Received: 16 January 2004Revised: 4 May 2004Accepted: 7 May 2004Published online: 17 July 2004© Springer-Verlag 2004

Abstract Comparison of volumecontrast US imaging with tissue har-monic imaging for the evaluation ofgallbladder lesions and determina-tion of the adequate slice thicknessin volume contrast US imaging wereperformed. Forty-one patients whohad gallbladder lesions (polyps in26, stones in 12, and sludge in 3)were enrolled in our study. A Voluson730 Expert US scanner was usedthroughout. Volume contrast US im-aging with slice thicknesses of 3, 5,10 and 15 mm and tissue harmonicimaging of the gallbladder were ob-tained. Two abdominal radiologistsreviewed the masked images andgraded by consensus these imagesusing a five-point scale [from grade1, (best), to grade 5, (worst)], basedon the sharpness of the anterior gall-bladder wall, internal artifact, lesionconspicuity and acoustic shadowingfrom stone. Volume contrast US im-aging with thin slice thicknesses (3 or 5 mm) was judged superior toboth tissue harmonic imaging and

with thick slice thicknesses (10 or15 mm), with respect to the sharp-ness of the anterior wall and lesionconspicuity (P<0.001). In terms ofinternal artifact, volume contrast imaging with thin slice thicknesseswas significantly superior to both tissue harmonic imaging and volumecontrast imaging with a 15 mmthickness (P<0.001) and was judgedto be marginally better than with a10 mm thickness (p>0.01). With re-gard to acoustic shadowing, volumecontrast imaging with thin slicethicknesses was also significantlybetter than with thick slice thick-nesses (P<0.01), and it was alsomarginally better than tissue har-monic imaging (P>0.01). Volumecontrast US imaging with thin slicethicknesses provides a better imagequality with fewer artifacts thanthree other types of images for theevaluation of gallbladder diseases.

Keywords Gallbladder · US · Technology

Eur Radiol (2004) 14:1657–1664DOI 10.1007/s00330-004-2376-3 H E PAT O B I L I A RY- PA N C R E A S

Se Hyung KimJeong Min LeeKyoung Ho LeeYoung Jun KimSu Kyung AnChang Jin HanJoon Koo HanByung Ihn Choi

Four-dimensional volume contrast ultrasoundimaging of the gallbladder compared with tissueharmonic imaging: preliminary experience

Introduction

Although CT or MR imaging could be used in patientssuspected to have gallbladder (GB) diseases, ultrasonog-raphy (US) remains the first line explorative modality forGB diseases and produces excellent results depicting theGB and its pathology [1–3]. However, reverberation andside lobe artifact, which are intrinsic features of conven-tional US, make the diagnosis and differentiation of GBlesions difficult [4]. Two harmonic techniques, namely,

tissue harmonic imaging (THI) and phase inversion har-monic imaging (PIHI) were introduced and seemed toovercome these problems [5–8]. In THI, however, abandpass filter is used to process the received signal sothat only the returning high-frequency harmonic signal isused to produce the image. Resolution and sensitivity arethus limited by a fundamental compromise in the fre-quency-filtering approach. Furthermore, harmonic echoesare relatively weak when compared with conventionalechoes, so that electronic noise may be an issue [6].

S. H. Kim · J. M. Lee · Y. J. KimS. K. An · C. J. Han · J. K. HanB. I. Choi (✉)Department of Radiology, College of Medicine,Seoul National University Hospital,28, Yongon-dong, Chongno-gu, Seoul, 110-744, Koreae-mail: choibi@radcom.snu.ac.krTel.: +82-2-7602515Fax: +82-2-7436385

J. M. Lee · J. K. Han · B. I. ChoiInstitute of Radiation Medicine, College of Medicine,Seoul National University Hospital,28, Yongon-dong, Chongno-gu, Seoul, 110-744, South Korea

K. H. LeeSeoul National University Bundang Hospital,Gyunggi-do, South Korea

Moreover, problems derived from weak harmonic signalswere maximized in superficial or deep-seated organs.

Volume contrast imaging (VCI) is a new four-dimen-sional (4D) US technology based on a real-time volumeacquisition and has the potential benefit of contrast en-hancement and speckle suppression in the 2D US image.In this article, the word “contrast” does not mean the useof intravenous contrast material. The main aspect of VCIis the use of a thin volume consisting of multi-slices in-stead of a single B-plane. Furthermore, VCI can be usedin combination with harmonic imaging, can boost the ad-vantages of harmonic imaging and can overcome itsshortcomings. From our preliminary study on the feasi-bility of VCI with the harmonic technique as applied tobile duct diseases, a VCI with 3-mm slice thickness wasfound to better depict the ductal wall and the intraductalpathology and to better suppress internal artifact thanTHI alone (Kim SH, unpublished data). However, to ourknowledge, the usefulness of VCI combined with theharmonic technique for various GB lesions vs. THI alonehas not been reported upon previously. The purpose ofthis study was to compare VCI of various slice thick-nesses using the harmonic technique with THI alone andto determine the optimal slice thickness of VCI for theevaluation of GB lesions.

Materials and methods

A prospective study over a period of one month using two differentUS techniques, VCI and THI, was performed on 41 consecutivepatients who had gallbladder lesions. The patient population con-sisted of 25 men and 16 women (mean age: 52 years, age range:21–74 years). The causes of referral for ultrasonography werescreening in asymptomatic patients (n=9), follow-up of known liver tumors (n=6) or screening examination in liver cirrhosis orchronic hepatitis (n=21) and evaluation of abdominal symptoms(n=5). Our study was approved by the ethics committee of our hos-pital, and written informed consent was obtained from each patient.

In order to differentiate a polyp from a gallstone or sludge, GBpolyps on US were defined as a lesion projecting into the lumen ofthe GB that did not show an acoustic shadow and did not movewith a positional change of the patient. A GB stone was defined asa lesion lying in the dependent portion of the GB that showed aposterior acoustic shadow and changed in position with movement.

The final diagnosis included GB polyps (3–23 mm; mean 7 mm) in26 patients, gallstones (5–50 mm; mean 14 mm) in 12 and sludgein three. Of these, three patients with polyp were confirmed histo-pathologically by cholecystectomy. One of them was diagnosedmalignant melanoma with GB metastasis. A diagnosis of stoneswas verified by cholecystectomy in three and by CT in seven.

To investigate the feasibility of VCI in assessing GB diseasesand to determine an adequate slice thickness of VCI, we used bothUS techniques (VCI of 3-, 5-, 10- and 15-mm slice thickness com-bined with the harmonic technique and THI alone) with a Voluson730 Expert (GE Kretz Ultrasound, Zipf, Austria) and a 4–8 MHzvolume probe. US scanning was performed by one abdominal radiologist (S.H.K). When the lesion was localized on THI first,machine settings such as depth, focal zone and time-gain compen-sation (TGC) were optimized. A single focal zone was located onthe center of the lesions. GB lesions were then scanned by VCIwith a 3-mm slice thickness, which was followed by 5-, 10- and15 mm slice thickness in the same imaging plane. The default set-ting of other machine parameters for VCI was set to mechanicalindex (MI: 1.2), the frame rate ranged from 7 to 15 Hz (15, 13, 10and 7 Hz for 3-, 5-, 10- and 15-mm slice thicknesses, respec-tively). The mechanical index and frame rate for THI was set to1.2 and 40 Hz, respectively. US scanning parameters except framerate were the same for both THI and VCI combined with harmonictechniques. Sets of representative best images of the GB in eachpatient were selected, using both THI and VCI. All images weremasked to exclude annotation of the technique used and wereevaluated qualitatively on a computer monitor in a side-by-sidemanner. All images were analyzed in terms of the sharpness of theanterior GB wall, internal artifact, lesion conspicuity and acousticshadow behind a stone by consensus between two abdominal radi-ologists (J.M.L. and J.K.H) blind to the imaging techniques usedduring image interpretation. Each of these four parameters wasrated using a five-point scale: the best technique was ranked as 1and the worst as 5.

Statistical analysis was performed using the Friedman test andthe Wilcoxon signed ranks test using SPSS version 10.0 (SPSS,Inc., Chicago, IL). For the Friedman test, a P value of less than0.05 was required for null hypothesis rejection. For the Wilcoxonsigned ranks test, 0.01 was required as a P value for null hypothe-sis rejection to minimize the alpha error.

Results

Total grades for the four parameters are summarized inTable 1, with details of the comparative statistical analy-sis of five types of images using two techniques (VCIand THI) in Table 2. Representative examples are shownin Figs. 1, 2, 3, 4.

1658

Table 1 Total numbers of grades assigned to the five types of images for each parameter

Parameter Sharpness of anterior Internal artifact Lesion conspicuity Acoustic shadowing from GB wall (n=41) (n=41) (n=41) stones (n=12)

Grade 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

THI alone 0 5 16 10 10 0 2 2 12 25 0 6 15 12 8 0 6 4 1 1VCI (3) 24 12 3 1 1 14 20 4 2 1 30 8 2 0 1 9 1 1 1 0VCI (5) 14 17 5 4 1 18 15 7 0 1 15 18 8 0 0 3 6 3 0 0VCI (10) 5 7 9 18 2 8 7 21 2 3 2 3 14 20 2 0 1 5 5 1VCI (15) 1 3 5 8 24 2 1 5 22 11 0 0 4 8 29 0 0 3 3 6

THI, tissue harmonic imaging; VCI (3), volume contrast imaging with 3-mm thickness; VCI (5), volume contrast imaging with 5-mmthickness; VCI (10), volume contrast imaging with 10-mm thickness; VCI (15), volume contrast imaging with 15-mm thickness.

Sharpness of the anterior GB wall

The order of median grade for the five types of imageswas VCIs with 3-mm (1.6), 5-mm (2.0), 10-mm (3.1),THI (3.6) and finally 15-mm thicknesses (4.2). VCI witha thin slice thickness (3 or 5 mm) was judged superior toboth THI and VCI with a thick slice thickness (10 or15 mm) (P< 0.001). VCI with 3 mm tended to be betterthan VCI with 5 mm, but the difference was not statisti-cally significant (P=0.066). VCI with 10 mm and THIwere better than THI and VCI with 15 mm, respectively,but the difference was also not statistically significant(P=0.127 and 0.014, respectively). However, VCI with10 mm was significantly better than with 15 mm(P<0.001). Most of the VCIs with a thin slice thickness(3 or 5 mm) were ranked as 1 or 2, whereas THI andVCI with a thick slice thickness (10 or 15 mm) wereranked as 3, 4 or 5. In addition, THI received no grade 1scores.

Internal artifact

The order of median grade for the five types of imageswas VCIs with 5 mm (1.8), 3 mm (1.9), 10 mm (2.6) and15 mm (4.0), and THI (4.5). VCI with a thin slice thick-ness was significantly superior to both THI and VCIwith 15 mm (P<0.001). VCIs with 5-, 3- and 15-mmthickness were better than VCIs with 3-, 10-mm thick-ness, and with THI, respectively, but the difference wasnot statistically significant (P=0.511, 0.017 and 0.029,respectively). However, VCI with 5- or 10-mm thicknesswas significantly better than VCI with 10- or 15-mmthickness, respectively (P=0.004 and P<0.001). Most ofVCIs with a thin slice thickness (3 or 5 mm) were rankedas 1 or 2, whereas VCI with a thick slice thickness (10 or15 mm) and THI were mainly ranked as 3, 4 or 5. THIreceived no grade 1 scores.

Lesion conspicuity

The order of median grade for the five types of imageswas VCIs with 3 mm (1.4), 5 mm (1.8), 10 mm (3.4) andTHI (3.5), followed by 15-mm thickness (4.6). VCI witha thin slice thickness (3 or 5 mm) was judged superior toboth THI and VCI with a thick slice thickness (10 or15 mm) (P<0.001). VCIs with 3 and 10 mm were betterthan VCI with 5 mm and THI, respectively, but this wasnot statistically significant (P=0.012 and 0.685, respec-tively). However, THI was significantly better than VCIwith 15-mm thickness (P<0.001).

Most of the VCIs with a thin slice thickness (3 or5 mm) were ranked as 1 or 2, whereas THI and VCI witha thick slice thickness (10 or 15 mm) were mainlyranked as 3, 4 or 5. THI and VCI with 15-mm thicknessreceived no grade 1 scores, and VCI with 5-mm thick-ness received no grade 5 scores.

Acoustic shadowing from stones

The order of median grades for the five types of imageswas VCIs with 3-mm (1.5), 5-mm (2), THI (2.75),

1659

Table 2 Comparative statistical analyses of the five types of images as used to evaluate each parameter

Parameters Results

Sharpness of anterior GB wall VCI (3) = VCI (5) > VCI (10) = THI = VCI (15), VCI (10) > VCI (15)Internal artifact VCI (5) = VCI (3) = VCI (10) > VCI (15) = THI, VCI (5) > VCI (10)Lesion conspicuity VCI (3) = VCI (5) > VCI (10) = THI > VCI (15)Acoustic shadowing from stones VCI (3) = VCI (5) = THI = VCI (10) = VCI (15), VCI (3) = VCI (5) > VCI (10) = VCI (15)

VCI (3)=THI, THI > VCI (15)

VCI (3), volume contrast imaging with 3-mm thickness; VCI (5),volume contrast imaging with 5-mm thickness; VCI (10), volumecontrast imaging with 10-mm thickness; THI, tissue harmonic im-aging; VCI (15), volume contrast imaging with 15-mm thickness.

Data were analyzed using the Wilcoxon signed rank test. A > B: Ais better than B with statistical significance (P<0.01); A=B: Nostatistically significant difference between A and B.

Fig. 1 Acquisition of a thin volume by volume contrast imaging(VCI). Within this volume, there is a render box with a large sur-face, but the thickness of the render box is kept small. The thick-ness of the rendered box can be chosen by the user as either 3, 5,10 or 15 mm to suit the purpose at hand. A thin volume within therender box is projected onto a 2D screen in the direction of therendering process (arrow)

1660

Fig. 2 A gallstone in a 45-year-old man. VCIs with a 3-, b 5-, c 10- and d 15-mm slice thicknesses and e THI, the sharpness ofanterior GB wall is clearest in the VCIs with thin slice thicknesses(a, b). Echo-free acoustic shadowing with a clear margin behind astone was obtained at VCI with 3-mm thickness (a). However, as agroup, no statistical difference was observed between VCIs withthin slice thicknesses and THI alone in that parameter. When thethicker slice thickness was applied, grayer acoustic shadowing

was obtained (arrows, in c, d). There was a tendency toward ablurred margin of the GB wall (arrowheads) and obscuring poste-rior shadowing from a stone on VCIs with thicker slice thickness-es (c, d) due to partial volume artifact. The internal artifact waswell suppressed by VCI compared with THI alone. Note the in-creased echogenicity and improved lesion conspicuity of the stoneby VCI compared with THI alone

1661

Fig. 3 A small (4 mm) polyp (arrow) in the anterior wall of GB ina 43-year-old woman. Lesion conspicuity and sharpness of the an-terior GB wall were better on VCI with a 3- and b 5-mm thicknessthan VCI with c 10- and d 15-mm thickness and e THI alone.

Therefore, small GB polyp was clearly visualized on VCI with athin slice thickness. Internal artifacts were more suppressed byVCI with 3- (a) and 10-mm (c) slice thicknesses than by THIalone (e)

1662

Fig. 4 Multiple polyps of the gallbladder in a 44-year-old man. a b c d, VCI with 3-, 5-, 10- and 15-mm slice thickness; e THI.The sharpness of the anterior GB wall and lesion conspicuity (arrows) were best on VCI with 5-mm thickness (b). However, asa group, statistical differences were not found between VCIs with 3- (a) and 5-mm (b) thickness. Internal artifacts including rever-beration artifacts were more suppressed by VCI with thin slice

thicknesses (a, b) than by THI alone (e). Note the near total obliteration of the GB lumen and inner multiple polyps by VCIwith 15-mm thickness (d) due to partial volume artifact. Theechogenicity of polyps was higher in VCI with 10-mm thickness(c) than in THI alone (e), but lesion conspicuity was worse in VCIwith 10-mm thickness due to partial obliteration of polyps. Thisobliteration was caused by partial volume averaging artifact

10-mm (3.5) and 15-mm thickness (4.3). VCIs with 3-and 5-mm thickness were significantly superior to bothVCI with 10- (P=0.003 for 3 mm, P=0.007 for 5 mm)and 15-mm thickness (P=0.002 for 3 mm and P=0.005for 5 mm), and better than THI, but differences were notstatistically significant (P=0.034 for 3 mm and P=0.059for 5 mm). THI and VCI with 10-mm thickness werebetter than VCI with 10- and 15-mm thickness, respec-tively, but differences were not significant (P=0.143 and0.034, respectively). However, THI was found to be sig-nificantly superior to VCI with 15-mm thickness(P=0.007). Most of the VCIs with a thin slice thickness(3 or 5 mm) were ranked as 1 or 2, whereas THI andVCI with a thick slice thickness (10 or 15 mm) weremainly ranked as 3, 4 or 5. THI, VCI with a thick thick-ness (10 or 15 mm) received no grade 1 scores, and VCIwith a thin thickness (3 or 5 mm) thickness received nograde 5 scores.

Discussion

VCI is a new 4D and “multi-slice” US imaging technolo-gy that utilizes volume probes to acquire slices of tissuecontinuously and rapidly in real time. Four-dimensionalUS is a dynamic real-time 3D US with a continuous volume data set acquisition and rendering process. InVCI of the present study, a continuous real-time volumeacquisition and 3D rendering process—even though itoccurs in an axial direction—are also performed withreal-time operation; therefore, VCI technique can meetthe requirements of 4D US.

The basic concept of VCI utilizes a thin volume,which consists of about 10–25 B-planes depending onthe thickness setting, for example, 3, 5, 10 or 15 mm in-stead of a single B-slice. It provides contrast enhance-ment and speckle suppression compared with 2D US in asingle B-plane. This also makes it possible to overcomethe impaired axial resolution that is one of disadvantagesof THI, caused by filtering out the fundamental compo-nent with a narrow bandpass filter. This was supportedby the findings of the present study, in which VCI withthinner slice thicknesses showed better image qualitycompared to THI alone. In the present study, VCI with athin slice thickness allowed the better depiction of ante-rior wall of GB and its pathology, improved the enhance-ment of posterior shadowing behind stones and intro-duced fewer internal artifacts than THI alone. These su-perior results agreed well with the theoretical advantagesof VCI. However, in most cases, VCI with thick slicethicknesses produced poorer results than THI alone. OnCT scan, partial volume artifacts are known to be a mainproblem in the interpretation of small cystic lesions ortransversely running structures [9, 10]. These artifactsare increased with thick slice thickness and high pitchand, therefore, they can be overcome by adequately ad-

justing scan parameters, such as by narrowing the colli-mation, applying low pitch, an overlapped reconstructionand summing thin slices [11, 12]. In our study, the poorresult for VCI with thick slice thicknesses seems to bemainly responsible for these artifacts. Polyps and stonescan be totally or partially obscured by a thick GB wallbecause of partial volume artifacts. We also presume thatthese artifacts in VCIs with thicker slice thicknesses areinfluenced by lesion size. But, in our study, the size ofthe lesion was not at issue. In addition, acoustic shadow-ing from stones was obscured in VCIs with thick slicethicknesses. This observation is also presumed to be theeffects of partial volume artifacts.

The mechanism of VCI has not yet been explained inthe literature. The information obtained from multi-sliceVCI technology is processed by volume rendering inwhich combined surface and transparent maximum gra-dient rendering are performed. The basic formula forvolume rendering is given by

where Ix,y is the result pixel of the image, gi is the grayvalue orthogonal to image plane, p is the probability(transparency) value for one gray value.

The idea behind this formula is to project a volumeonto a 2D screen. The function given for the transparen-cies p(gi) is set in case of US to a linear function, in-creasing with the gray values and set to zero for valuesbeyond a certain gray threshold. The default setting ofthe volume rendering for VCI is a mixture of 70% sur-face texture rendering and 30% transparent maximumgradient. One hundred percent surface rendering wouldonly display the signals of the first B-slice, and onlywhen the render ray meets a gap in the speckle patternwould it pass on the second B-slice for analysis. Thismeans that the 30% transparent maximum provides anamount of tissue information from the directly neighbor-ing B-slices. This causes the speckle pattern, which isseen in any conventional B-mode image to be suppressedin VCIs without averaging, which would create a blurredimage. Furthermore, these rendering processes have avery high sample rate, which means the density of therendering analysis is higher than the density of pixels ina single B-slice. This means that gaps in the speckle pat-tern are filled up with information from adjacent slices inthe multislices. Speckle suppression also produces a verysmooth image, which has not lost information detail be-cause it is not just a 2D filter. These apparent advantagesmay allow VCI to play an important role in the diagnosisof GB diseases, especially in the detection of subtle andanteriorly located lesions, such as small polyps or adeno-myomatosis in GB, in which reverberation or a side lobeartifact may cause diagnostic problems.

Our results for THI alone appear poor, which is in-consistent with several previous reports in which THI

1663

1664

shows clear advantages in examinations of the biliarysystem [5, 6, 13]. This discordance might be caused byseveral factors, for example, THI results are likely to bemachine-specific. Moreover, in the other studies, THIwas compared with conventional imaging and showedexcellent results with respect to contrast and side lobe ar-tifact reduction; however, our study compares THI withVCI using THI. It was not surprising that we obtainedbetter results when VCI was combined with THI than using THI alone due to the synergism of the technologyfusion.

We should mention several limitations of this study.First, the number of cases was relatively small, whichprevents our generalizing these preliminary results. Fur-ther studies with larger cases are warranted to investigatethe feasibility of VCI in various clinical applications.Second, we did not consider the size of the lesion in de-termining the effect of slice thickness on image quality.However, the size of most lesions in this study was lessthan 1 cm; thus, VCI with slice thickness greater than

10 mm would be expected to be significantly influencedby partial volume artifacts. Third, our study populationis limited by the recruitment of consecutive patients withGB diseases of low clinical significance; therefore, his-topathologic proof was not available in most cases.

The objective of our study was to evaluate the feasi-bility of VCI for assessing GB lesions compared withTHI alone, not to evaluate its diagnostic performance atdetecting GB lesions. Although our results seem to beexcellent, further prospective studies are needed to as-sess the diagnostic performance of VCI. In addition,many reports about the clinical applications of VCI tech-nology as applied to other organs are also needed.

In conclusion, VCI combined with the harmonic tech-nique, especially with thin slice thicknesses, provides abetter image quality and fewer artifacts than THI alonefor the evaluation of GB diseases.

Acknowledgements We thank Helmut Brandl, PhD, R&D engi-neer of GE Kretz Ultrasound, for helpful discussion about techni-cal aspects of volume contrast imaging.

References

1. Rooholamini SA, Tehrani NS, RazaviMK et al (1994) Imaging of gallblad-der carcinoma. Radiographics14:291–306

2. Yarmenitis SD (2002) Ultrasound ofthe gallbladder and the biliary tree. Eur Radiol 12:270–282

3. Bennett GL, Balthazar EJ (2003) Ultra-sound and CT evaluation of emergentgallbladder pathology. Radiol ClinNorth Am 41:1203–1216

4. Mostbeck GH, Mallek R (1993) Possibilities of error and pitfalls in sonography of the gallbladder and bileducts (German). Radiologe 33:19–26

5. Flunker S, Aube C, Anglade E et al(2001) Value of tissue harmonic imag-ing in biliary lithiasis. GastroenterolClin Biol 25:589–594

6. Hong HS, Han JK, Kim TK et al(2001) Ultrasonographic evaluation ofthe gallbladder: comparison of funda-mental, tissue harmonic, and pulse inversion harmonic imaging. J Ultra-sound Med 20:35–41

7. Lencioni R, Cioni D, Bartolozzi C(2002) Tissue harmonic and contrast-specific imaging: back to gray scale inultrasound. Eur Radiol 12:151–165

8. Hohl C, Schmidt T, Haage P et al(2004) Phase-inversion tissue harmonicimaging compared with conventionalmode ultrasound in the evaluation ofpancreatic lesions. Eur Radiol14:1109–1117

9. Siegel CL, Fisher AJ, Bennett HF(1999) Interobserver variability in determining enhancement of renalmasses on helical CT. Am JRoentgenol 172:1207–1212

10. Diederichs CG, Keating DP, GlattingG, Oestmann JW (1996) Blurring ofvessels in spiral CT angiography: effects of collimation width, pitch,viewing plane, and windowing in maximum intensity projection. JCAT20:965–974

11. Heuscher DJ, Vembar M (1999) Reduced partial volume artifacts usingspiral computed tomography and an integrating interpolator. Med Phys26:276–286

12. Kasales CJ, Hopper KD, Ariola DN et al (1995) Reconstructed helical CTscans: improvement in z-axis resolu-tion compared with overlapped andnonoverlapped conventional CT scans.Am J Roentgenol 164:1281–1284

13. Ortega D, Burns PN, Hope Simpson D,Wilson SR (2001) Tissue harmonic imaging: is it a benefit for bile duct sonography? Am J Roentgenol176:653–659