characterization of focal liver lesions: comparison of pulse-inversion harmonic contrast-enhanced...

Characterization of Focal Liver Lesions:Comparison of Pulse-Inversion HarmonicContrast-Enhanced Sonography withContrast-Enhanced CT

Rui Li, MD,1 Yanli Guo, MD,1 Xing Hua, MD,1 Yun He, MMed,1 Jun Ding, MMed,1 Aiming Guo, MMed,1

Mingkui Luo, PhD2

1 Department of Ultrasound, Southwest Hospital, Third Military Medical University, Chongqing 400038, China2 Department of Mathematics, Third Military Medical University, Chongqing 400038, China

Received 17 October 2005; accepted 29 August 2006

ABSTRACT: Purpose. To compare the efficacy of

contrast-enhanced pulse-inversion harmonic sonog-

raphy for the characterization of focal liver lesions

with that of contrast-enhanced helical CT.

Methods. Real-time contrast-enhanced sonography

(CEUS) using Sonovue and contrast-enhanced CT

(CECT) were performed on 109 patients with focal liver

lesions, including 61 hepatocellular carcinomas, 15 liver

metastases, 5 cholangiocellular carcinomas, 12 heman-

giomas, 5 regenerative nodules, 3 adenomas, 3 focal

nodular hyperplasias, 4 focal necroses, and 1 angio-

myolipoma prior to surgery or percutaneous needle bi-

opsy. The diagnostic performance was assessed by

using histopathological results as reference standards.

Results. Three cases were missed on CEUS, and 7

cases were missed on CECT. These 10 missed cases

were excluded from paired statistical comparison. Ten

cases were misdiagnosed on CEUS and 17 cases were

misdiagnosed on CECT. The overall accuracy was

89.9% (89/99) for CEUS and 82.8% (82/99) for CECT.

The difference between CEUS and CECT was not stat-

istically significant. Concordance between CEUS and

CECTwas observed in 90.9% (90/99) cases.

Conclusion. Real-time pulse-inversion harmonic

CEUS with Sonovue is comparable with CECT in the

characterization of focal liver lesions. VVC 2007 Wiley

Periodicals, Inc. J Clin Ultrasound 35:109–117, 2007;

Published online in Wiley Interscience (www.

interscience.wiley.com). DOI: 10.1002/jcu.20310

Keywords: liver; neoplasms; contrast media; ultra-

sonography; computed tomography

Characterization of various hepatic tumors isof great therapeutic and prognostic relevance

and has thus been one of the focuses of multipleimaging research studies. The key to characteriz-ing focal liver lesions lies in the temporal courseof enhancement and the vascular architecture,because they clearly differ between various typesof tumor.1 Studies by Bezy-Wendling et al2 andNino-Murcia et al3 have demonstrated that con-trast-enhanced CT (CECT) can show type-specificdiagnostic differences between different types ofliver tumor. However, detecting slow or low vol-ume flow has been limited with conventionalcolor Doppler or power Doppler sonography. Thisproblem has been resolved through the use ofultrasound contrast agents.

Sonovue (Bracco, Geneva, Switzerland) is asulfur hexafluoride gas stabilized with phospholi-pids that presents a high reflectivity at a low me-chanical index and persists in the blood streammuch longer after intravenous injection thanprevious contrast agents. This mechanism ofenhancement allows real-time imaging of themicrocirculation of the liver for several minutes.Thus, the temporal course of contrast mediumenhancement and the vascular architecture canbe analyzed in a continuous manner.

The purpose of this study was to investigatethe efficacy of contrast-enhanced sonography(CEUS) using the contrast agent Sonovue andpulse-inversion imaging for the characterizationof focal liver lesions by comparing the results tothose obtained with CECT.

Correspondence to: R. Li

' 2007 Wiley Periodicals, Inc.

VOL. 35, NO. 3, MARCH/APRIL 2007—DOI 10.1002/jcu 109

PATIENTS ANDMETHODS

In this prospective study, 109 patients (37women, 72 men; age range 18–79 years, meanage 46 6 12 years) were examined with conven-tional sonography and unenhanced CT scan. Thepatients were admitted to our hospital and exam-ined via both real-time pulse-inversion harmonicCEUS and CECT.

Abdominal CTwas performed using 16-slice spi-ral CT (Somatom Sensation; Siemens, Erlangen,Germany) using a 3-phase contrast-enhancedprotocol. First, an unenhanced scan of the liverwas obtained. Next, after intravenous infusion of100–120 ml of a nonionic iodine-containing con-trast agent (Ultravist 370; Schering, Berlin, Ger-many) using a power injector (Stellant CT InjectionSystem; Medrad, Indianola, PA) at a rate of 4 ml/sec, CECT scans were obtained in arterial, portalvenous, and late phase with bolus test trigger foroptimal characterization of focal hepatic lesions.Data were obtained through the whole liver in acraniocaudal direction during a single breath-holdhelical acquisition of 6–8 seconds.

Helical CT findings were evaluated subjec-tively by 2 radiologists who were blinded to theCEUS findings. Established CT criteria3 wereused to characterize the lesions.

The first step of sonographic examination con-sisted of conventional baseline sonography of theentire liver. This examination was performed todetect and localize focal hepatic lesions. Inpatients with more than 1 lesion in the liver, thelargest and most conspicuous lesion was sub-jected to the following examination.

The second step of sonographic examinationwas CEUS with the contrast agent Sonovue(Bracco). This second-generation contrast agentconsists of microbubbles of sulfur-hexafluoride(SF6).

Sonovue is licensed for use in abdominal andvascular imaging in China. Oral informed con-sent was obtained from all patients after the na-ture of the procedure had been thoroughlyexplained. The ethical principles of Declarationof Helsinki4 were strictly followed.

The patient’s history was reviewed with spe-cial attention to severe cardiovascular and pul-monary diseases, known allergies, and preg-nancy. A volume of 2.4 ml Sonovue was injectedinto the cubital vein in bolus via a 20-gauge nee-dle followed by a 5-ml saline flush. All examina-tions were performed with an HDI 5000 ultra-sound unit (Philips Ultrasound, Bothell, WA).CEUS was performed using a low mechanicalindex (0.09–0.15) pulse-inversion harmonic imag-ing setting that was automatically adjusted by

the contrast-specific software of the scanner witha 2–5-MHz convex-array broadband transducer.Images were stored on magnetic optical discs andsuper-VHS videotapes.

After contrast medium injection, hepaticlesions were scanned continuously for up to5 minutes until the enhancement effect began tosubside. The entire vascular phase was studied,consisting of the arterial phase (0–40 secondsfrom beginning of contrast agent bolus injection),portal venous phase (41–100 seconds after theinjection), and late phase (101–300 seconds afterthe injection). Contrast enhancement was subjec-tively assessed in consensus between 2 sonolo-gists who were unaware of the CECT findingsusing the criteria reported in Table 1, which weredeveloped on the basis of enhancement patternsrecently described by Quaia et al5 and Dietrich.6

TABLE 1

Diagnostic Criteria of CEUS for Focal Hepatic Lesions

Lesion CEUS Criteria

HCC Tortuous intratumoral vessels and diffuse

enhancement during arterial phase that

decreases during portal and late phase

(hypoechoic appearance)

Metastasis Enhancing peripheral rim, variable

intratumoral enhancement during arterial

phase that decrease during portal and late

phase (hypoechoic appearance)

CCC Variable intratumoral vessels and

heterogeneous peripheral enhancement

during arterial phase that decrease during

portal and late phase (hypoechoic

appearance), dilatation of the bile ducts

near the tumor may be accentuated after

enhancement

Hemangioma Nodular peripheral enhancement during arterial

phase, with centripetal progression during

arterial, portal venous, and late phase

FNH Central spoke-wheel enhancement during early

arterial phase that becomes homogeneous

quickly during arterial phase, homogeneous

enhancement similar to that of the liver

parenchyma during portal and late phase,

star-shaped hypoechoic area corresponding

to central scar during late phase

Regenerative

nodule

Enhancement lower than or similar to that of

the surrounding liver parenchyma during

all phases

Adenoma Diffuse homogeneous or heterogeneous

enhancement during arterial phase,

enhancement similar to that of the

surrounding liver parenchyma during

portal venous and late phase

Focal necrosis Lack of internal enhancement during all

phases, sharp demarcation with

surrounding enhanced liver parenchyma

Angiomyolipoma Heterogeneous enhancement during arterial

phase, slightly more echogenic than

surrounding liver parenchyma during

portal venous and late phase

Abbreviations: CCC, cholangiocellular carcinoma; CEUS, contrast-

enhanced sonography; FNH, focal nodular hyperplasia; HCC, hepato-

cellular carcinoma.

LI ET AL

110 JOURNAL OF CLINICAL ULTRASOUND—DOI 10.1002/jcu

The gold standard was the histopathologicdiagnosis, which was established in 56 casesby means of surgical resection and in 53 cases

via percutaneous needle biopsy with an 18-gauge needle within 2 weeks after CEUS andCECT.

FIGURE 1. A 57-year-oldman with a small 1.8-cm HCC in the right lobe of the liver. (A) Transverse CEUS sonogram of the right lobe (RL) 10 seconds

after Sonovue injection shows HCC with diffuse contrast enhancement (arrow). (B) Sonogram obtained 1minute and 7 seconds after contrast injec-

tion shows HCC to be hypoechoic (arrow) compared with the adjacent liver. The lesion was correctly characterized as HCC by CEUS. (C, D) CECT

shows no enhancement during arterial phase (C), portal venous phase, and late phase (D). The lesionwasmisdiagnosed as a hepatic cyst on CECT.

TABLE 2

Comparison of Results of CEUS and CECT in the Characterization of 109 Liver Lesions

Pathologic

Diagnosis No.

CECT CEUS

Correct Incorrect Lesion Not Visualized Correct Incorrect Lesion Not Visualized

HCC 61 51 7 3 54 5 2

Metastases 15 13 0 2 14 0 1

CCC 5 3 2 0 4 1 0

Hemangioma 12 8 3 1 11 1 0

Regenerative nodule 5 2 2 1 4 1 0

FNH 3 1 2 0 2 1 0

Adenoma 3 2 1 0 2 1 0

Focal necrosis 4 2 2 0 4 0 0

Angiomyolipoma 1 1 0 0 1 0 0

Total 109 83 19 7 96 10 3

Abbreviations: CCC, cholangiocellular carcinoma; CECT, contrast-enhanced CT; CEUS, contrast-enhanced sonography; FNH, focal nodular

hyperplasia; HCC, hepatocellular carcinoma.

LIVER LESIONS: ENHANCED SONOGRAPHY AND CT


Differences within groups were evaluatedusing McNemar’s test. A p value of less than 0.01was considered statistically significant.

RESULTS

The final pathologic diagnosis included 61 hepa-tocellular carcinomas (HCCs), 15 metastases, 5cholangiocellular carcinomas (CCCs), 12 heman-giomas, 5 regenerative nodules, 3 adenomas, 3focal nodular hyperplasias (FNHs), 4 cases offocal necrosis, and 1 angiomyolipoma. The 109focal hepatic lesions ranged in size from 0.9 to12.8 cm (mean 2.9 6 1.3 cm). The results ofCEUS and CECT are summarized in Table 2.

Nineteen focal hepatic lesions were misdiag-nosed on CECT. Nine malignant tumors (7 HCCs,2 CCCs) were misdiagnosed as benign on CECT(Figure 1) and 6 benign lesions were misdiag-

nosed as malignant on CECT (Figures 2, 3). OneFNH was misdiagnosed as adenoma, while 2 focalnecroses and 1 regenerative nodule were mis-diagnosed as hepatic cysts on CECT.

Ten focal hepatic lesions were misdiagnosed onCEUS. Two HCCs showed dotted contrast en-hancement during the arterial phase, followed byisoechoic appearance during the late phase, andwere then misdiagnosed as regenerative nodules.One HCC had nodular peripheral enhancementduring the arterial phase, with progressive cen-tripetal fill-in during portal venous phase wasincorrectly diagnosed as hemangioma. Two HCCshad diffuse enhancement during arterial phase,appeared isoechoic relative to the adjacent liverduring the late phase, and were incorrectly char-acterized as benign lesions. One CCC that had nointernal enhancement during the arterial phasewith irregular coalescent appearance, had a sharpboundary between the lesion and surrounding

FIGURE 2. A 39-year-old woman with a 2.5-cm hemangioma in the right lobe of the liver. (A) Transverse CEUS scan 6 seconds after Sonovue

injection during early arterial phase shows nodular peripheral enhancement (arrow). (B) CEUS scan 12 seconds after injection shows very quick

centripetal fill-in (arrow) of the lesion in the right lobe of the liver (RL). The lesion was correctly diagnosed as hemangioma on CEUS. (C, D) CECT

shows only diffuse homogeneous enhancement pattern (arrow) during the arterial phase (C) and hypodensity (arrow) compared with the sur-

rounding liver during the late phase (D). The lesion was misdiagnosed as HCC on CECT.

LI ET AL


hepatic parenchyma, and had persistent hypoe-choic appearance during portal venous and latephase was incorrectly diagnosed as a hepatic ab-scess (Figure 4). One FNH (2.7 cm in diameter)that had diffuse and homogeneous enhancementduring the arterial phase and was slightly hypere-choic during the late phase, but had no typicalspoke-wheel contrast enhancement, was misdiag-nosed as hepatocellular adenoma. One cholangio-cellular adenoma and 1 regenerative nodule witha dysplastic histologic pattern that had diffusecontrast enhancement during the arterial phaseand appeared hypoechoic during the late phasewere incorrectly diagnosed as HCCs (Figure 5).CEUS correlated with CECT in 9 lesions, except 1HCC that was misdiagnosed as a regenerativenodule on CEUS but was correctly diagnosed asan atypical HCC on CECT.

Seven lesions (3 HCCs, 2 metastases, 1 heman-gioma, and 1 regenerative nodule, average size2.0 6 0.7 cm) were missed (ie, not visible) onCECT and were correctly characterized byCEUS. Three lesions (2 HCCs and 1 metastasis,average size 2.0 6 0.3 cm) were missed (ie, notvisible) on CEUS (1 located adjacent to the dia-phragm and 2 near the colon). The metastasiswas correctly diagnosed, and the 2 HCCs wereincorrectly characterized on CECT. These 10missed lesions were excluded from statisticalcomparison, because this study concentrated onthe characterization of focal hepatic lesions. Thepaired outcomes of the remaining 99 lesions areshown in Table 3.

The overall accuracy in the diagnosis of focalhepatic lesion was 89.9% (89/99) for CEUS and82.8% (82/99) for CECT. The difference between

FIGURE 3. A 42-year-old man with a well-defined homogeneous hypoechoic 3.1-cm mass in the right lobe. (A) Transverse CEUS sonogram shows

a central spoke-wheel enhancement (arrow) 10 seconds after the injection of Sonovue in the right lobe of the liver (RL). (B) The contrast agent is

filling rapidly from the central arterial supply toward the periphery, and enhancement of the lesion is diffuse and homogeneous (arrow) 17 sec-

onds after Sonovue injection. The lesion was correctly characterized as FNH on CEUS. (C) CECT does not show a spoke-wheel but a homogeneous

enhancement pattern of the nodule (arrow), and the lesion was misdiagnosed as HCC. (D) Bisected surgical specimen shows the nodule with typi-

cal central scar and fibrous capsule. The diagnosis of FNH was confirmed histologically.



CEUS and CECTwas not statistically significant.Concordance between CEUS and CECT wasobserved in 90/99 cases (90.9%).

DISCUSSION

With the advent of new equipment and software,sonography has improved in recent years. Theimage resolution of unenhanced sonography ishigh enough to detect small focal lesions in theliver, especially when new techniques such as tis-sue harmonic imaging are used. Although sonog-raphy can reveal focal lesions of the liver, its abil-ity to assess the nature of the lesion is disap-pointing.7–9 Although color and power Dopplersonography have already revealed a characteris-tic pattern of vascularization in some cases, onlythe introduction of contrast agents opens up the

possibility to visualize small vessels withintumors. CEUS was found to increase the visibil-ity of tumor vessels in 95% of HCCs10 and wassuitable to demonstrate vascularity in lesionsthat were vascular on CECT but not on colorDoppler sonography.11 The use of color doppler toprovide vascular information is often limited inhepatic tumors that are small, deep in location,or subject to motion artifacts from either respira-tory or cardiac activity, because Doppler artifactsdue to moving solid tissue are often stronger thanthe Doppler signals reflecting the blood flow insmall vessels.

Contrast-enhanced pulse-inversion sonogra-phy using the contrast agent Levovist (SHU508A) has been shown to improve the characteri-zation of focal liver lesions.12–14 Levovist can beinsonated only by using high acoustic power to

FIGURE 4. A 69-year-old man with a 5.7-cm cholangiocellular carcinoma in the right lobe of the liver. (A) Oblique CEUS scan during the arterial

phase shows no enhancement inside the lesion with irregular coalescent appearance and sharp boundary (arrow) 8 seconds after Sonovue injec-

tion. (B) CEUS sonogram 3 minutes and 28 seconds after injection shows persistent hypoechoic appearance of the mass (arrow) in the right lobe

of the liver (RL). (C) CECT scan shows no enhancement inside the irregular low density lesion (arrow) during the arterial phase. (D) There is no

enhancement in the lesion with coalescent appearance and sharp delineation (arrow) from the surrounding hepatic parenchyma during the late

phase. The lesion was incorrectly diagnosed as hepatic abscess on both CEUS and CECT.

LI ET AL


produce extensive bubble destruction with emis-sion of a wideband frequency signal that isdetectable with contrast-specific imaging set-tings. Sonographic examination has to be per-formed intermittently, because microbubbles arerapidly destroyed. Destructive high acousticpower scanning produces more tissue harmonics,with a more marked mixture of tissue and micro-bubble harmonics and the production of a sub-stantial tissue background on which the bubblesignal is superimposed.5 Nondestructive sono-

graphic examination performed with low acousticpower and second-generation contrast agentssuch as Sonovue provides new possibilities. Theadvantage of this technique is its ability to allowthe continuous visualization of the arterial, por-tovenous, and late phases in 1 session after a sin-gle administration of contrast agent.

Ultrasound contrast agents do not disperse intothe extracellular space and therefore permit amore accurate demonstration of persisting bloodflow in untreated tumors.15 Sonovue is character-ized by a low solubility in water and a low diffu-sion in blood and persists in the blood streammuch longer after intravenous injection than thecontrast agents used previously in sonography, soit can actually be considered a blood pool contrastagent.16 CEUS performed with Sonovue andpulse-inversion under low acoustic power sonog-raphy makes it possible to show the vasculararchitecture and temporal course of enhancementin various types of liver tumor in real-time. Sono-

FIGURE 5. Cholangiocellular adenoma in a 46-year-old woman. (A) Conventional sonogram shows a well-defined heterogeneous hypoechoic

nodule (arrow) in the right lobe of the liver (RL) on a sagittal plane. (B) The nodule is diffusely and homogeneously enhanced 7 seconds after

Sonovue injection (arrow). (C) The lesion becomes progressively hypoechoic (arrows) relative to the adjacent liver parenchyma during the late

phase (158 seconds). The lesion was misdiagnosed as HCC on both CEUS and CECT. (D)Microphotograph of resected specimen (hematoxylin-eo-

sin; original magnification 200�). The pathologic diagnosis was cholangiocellular adenoma with grade 2 dysplastic change.

TABLE 3

Paired Contingency of CEUS and CECT Diagnosis

of 99 Focal Liver Lesions

CECT Correct CECT Incorrect

CEUS correct 81 8

CEUS incorrect 1 9

Abbreviations: CECT, contrast-enhanced CT; CEUS, contrast-

enhanced sonography.



vue-enhanced sonography improved sensitivityfrom 78% to 100% and specificity from 23% to 92%in the diagnosis of benign versus malignant liverlesions compared with baseline sonography.17

Overall diagnostic accuracy increased from 51%to 88%, and the diagnostic confidence—that is, theROC curve area (Az)—was 0.83 before versus 0.98after review of CEUS.5

In our study, real-time CEUS with Sonovueincorrectly diagnosed 10.1% (10/99) of focal he-patic lesions. Four HCCs had diffuse or dottedenhancement during arterial phase, appearedisoechoic during the late phase and were mis-diagnosed as benign lesions. It is not clear whysome HCCs have this peculiar enhancing behav-ior. A possible explanation is that well-differenti-ated HCCs have higher probability of isoechoicappearance with respect to the surrounding liverduring the late phase than poorly and moderatelydifferentiated HCCs.18 Therefore, further studiesare necessary to determine whether cellular dif-ferentiation, vascular architecture and contrastenhancement differences exist in HCCs. OneCCC showed some features of a pyogenic hepaticabscess, including lack of internal enhancement,irregular coalescent appearance, and sharp boun-dary between the lesion and surrounding hepaticparenchyma and was misdiagnosed as a hepaticabscess. Quaia et al5 reported that 17 of 17 (100%)FNHs larger than 3 cm but only 3 of 9 (33.3%)FNHs smaller than 3 cm in diameter had a patho-gnomonic spoke-wheel contrast enhancement pat-tern. In this study, 1 FNH (2.7 cm in diameter)had diffuse and homogeneous enhancement dur-ing the arterial phase and slightly hyperechoicduring the late phase instead of a typical spoke-wheel contrast enhancement, and was conse-quently misdiagnosed as hepatocellular adenoma.One cholangiocellular adenoma and 1 regenera-tive nodule with dysplastic histologic pattern haddiffuse contrast enhancement during the arterialphase and appeared hypoechoic during the latephase, and were therefore incorrectly diagnosedas HCC. Since the distinction betweenmacrorege-nerative nodules and well-differentiated HCCs isoften difficult even at histologic diagnosis,19 anoverlap in CEUS characterization is understand-able. CEUS correlated with CECT in 9 lesionsexcept 1HCC that wasmisdiagnosed as a regener-ative nodule on CEUS but correctly diagnosed asan atypical HCC on CECT.

Seventeen focal hepatic lesions (17.2%) weremisdiagnosed on CECT. Although the differencewas not statistically significant when comparedwith that of CEUS, there were actually 7 morelesions incorrectly characterized on CECT than

on CEUS. The superiority of CEUS over CECTcould be explained by 2 factors. First, the ultra-sound contrast agent acts like a blood pool agentthat enhances the entire intravascular space,and Sonovue CEUS using low mechanical indeximaging reflects arterial enhancement better inmalignant liver tumors than CECT.20 In thisstudy, 5 of 56 HCCs (8.9%) were seen as hyper-vascular lesions on CEUS but showed hypovascu-lar enhancement pattern on CECT, and werethen incorrectly diagnosed as cysts, focal necro-sis, or hepatic abscess. Giorgio et al21 alsoreported that 10% of small HCCs were hypervas-cular lesions on CEUS but showed a hypovascu-lar pattern on CECT. Second, continuous real-time imaging at a low mechanical index, whichpermits a continuous monitoring of the temporalcourse of enhancement and visualizes more deli-cately vascular architecture that may be missedon CT, which is dependent on interval-delayimaging acquisition even during the first secondsof the arterial phase (eg, rapid centripetal fillingin some small hemangiomas, central spoke-wheelHCCs vessel in focal nodular hyperplasias, tortu-ous irregular intratumoral vessel branches insmall HCCs. Also, the interval between bolusinjection of contrast agent and tumor enhance-ment may vary from patient to patient. The peaktumor enhancement time may be outside thetime window of the arterial phase of routine heli-cal CT in some patients. Therefore, some type-specific enhancement patterns during arterialphase may not be captured on CECT.

Nevertheless, there are still some limitationsto CEUS, such as operator dependence, limiteddepth, superimposed air and bone, and the factthat in patients with multiple lesions, the lesionscan only be studied one at a time because of rapidcontrast filling and washout of most hypervascu-lar lesions.

From the results of this study, we concludethat pulse-inversion harmonic Sonovue CEUS iscomparable to CECT in the characterization offocal liver lesions, and represents a complementto unenhanced conventional sonography for thatpurpose.

ACKNOWLEDGMENT

This study was financially supported by the Clin-ical New Technology Foundation of SouthwestHospital (SWH2005A004).

REFERENCES

1. Folkman J. Role of angiogenesis in tumor growthand metastasis. Semin Oncol 2002;29:15.

LI ET AL


2. Bezy-Wendling J, Kretowski M, Rolland Y. Hepatictumor enhancement in computed tomography:combined models of liver perfusion and dynamicimaging. Comput BioMed 2003;33:77.

3. Nino-Murcia M, Olcott EW, Jeffrey RB, et al. Focalliver lesions: pattern-based classification schemefor enhancement at arterial phase. Radiology 2000;215:746.

4. World Medical Association Declaration of Helsinki.Ethical principles for medical research involvinghuman subjects. Available at: www.wma.net/e/policy/pdf/17c.pdf, accessed December 8, 2003.

5. Quaia E, Calliada F, Bertolotto M, et al. Character-ization of focal liver lesions with contrast specificUS modes and a sulfur hexafluoride-filled micro-bubble contrast agent: diagnostic performance andconfidence: Radiology 2004;232:420.

6. Dietrich CF. Characterization of focal liver lesionswith contrast enhanced ultrasonography. Eur JRadiol 2004;51 (Suppl 1):S9.

7. Rode A, Bancel B, Douek P, et al. Small noduledetection in cirrhotic livers: evaluation with US,spiral CT, and MRI and correlation with pathologicexamination of explanted liver. J Comput AssistTomogr 2001;25:327.

8. Bennett GL, Krinsky GA, Abitbol RJ, et al. Sono-graphic detection of hepatocellular carcinoma anddysplastic nodules in cirrhosis correlation of pre-transplantation sonography and liver explant pa-thology in 200 patients. AJR Am J Roentgenol2002;179:75.

9. Teefey A, Hildeboldt CC, Dehdashti F, et al. Detec-tion of primary hepatic malignancy in liver trans-plant candidates: prospective comparison of CT,MR imaging, US and PET. Radiology 2003;226:533.

10. KimAY, Choi BI, Kim TK, et al. Hepatocellular cari-noma: power Doppler US with a contrast agent—preliminary results. Radiology 1998;209:135.

11. Cedrone A, Pompili M, Sallustio G, et al. Compari-son between color power Doppler ultrasound withecho-enhancer and spiral computed tomography inthe evaluation of hepatocellular carcinoma vascu-larization before and after ablation procedures.Am J Gastroenterol 2001;96:1854.

12. Dietrich CF, Ignee A, Trojan J, et al. Improvedcharacterisation of histologically prove tumors bycontrast enhanced ultrasonography during theportal venous and specific late phase of SHU 508A.Gut 2004;53:401.

13. Bryant TH, BlomleyMJ, Albrecht T, et al. Improvedcharacterization of liver lesions with liver-phaseuptake of liver-specific microbubbles: prospectivemulticenter study. Radiology 2004;232:799.

14. Isozaki T, Numata K, Kiba T, et al. Differential di-agnosis of hepatic tumors by using contrast en-hancement patterns at US. Radiology 2003;229:798.

15. Solbiati L, Goldberg SN, Lerace T, et al. Radio fre-quency ablation of hepatic metastases: post proce-dural assessment with a US microbubble contrastagent-early experience. Radiology 1999;211:643.

16. Morel DR, Schwieger I, Hohn L, et al. Humanpharmacokinetics and safety evaluation of Sono-vue, a new contrast agent for ultrasound imaging.Invest Radiology 2000;35:80.

17. Herbay AV, Vogt C, Willers R, et al. Real-timeimaging with the sonographic contrast agent SonoVue. J Ultrasound Med 2004;23:1557.

18. Nicolau C, Catala V, Vilana R, et al. Evaluation ofhepatocellular carcinoma using Sono Vue, a secondgeneration ultrasound contrast agent: correlationwith cellular differentiation. Eur Radiol 2004;14:1092.

19. Longchampt E, Patriarche C, Fabre M. Accuracyof cytology vs microbiopsy for the diagnosis of well-differentiated hepatocellular carcinoma and mac-roregenerative nodule: definition of standardizedcriteria from a study of 100 cases. Acta Cytol 2000;44:515.

20. Krix M, Kiessling F, Essig M, et al. Low mechani-cal index contrast-enhanced ultrasound betterreflects high arterial perfusion of liver metastasesthan arterial phase computed tomography. InvestRadiology 2004;39:216.

21. Giorgio A, Ferraioli G, Tarantino L, et al. Con-trast-enhanced sonographic appearance of hepato-cellular carcinoma in patients with cirrhosis: com-parison with contrast-enhanced helical CT appear-ance. AJR Am J Roentgenol 2004;183:1319.



characterization of focal liver lesions: comparison of pulse-inversion harmonic contrast-enhanced...

Documents