Clinical Studies

ARTICLE IN PRESS

Doppler Vibrometry: Assessment of ArterialStenosis By Using Perivascular TissueVibrations without Lumen VisualizationSiddhartha Sikdar, PhD, Sandeep Vaidya, MD, Manjiri Dighe, MD, Orpheus Kolokythas, MD, Jae Hwan Kim, BS,

Kirk W. Beach, MD, PhD, and Yongmin Kim, PhD

PURPOSE: To correlate vibration frequency and duration at Doppler vibrometry with stenosis severity determined atcatheter angiography.

MATERIALS AND METHODS: Sixteen patients (eight women) scheduled to undergo abdominal or pelvic angiog-raphy were recruited after providing informed consent. An ultrasonography (US) scanner was customized to acquireraw echo data before conventional Doppler processing. Data were acquired from perivascular tissue regions proximalto stenoses, close to the most narrow lumen, and distal to stenoses in the renal, hepatic, common iliac, and superiormesenteric arteries. The data were processed to quantify vibration frequency and duration. The minimum lumendiameter and the pre- and poststenotic lumen diameters were quantified from angiograms. One patient with a hepaticartery stenosis did not yield measurable vibrometry data due to significant bowel gas.

RESULTS: Stenoses (diameter reduction, 43%–91%) were angiographically measured in the six renal arteries, fourhepatic arteries, three iliac arteries, and one superior mesenteric artery yielding vibrometry data. Three iliac arterieswere normal (<30% diameter reduction at angiography). For these 17 arteries, the vibration frequency was higher witha greater percentage of stenosis [Pearson r � .75; P < .001) and a smaller minimum lumen diameter (r � .72; P < .001).The vibration duration increased with a greater percentage of stenosis (r � .7; P < .001).

CONCLUSIONS: Preliminary results indicate that the vibration frequency and duration can be used to quantitativelyestimate stenosis severity. Doppler vibrometry is complementary to duplex US and does not require lumenvisualization.

J Vasc Interv Radiol 2009; xx:xxx

DUPLEX ultrasonography (US) is thediagnostic method of choice for theinitial assessment of abdominal andpelvic artery stenoses (1). The duplex

From the Department of Electrical and ComputerEngineering (S.S.), George Mason University, 4400University Dr, MS 1G5, Fairfax, VA 22030; and theDepartments of Radiology (S.V., M.D., O.K.), Bio-engineering (J.H.K., K.W.B., Y.K.), and Surgery(K.W.B.), University of Washington Medical Center,Seattle, Washington. Received October 29, 2008; fi-nal revision received May 13, 2009; accepted June 8,2009. Address correspondence to S.S.; E-mail:ssikdar@gmu.edu

From the 2008 SIR annual meeting.

None of the authors have identified a conflict ofinterest.

© SIR, 2009

DOI: 10.1016/j.jvir.2009.06.003

examination can be technically chal-lenging and is often time-consumingowing to the need to visualize theartery lumen and stenosis site. Al-though the negative predictive valueof the duplex examination is typi-cally high, positive findings oftenneed further confirmation with useof conventional angiography, mag-netic resonance angiography, orcomputed tomographic angiographybefore intervention (2,3).

Several clinical studies performedduring the past 3 decades have dem-onstrated the utility of acousticmethods for the diagnosis of arterialstenoses (4 –7). Vascular sounds, orbruits, are audio frequency tissue vi-brations caused by turbulent flow

downstream of arterial stenoses.

These vibrations propagate throughthe soft tissue surrounding the ste-nosis site. Quantitative analysis ofthe bruit, known as phonoangiogra-phy, has shown that the frequency ofthe bruit is related to the minimumstenosis diameter. We have devel-oped a method, called Doppler vi-brometry, to locate and quantifythese tissue vibrations (8). Bruits ap-pear in the spectral Doppler wave-form as double-sided clutter (Fig 1).Conventional Doppler instrumentsattempt to suppress these tissue signalsby employing wall filters. Doppler vi-brometry enables the quantification ofthese tissue vibration signals from tissuenear the stenosis site. Vibrometry is asignificant improvement over phonoan-

giography, which is limited only to

1

2 • Doppler Vibrometry Assessment of Arterial Stenosis Month 2009 JVIR

ARTICLE IN PRESS

bruits audible at the skin and thus lacksspecificity about the origin of the bruit.Vibrometry allows the measurement ofvibration amplitude within the tissueadjacent to the artery, allowing compu-tation of acoustic intensity as well asfrequency. Vibrometry does not requirevisualization of the artery lumen yet canprovide quantitative information aboutpoststenotic turbulence. Thus, vibrom-etry can potentially complement duplexUS in technically challenging cases. Thepurpose of this study was to correlatethe quantitative vibration characteristicsdetermined with Doppler vibrometry inabdominal and pelvic arteries with thestenosis severity determined with con-ventional angiography.

MATERIALS AND METHODS

Study Population

In this prospective feasibilitystudy, 16 patients (eight women)who were scheduled to undergoabdominal or pelvic angiography atthe University of Washington inter-ventional radiology laboratory wererecruited for examination before an-giography. Angiography was indi-cated in these patients to evaluatesuspected stenoses on the basis of aprevious conventional duplex US ex-

Figure 1. Doppler signature of a systolic breduction, arrow). (b) Routine duplex US ssignal (red arrow) corresponding to a bruitof the Doppler signals from the superior pperivascular tissue (c.4). The multigate ansignal. Note that due to a lower pulse repvelocity signals are aliased.

amination. The symptoms that

prompted the initial duplex exami-nation were varied and related todifferent vascular beds (eg, hyper-tension, liver dysfunction, chronicabdominal discomfort, and claudica-tion). Written informed consent wasobtained from all patients. The Uni-versity of Washington institutionalreview board approved all the studyprocedures.

US Data Acquisition

A commercial US instrument (Hi-Vision 5500; Hitachi Medical SystemsAmerica, Twinsburg, Ohio) was modi-fied with custom software to acquireraw US data before conventional Dopp-ler signal processing. A 3�5-MHz con-vex transducer was used to acquire datafrom a 15-mm range gate placed in theperivascular tissue proximal to stenoses,close to the most narrow lumen, anddistal to the suspected stenoses in renal,hepatic, common iliac, and superiormesenteric arteries. The acquired Dopp-ler data were synchronized with the Rwave of the electrocardiographic signal.Figure 2a.1 shows the approximateplacement of the 15-mm Doppler rangegate in the perivascular tissue surround-ing a stenosis site in an iliac artery. Thedouble-sided spectral signature of thebruit is clearly visible in the spectral

t in the common iliac artery. (a) Angiogramdemonstrates elevated peak systolic velocid forward diastolic flow, all indications ofascular tissue (c.1), superior vessel wall (cis demonstrates that perivascular tissue v

ion frequency used to highlight the tissue

Doppler waveform (Fig 2a.2).

Quantification of Tissue Vibrations

The raw US data were processed byusing custom signal processing meth-ods that have been described in detailelsewhere (8). These methods were de-signed to estimate and analyze the tis-sue motion surrounding the artery asopposed to the blood velocity inside thelumen. Briefly, the instantaneous tissuevelocity was estimated by using a con-ventional autocorrelation method with-out employing any wall filters (Fig2b.1). The velocity was integrated to ob-tain a tissue displacement signal. Theaudio frequency tissue vibration wasisolated by filtering the tissue displace-ment signal between the frequencies of100 and 1,000 Hz, which corresponds tothe expected frequency range of postste-notic bruits (Fig 2b.2). The duration andfrequency of the vibration signal werethen quantified. The spectral waveformof the vibration demonstrates the broad-band high-frequency content (Fig 2b.3).The frequency was estimated as the sec-ond moment of the vibration powerspectral density (9).

Quantification of AngiographicSeverity

Transcatheter digital subtraction an-giography was performed in each pa-

a severe iliac artery stenosis (82% diameterand a strong double-sided systolic clutter

nosis. (c) Simultaneous multigate analysisinside the vessel lumen (c.3), and inferior

ations are the source of the strong clutterrations in the multigate analysis, the flow

rui ofcan ties

an steeriv .2),alys ibretit vib

tient by using the Axiom Artis dTA

we e.

Sikdar et al • 3Volume xx Number x

ARTICLE IN PRESS

(Siemens Medical Systems, Malvern,Pennsylvania) angiography system. Anonselective aortogram was obtainedinitially followed by selective angio-grams of the vessel of interest. Imag-ing was done in multiple planes with aminimum of two orthogonal planes.The image that showed the lesion inthe best profile was selected for mea-surements. Measurements were car-ried out by using the internal calibra-tion standards preset in theequipment. The opacified luminal di-ameters were measured at the lesionas well as before and after the lesion.The measurements were done by oneoperator (S.V.), who was blinded tothe results of the Doppler and vibro-metry studies. In this study, the steno-sis severity was treated as a continu-ous variable (percentage diameterreduction) and the correlation of thevibration characteristics with stenosisseverity was estimated.

RESULTS

A total of 18 arteries were imaged inthe 16 patients. Of the 18 arteries, steno-ses (diameter reduction, 43%–91%) wereangiographically measured in six renal

Figure 2. Illustration of the Doppler vibrothe perivascular tissue surrounding an iliacThe tissue velocity waveform is derived frby bandpass filtering of the velocity wavefothe distribution of the vibration power bet

arteries, four hepatic arteries, three iliac

arteries, and one superior mesenteric ar-tery yielding vibrometry data. One pa-tient with a 79% diameter reduction he-patic artery stenosis did not yieldmeasurable vibrometry data due to sig-nificant bowel gas precluding adequate

try method in a stenosed common iliac artery stenosis. (a.2) The vibration occurs immthe raw US data before clutter filtering. (bbetween 100 and 1,000 Hz. (b.3) The vibra

en different frequency components over tim

Figure 3. Representative vibrometry wavseverities. (a) Iliac bifurcation (range gate wpercent common iliac stenosis on the rigstenosis on the right (arrow). The vibratstenosis severity increases.

US data acquisition. Three iliac arteries

were normal (�30% diameter reductionat angiography).

Perivascular tissue velocities in an-giographically normal arteries (�30%diameter reduction) did not show anysignificant vibrations (the frequency

. (a.1) The Doppler range gate is placed iniately following the R wave of systole. (b.1)The vibration signal waveform is isolatedspectral waveform. This display indicates

rms from iliac artery stenoses of differentplaced just after the bifurcation). (b) Sixty

(arrow). (c) Eighty percent common iliacfrequency and duration increase as the

me eryart ed

om .2)rm tion

efoas

htion

content of tissue displacement signals

diameter.

4 • Doppler Vibrometry Assessment of Arterial Stenosis Month 2009 JVIR

ARTICLE IN PRESS

was �250 Hz). Figure 3a shows thevibrometry signals and spectral wave-forms over one complete cardiac cyclefrom a representative common iliac ar-tery with no angiographic stenosis. Asthe stenosis severity increases, the vi-bration frequency and duration in-crease. Figures 3b and 3c show rep-resentative vibrometry signals andspectral waveforms for two differentseverities of common iliac artery ste-noses. Figure 4 shows scattergramsof the vibration frequency and dura-tion versus the angiographic per-centage stenosis and the reciprocal ofthe minimum lumen diameter. Thevibration frequency was correlatedwith the percentage stenosis (Pear-son r � 0.75; P � .001) and inverselycorrelated with the minimum lumendiameter (r � 0.72; P � .001). Thecorrelation between the vibrationduration and the percentage stenosiswas 0.7 (P � .001). The vibration du-ration was correlated with the vibra-tion frequency (r � 0.82; P � .001)(Fig 5). Stepwise regression showedthat, among vibration frequency andduration, the vibration frequencywas the primary independent signif-icant predictor of the angiographicpercentage stenosis (P � .05).

DISCUSSION

Lees and colleagues (4,10,11) devel-oped the phonoangiography methodfor predicting the minimum lumen di-ameter of a carotid stenosis on the ba-sis of a quantitative analysis of thefrequency spectrum of cervical bruitsdetected on the skin surface. The hy-pothesis behind quantitative phonoan-giography is that the frequency spec-trum of the sound produced by vesselwall vibrations is dependent on theeffective diameter of the turbulent jet,assuming that the systolic flow rate tothe brain is the same in all patients. Itwas shown that the bruit spectrum ex-hibits a peak frequency beyond whichthe energy falls off rapidly with in-creasing frequency.

From theoretical fluid dynamicconsiderations, the break frequency ofturbulent fluctuations is inversely pro-portional to the length scale of turbu-lence (the residual lumen diameter atthe stenosis), provided that the Strou-hal number and the mean down-stream velocity (blood velocity in the

unobstructed vessel) remain constant.

Figure 4. Scattergrams demonstrate the correlations between vibration characteristicsand angiographic measurements. (a) Vibration frequency and percentage stenosis, (b)vibration frequency and reciprocal of minimum lumen diameter, (c) vibration durationand percentage stenosis, and (d) vibration duration and reciprocal of minimum lumen

Figure 5. Scattergram shows correlation between vibration duration and vibration

frequency.

he

Sikdar et al • 5Volume xx Number x

ARTICLE IN PRESS

It was empirically observed that in thecarotid arteries a simple relationshipexists between the break frequencyand the residual lumen diameter, byassuming that the unobstructed meanvelocity in the carotid artery is approx-imately constant at 500 mm/sec (6,11).The minimum lumen diameter (in mil-limeters) could then be obtained bydividing 500 by the break frequency(in hertz).

Several clinical studies demonstratedthe validity of this method (5–7), and itwas shown that the effective residuallumen diameter in carotid artery steno-ses was accurately estimated with thistechnique to within 1 mm of the angio-graphic value in 93% of 170 cases (12).Several researchers have experimentallyconfirmed the production of vessel wallvibrations due to underlying flow tur-bulence (9,13–15). However, phonoan-giography lacks specificity about thesource of the bruit. For example, a cer-vical bruit could be caused by an inter-nal or external carotid stenosis or a con-ducted cardiac murmur (6). In addition,a substantial number of bruits recordedat the skin surface are too weak to beanalyzed (6). These limitations madephonoangiography less attractive forroutine clinical use compared to duplexDoppler US.

The presence of the bruit signatureon spectral Doppler waveforms andcolor Doppler images is well known(1,16–18). The common presentationof a bruit is a double-sided clutter onspectral Doppler waveforms and a fo-cal random mixture of red (motion to-ward the US transducer) and blue(motion away from the transducer)colors forming a mosaic pattern oncolor Doppler images. The perivascu-lar bruit artifacts have been shown tohave clinical utility in the identifica-tion of hemodynamically significantstenoses (19).

Our results provide preliminary ev-idence that the vibration frequency es-timated from the Doppler bruit spec-trum can be used as a predictor ofstenosis severity. Doppler vibrometrycan complement conventional duplexexaminations, especially in technicallychallenging cases, because vibrometrydoes not require visualization of thevessel lumen (20). The measurementsof the vibration frequency and dura-tion with use of Doppler vibrometryare independent of the Doppler angle

(21). Furthermore, echoes from tissue

are significantly stronger (60 dB ormore) than those from blood; thus,Doppler vibrometry can provide anadvantage over duplex US in hard-to-image patients.

The vibration characteristics variedwith the placement of the Dopplerrange gate. Figure 6 shows the changein the vibration characteristics as theDoppler range gate was placed up-stream, immediately downstream, andfarther downstream of a severe com-mon iliac artery stenosis. The vibrationamplitude is highest immediatelydownstream of the stenosis, as ex-pected. Therefore, the vibration ampli-tude can be used as a guide for stenosislocation. Downstream of the stenosis,the high-frequency components of thevibration are attenuated. This is ex-pected because only the larger, lower-frequency eddies are convected down-stream.

Figure 6. Variation of the vibration characdiameter reduction common iliac artery stegate was placed proximal to the stenosisDoppler range gate was placed immediategate was placed �15 mm downstream of t

In addition to the frequency, the

duration of the bruit has been pro-posed as an important diagnostic pa-rameter (7). In our present study, thevibration duration was correlatedwith the stenosis severity but did nothave any additional predictive valuefor determining the angiographic se-verity compared with the vibrationfrequency. We believe that the dura-tion of the bruit could be an indirectmeasure of the hemodynamically-significant poststenotic pressuredrop. The components of the pres-sure drop across the stenosis are theBernoulli pressure depression due toan increase in velocity in the post-stenotic jet and the viscous loss dueto turbulence. The Bernoulli pressuredepression is recovered as the flowvelocity returns to normal fartherdownstream of the stenosis if turbu-lence is not present. The irrecover-able viscous loss due to turbulence is

stics with position along the artery for 80%sis (same case as Fig 1). (a) Doppler rangemediately after the iliac bifurcation. (b)fter the stenotic region. (c) Doppler rangestenosis.

terinoim

ly a

the component of the pressure drop

6 • Doppler Vibrometry Assessment of Arterial Stenosis Month 2009 JVIR

ARTICLE IN PRESS

that contributes to the reduction inthe end organ flow. The pressuredrop across the stenosis provides theenergy to sustain turbulence in post-stenotic flow and the resulting vibra-tions. Therefore, it may be arguedthat a short-duration vibration corre-sponds to a smaller pressure dropacross the stenosis that occurs onlyduring the peak systolic decelerationphase, whereas a longer-duration vi-bration corresponds to a larger pres-sure drop that is sustained for a longerduration in systole. Further studies areneeded to confirm this hypothesis.

Although renal artery, iliac artery,and superior mesenteric artery ste-noses showed good agreement withthe overall trend, hepatic artery ste-noses tended to be outliers both interms of vibration frequency and du-ration (note the large variability offrequency and duration values fortwo 80% diameter reduction hepatic

Figure 7. Variation in the vibration charahepatic artery stenosis (arrow). Both stenosThe hepatic artery stenosis is more difficuvarious factors.

artery stenoses in Fig 4, and the case

for which no measurable vibrationdata could be obtained). Figure 7shows two stenoses of similar sever-ity in the superior mesenteric arteryand proper hepatic artery. The vibra-tions in the hepatic artery stenosisare significantly weaker. We believethat three factors could have poten-tially contributed to this: (a) the an-atomic location of hepatic arteriesand their lack of a close couplingwith surrounding soft tissue com-pared to other arterial beds, such asrenal and iliac; (b) substantial move-ment of the hepatic artery and sur-rounding tissues during respirationcausing difficulties in placement ofthe range gate close to the stenosisregion; and (c) the presence of bowelgas blocking the vibrometry signal,either preventing the acquisition ofmeasurable data or leading to under-estimation of the vibration fre-quency. Hepatic artery flow rates are

ristics with vascular bed. (a) Superior mesre of similar severity, but their measured vo image, and it is possible that the vibrati

also variable depending on the fast-

ing state of the patient, which couldexplain some of the variability in thevibrometry findings (22).

One potential application of Dopplervibrometry could be as an adjunct toduplex US examination for renal arterystenosis. A major challenge for diagnos-ing renal artery stenosis with duplex USis the difficulty in visualizing renal ar-teries, especially in patients with sub-stantial body fat. A further complicationis the complex anatomy of the renal ar-teries in many patients due to the highprevalence of accessory renal arteries.Stenoses in accessory renal arteriescould contribute to renovascular hyper-tension, yet they are often missed inroutine duplex examinations. Further-more, the Doppler range gate must beplaced at multiple locations along therenal artery and renal parenchyma for acomplete evaluation of all possible ste-notic sites, which makes the examina-tion time-consuming. In addition to

eric artery stenosis (arrow) and (b) propertion characteristics are markedly different.characteristics are underestimated due to

cte entes a ibralt t on

these technical challenges, the choice of

Sikdar et al • 7Volume xx Number x

ARTICLE IN PRESS

an appropriate diagnostic parameter forrenal artery stenosis continues to be con-troversial because the commonly usedparameters (eg, renal artery peak sys-tolic velocity, renal aortic ratio, renal re-sistive index) have been shown to havelimitations (23). Bruits have the highestpositive predictive value among theclinical signs for renovascular disease(24,25) and are prevalent in 77%–87% ofpatients with angiographically provedrenal artery stenosis. The diagnosis ofrenovascular hypertension with stetho-scope auscultation of abdominal bruitshas a high specificity (90%–99%) and asomewhat low sensitivity (39%–63%)compared with angiography (24). Webelieve that ultrasonic vibrometry couldhave greater sensitivity for the detectionof abdominal bruits compared to stetho-scope auscultation, which would in-crease the sensitivity of bruit analysis. Inaddition, the quantitative analysis ofDoppler bruit frequency, duration, andamplitude would contribute to the clin-ical diagnosis of renal artery stenosis byproviding complementary measure-ments to the conventional indexes onthe basis of blood flow velocity.

Some limitations of our studyshould be acknowledged. The smallsample size of this study prevented adetailed evaluation of the differencesin vibration characteristics from dif-ferent vascular beds. Future studieswith a larger sample size are neededto determine the inter- and intraob-server variability and repeatabilityof the vibrometry method.

In conclusion, advances in Dopplertechnology and the development of di-agnostic criteria based on blood flowvelocities have overshadowed the clin-ical use of phonoangiography andquantitative analysis of bruits. How-ever, the principles of phonoangiogra-phy can be applied to US echo signalsand have the advantages of greatersignal strength, no visualization of thevessel lumen, and no dependence onangle in situations where conventionalDoppler methods have limitations. Ourpreliminary results show that quantita-tive analysis of Doppler bruits with vi-

brometry can be used to predict the se-

verity of stenoses in renal, iliac, andsuperior mesenteric arteries.

References1. Strandness DE Jr, ed. Doppler ultra-

sonic techniques in vascular disease.4th ed. St. Louis, Mo: Mosby; 1993.

2. Soares GM, Murphy TP, Singha MS,Parada A, Jaff M. Renal artery duplexultrasonography as a screening andsurveillance tool to detect renal arterystenosis: a comparison with currentreference standard imaging. J Ultra-sound Med 2006; 25:293–298.

3. Williams GJ, Macaskill P, Chan SF, etal. Comparative accuracy of renal du-plex sonographic parameters in the di-agnosis of renal artery stenosis: pairedand unpaired analysis. AJR Am JReontgenol 2007; 188:798–811.

4. Lees RS, Dewey CF Jr. Phono-angiography: a new noninvasive di-agnostic method for studying arterialdisease. Proc Natl Acad Sci USA1970; 67:935–942.

5. Duncan GW, Gruber JO, Dewey CF Jr,Myers GS, Lees RS. Evaluation of ca-rotid stenosis by phonoangiography.N Engl J Med 1975; 293:1124–1128.

6. Knox R, Breslau P, Strandness DE Jr.Quantitative carotid phonoangiogra-phy. Stroke 1981; 12:798–803.

7. Tavel ME, Bates JR. The cervicalbruit: sound spectral analysis related toseverity of carotid arterial disease. ClinCardiol 2006; 29:462–465.

8. Sikdar S, Beach KW, Vaezy S, Kim Y.Ultrasonic technique for imaging tissuevibrations: preliminary results. Ultra-sound Med Biol 2005; 31:221–232.

9. Jones SA, Fronek A. Analysis ofbreak frequencies downstream of aconstriction in a cylindrical tube. J Bio-mech 1987; 20:319–327.

10. Miller A, Lees RS, Kistler JP, AbbottWM. Spectral analysis of arterial bruits(phonoangiography): experimental vali-dation. Circulation 1980; 61:515–520.

11. Kistler JP, Lees RS, Miller A, CrowellRM, Roberson G. Correlation of spec-tral phonoangiography and carotid an-giography with gross pathology in ca-rotid stenosis. N Engl J Med 1981; 305:417–419.

12. Lees RS, Myers GS. Noninvasive di-agnosis of arterial disease. Adv InternMed 1982; 27:475–509.

13. Khalifa AM, Giddens DP. Cha-

racterization and evolution postste-

notic flow disturbances. J Biomech1981; 14:279 –296.

14. Kirkeeide RL, Young DF, Cholvin NR.Wall vibrations induced by flow throughsimulated stenosis in models and arter-ies. J Biomech 1977; 10:431–441.

15. Plett MI, Beach KW, Dunmire B, BrownKG, Primozich JF, Strandness E Jr. Invivo ultrasonic measurement of tissuevibration at a stenosis: a case study.Ultrasound Med Biol 2001 Aug;27:1049–1058.

16. Middleton WD, Erickson S, MelsonGL. Perivascular color artifact: patho-logic significance and appearance oncolor Doppler US images. Radiology1989; 171:647–652.

17. Zweibel W. Introduction to vascularultrasonography. 5th ed. Philadelphia,Pa: Saunders, 2005.

18. Nilsson A. Artefacts in sonographyand Doppler. Eur Radiol 2001; 11:1308–1315.

19. Hatsukami TS, Primozich JF, ZierlerRE, Harley JD, Strandness DE Jr.Color Doppler imaging of infraingui-nal arterial occlusive disease. J VascSurg 1992; 16:527–531.

20. Sikdar S, Lee JC, Remington J, et al.Ultrasonic Doppler vibrometry: novelmethod for detection of left ventricularwall vibrations caused by poststenoticcoronary flow. J Am Soc Echocardiogr2007; 20:1386–1392.

21. Sikdar S, Bo Z, Zaccardi M, Leotta D,Beach KW, Yongmin K. UltrasonicDoppler measurement of tissue vi-brations: opportunities and limita-tions. Presented at the IEEE Confer-ence Proceedings, Los Alamitos,California, October 2– 6, 2006.

22. Numata K, Tanaka K, Kiba T, Mitsui K,Saito S, Sekihara H. Hemodynamicchanges in hepatic artery after glucoseingestion in healthy subjects and pa-tients with cirrhosis. J Clin Ultrasound1998; 26:137–142.

23. Leiner T, de Haan MW, Nelemans PJ,van Engelshoven JM, Vasbinder GB.Contemporary imaging techniques forthe diagnosis of renal artery stenosis.Eur Radiol 2005; 15:2219–2229.

24. Albers FJ. Clinical characteristics ofatherosclerotic renovascular disease.Am J Kidney Dis 1994; 24:636–641.

25. Svetkey LP, Helms MJ, Dunnick NR,Klotman PE. Clinical characteristicsuseful in screening for renovasculardisease. South Med J 1990; 83:743–

747.