.

17. Hennerici M, Aulich A, Sandmann W, et al: Inci-dence of asymptomatic extracranial arterial dis-ease. Stroke 1981;12:750.

18. Steinke W, Rautenberg W, Hennerici M: Cerebro-vascular ultrasonography. In Lanzer P, Rosch J,eds: Vascular Diagnostics. Berlin, Springer-Verlag,1994, p. 97.

19. Widder B, von Reutern GM, Neuerburg Heusler D:Morphologic and Doppler sonographic criteria for de-termining the degree of stenosis of the internal ca-rotid artery [in German]. Ultraschall Med 1986;7:70.

20. Steinke W, Kloetzsch C, Hennerici M: Carotid arterydisease assessed by color Doppler flow imaging: cor-relation with standard Doppler sonography and an-giography. AJNR Am J Neuroradiol 1990;11:259.

21. O’Leary DH, Gibbons GW, Pinel DF: Limitations ofnoninvasive testing in assessing the ‘‘occluded’’ ca-rotid artery. AJNR Am J Neuroradiol 1983;4:759.

22. Ringelstein EB, Berg Dammer E, Zeumer H: Theso-called atheromatous pseudoocclusion of the in-ternal carotid artery. A diagnostic and therapeuti-cal challenge. Neuroradiology 1983;25:147.

INTERNAL CAROTID ARTERY STENOSIS

89VOL. 26, NO. 2, FEBRUARY 1998