Gastrointestinal Imaging Original Research

Gastrointestinal Imaging Original Research Arita et al. Gastrointestinal Imaging Original Research Junichi Arita 1 Kiyoshi Hasegawa Michiro Takahashi Shojiro Hata Junichi Shindoh Yasuhiko Sugawara Norihiro Kokudo Arita J, Hasegawa K, Takahashi M, et al. Keywords: contrast-enhanced ultrasound, hepatocellular carcinoma (HCC), histologic grade, intraoperative ultrasound, microbubbles, perfluorobutane, Sonazoid DOI:10.2214/AJR.10.4310 Received January 21, 2010; accepted after revision November 10, 2010. This work was supported in part by a grant-in-aid (21791269) for scientific research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, Tokyo; a grant-in-aid (H20-hepatitis-general-001) for research on the development of early detection systems for liver cancer using molecular markers and diagnostic imaging from the Ministry of Health, Labor, and Welfare of Japan, Tokyo; and a grant from the Japanese Society for Advancement of Surgical Techniques, Osaka. 1 All authors: Divisions of Hepato-Biliary-Pancreatic Surgery and Artificial Organ and Transplantation, Department of Surgery, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8655, Japan. Address correspondence to J. Arita (jarita-tky@umin.ac.jp). AJR 2011; 196:1314 1321 0361 803X/11/1966 1314 American Roentgen Ray Society Correlation Between Contrast- Enhanced Intraoperative Ultrasound Using Sonazoid and Histologic Grade of Resected Hepatocellular Carcinoma OBJECTIVE. Our aim was to accurately assess the correlation between findings of contrast-enhanced intraoperative ultrasound using Sonazoid and histologic grade of hepatocellular carcinoma (HCC). SUBJECTS AND METHODS. We enrolled 239 consecutive patients who were undergoing surgery for HCC for this study. Because 33 extensively necrotic HCCs were excluded, a total of 374 histologically proven HCCs were detected in all resected specimens and were the study subjects (71 well-differentiated, 239 moderately differentiated, and 64 poorly differentiated HCCs). After a laparotomy and liver mobilization, contrast-enhanced intraoperative ultrasound in the harmonic mode was performed after a Sonazoid injection. The first minute was defined as the vascular phase, in which the vascularity of the 239 HCCs was assessed. After an approximately 15-minute delay, a thorough liver exploration was performed (Kupffer phase). Preoperative dynamic CT was routinely performed, and the findings were assessed for reference. RESULTS. The proportion of hypervascular tumors during the vascular phase tended to be lower among well-differentiated than among moderately and poorly differentiated HCCs (66% vs 80%, p = 0.058). The proportion of hypoechoic tumors during the Kupffer phase was significantly lower among well-differentiated than among moderately and poorly differentiated HCCs (54% vs 92%, p < 0.0001). In dynamic CT, the proportions of hypervascular tumors during the early phase and hypodense tumors during the late phase were significantly lower among well-differentiated HCCs than among moderately and poorly differentiated HCCs, respectively (early phase, 51% vs 87%, p < 0.0001; late phase, 59% vs 85%, p < 0.0001). CONCLUSION. Contrast-enhanced intraoperative ultrasound using Sonazoid is useful for estimating the histologic grade of HCC. H epatocellular carcinoma (HCC) is one of the most common cancers worldwide and is a major cause of death in patients with cirrhosis [1]. The histologic grade of HCC strongly correlates with tumor activity [2] and thus is an established prognostic factor [3, 4]. Well-differentiated HCCs less frequently exhibit vascular invasion and intrahepatic metastasis [5 8]. Therefore, estimating the histologic grade of HCC is helpful for selecting a treatment strategy. Contrast-enhanced ultrasound has been developed for a differential diagnosis of focal liver lesions found in cirrhotic patients [9 12]. Two widely used contrast agents for ultrasound, gaseous perflutren (Definity, Bristol-Myers Squibb) and gaseous sulfur hexafluoride (SonoVue, Bracco) are both pure intravascular contrast agents [13]. Gaseous perfluorobutane (Sonazoid, GE Healthcare) is also a second-generation contrast agent that has recently become commercially available in Japan. It is not a pure intravascular contrast agent but has the peculiarity of accumulating in Kupffer cells in the liver [13 15] and thus enables Kupffer imaging in addition to showing tumor vascularity [16]. Superparamagnetic iron oxide (SPIO) enhanced MRI, which also enables Kupffer imaging, has been reported to be well correlated with the histologic grade of HCC [17 19]. Two studies have compared Kupffer phase images of contrast-enhanced ultrasound using Sonazoid and SPIO-enhanced MRI and shown good agreement for both images [20, 21]. One of the two studies further compared the histologic grade of HCC with Kupffer phase images of contrast-enhanced ultrasound using Sonazoid and showed fair correlation; however, they assessed 1314 AJR:196, June 2011

only 48 HCCs, including as few as 13 welldifferentiated HCCs, and, moreover, the majority of the specimens for histologic grading were obtained by needle biopsy instead of surgical resection [21]. Histologic examination of a resected specimen is preferable to that of a needle biopsy specimen because the latter yields up to 30% of false-negative findings among small HCCs [22 24] and both methods show low compatibility in the histologic grading of HCC [25]. In this study, findings of contrast-enhanced intraoperative ultrasound using Sonazoid for surgically resected HCCs were compared with their histologic grade. Contrast enhanced ultrasound was performed in an intraoperative setting to secure a precise evaluation of ultrasound images, avoiding dead angles, breath motion artifact, and ultrasonic attenuation through the abdominal wall. Subjects and Methods Patients The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the institutional review board. This study was conducted in accordance with a prospective protocol. Two hundred ninety-four consecutive patients who were scheduled to undergo surgery for HCC between February 7 and September 9 were considered for enrollment in the study. Of them, 55 patients were excluded from the study for the following reasons: absence of informed consent in 30 patients, unavailability of the required devices in eight, or the number of histologically proven HCCs was more than nine in 17 (a precise assessment by histology or by imaging studies was impossible because of possible numerous tiny intrahepatic metastases). The remaining 239 patients were enrolled in the study. Written informed consent was obtained from each patient included in the study. The baseline characteristics of the enrolled patients are summarized in Table 1. Diagnosis of HCC All specimens resected in the enrolled patients were fixed in formalin, sectioned serially, stained with H and E, and examined by two expert pathologists per patient on the basis of the criteria of the International Working Party [26]. Tumors were not excluded on the basis of size or depth considering the advantages of intraoperative ultrasound [27]. In total, 407 histologically proven HCCs were detected among the 239 patients. For a precise assessment of histologic grade, 33 HCCs were excluded from the analysis because more than two thirds of the entire lesion was necrotic. Thus, the remaining 374 tumors were included in this study. TABLE 1: Characteristics of Patients Characteristics Values Sex Men 198 Women 41 Age a (y) 68 (21 86) HBsAg 48 HCVAb 120 Both HBsAg and HCVAb 1 Neither HBsAg or HCVAb 72 Background liver Normal liver 11 Chronic hepatitis 138 Liver cirrhosis 90 Number of HCCs b (1/2/3/4 or more) 1 158 2 45 3 20 4 16 Size of largest HCC a (cm) 2.7 (0.8 20.0) Note Values are expressed as number. HBsAg = hepatitis B surface antigen, HCVAb = hepatitis C virus antibodies, HCC = hepatocellular carcinoma. a Values are expressed as median (range). b Including extensively necrotic tumors. The cellular differentiation of HCC was graded as either of well-, moderately, or poorly differentiated on the basis of the above-mentioned criteria [26]. When a tumor had several components with the various cellular differentiations, the histologic grade of the tumor was decided according to the lowest cellular differentiation among all components. This arrangement is based on the following consideration: a component of lower cellular differentiation is more hypervascular [10, 28, 29] and has fewer Kupffer cells [17 19, 30, 31] in general and thus more strongly influences the contrast-enhanced intraoperative ultrasound characterization of the tumor. Consequently, 374 tumors were classified as 71 well-, 239 moderately, and 64 poorly differentiated HCCs; the median (range) size of the tumors was 1.4 cm (0.2 9.5 cm), 1.8 cm (0.2 20.0 cm), and 2.5 cm (0.2 17.0 cm), respectively. Dynamic CT Preoperative dynamic CT was routinely performed in the patients with HCC, and the findings were also assessed as a reference. The CT systems used in this study consisted of 4-, 8-, or 16-MDCT scanners (Aquilion 4 and Aquilion 16, Toshiba; LightSpeed Qx/i and LightSpeed Ultra, GE Healthcare). After conventional CT, dynamic contrast-enhanced CT was performed with a bolus injection of iodinated contrast agent using a mechanical power injector; 2 ml/kg body weight (maximum, 100 ml) of contrast agent at a concentration of 300 370 mg I/mL was injected in 30 seconds through the antecubital vein. CT was performed at 25, 45, and 120 seconds after starting the injection. The former was referred to for the assessment in the early phase and the latter in the late phase. Reconstruction images were obtained with a section thickness of 2.0 2.5 mm and a section interval of 1.5 2.0 mm. A focal liver lesion was categorized as either hyper-, iso-, or hypovascular during the early phase and categorized as either hyper-, iso-, or hypodense during the late phase compared with the surrounding liver parenchyma in consensus between two expert radiologists who were in charge for a given patient. A focal liver lesion that was hard to differentiate from the surrounding liver was categorized as iso- in both phases. Contrast-Enhanced Intraoperative Ultrasound The ultrasound system Prosound α10 (Aloka) was used for all fundamental intraoperative ultrasound and contrast-enhanced intraoperative ultrasound procedures. A microconvex probe dedicated to intraoperative ultrasound, UST-9132 (Aloka), with a 3.75-MHz transmit and receive center frequency, was used for all baseline fundamental intraoperative ultrasound examinations. A microconvex probe dedicated to intraoperative contrast-enhanced AJR:196, June 2011 1315

Arita et al. A Fig. 1 55-year-old man with liver cirrhosis who had moderately differentiated hepatocellular carcinoma in segment V of liver, 1.8 cm in diameter. Left side of ultrasound images shows referential fundamental view and right side shows harmonic view. A, Intraoperative contrast-enhanced ultrasound image obtained during vascular phase shows focal liver lesion strongly enhanced at 40 seconds after contrast injection (arrows). B, Contrast-enhanced intraoperative ultrasound during Kupffer phase approximately 16 minutes after contrast injection shows well-delineated hypoechoic focal liver lesion (arrowheads). Fig. 2 66-year-old man with chronic hepatitis who had hepatocellular carcinoma in segment VI of liver, 6.5 cm in diameter. Left side shows fundamental view and right side shows harmonic view. During vascular phase, unpaired arteries (arrows) are seen in tumor with margins indicated by arrowheads. ultrasound, UST-9133 (Aloka), was used for all contrast-enhanced intraoperative ultrasound procedures; the transmit center frequency of this probe was 1.88 MHz and the receive center frequency was 3.75 MHz. Contrast-enhanced intraoperative ultrasound was performed in the Extended Pure Harmonic Detection Mode (Aloka), a harmonic mode dedicated to microbubble detection. The frame rate was 63 Hz in the fundamental mode and 15 Hz in the harmonic mode. The mechanical index was set at 0.2 in all contrast-enhanced intraoperative ultrasound procedures. The acoustic power was automatically altered to keep the mechanical index at 0.2 as the depth of the focus point altered. During all contrast-enhanced intraoperative ultrasound procedures, a real-time fundamental mode image was displayed simultaneously with a harmonic mode image side-by-side for reference (Dual Dynamic Monitor Mode, Aloka) (Figs. 1 and 2). Thus, the target focal liver lesions were not missed even if they were hard to recognize in the harmonic mode. The gain and time-gain controls were set at the optimal settings before the contrast injection and were not altered thereafter. In the harmonic mode, an echo signal from within 1 cm of the probe was hard to recognize, so a 1-cm-thick sterilized gelatin sheet (Sonar Aid, Geistlich Pharma) was placed on the liver surface as an ultrasonic medium. All ultrasound procedures during operations were performed by one of four physicians who each had more than 5 years of experience performing intraoperative ultrasound examinations. After a laparotomy and liver mobilization, fundamental intraoperative ultrasound was performed to confirm the preoperative tumor staging. Then, Sonazoid was injected at a dose of 0.12 μl of microbubbles/kg of body weight through the antecubital vein in about 1 second. After injection, one focal liver lesion per injection was continuously observed in the harmonic mode for 1 minute (vascular phase), and the tumor vascularity was categorized as hypervascular, isovascular, or hypovascular. A focal liver lesion was categorized as hypervascular when all or a part of it turned hyperechoic compared with the surrounding liver parenchyma (Fig. 1A) or when unpaired dysmorphic vessels were observed within it [32, 33] (Fig. 2). Isovascular and hypovascular were defined as when a focal liver lesion was isoechoic and hypoechoic, respectively, at its maximal enhancement compared with the surrounding liver parenchyma. The focus was set at the level just below the approximate bottom of the target focal liver lesion. The ultrasonic observation was not extended beyond 1 minute after injection. One reason is that as microbubbles accumulated in the Kupffer cells the surrounding liver obviously increased its echogenicity in approximately 1 minute after injection, thus interfering with an assessment of tumor vascularity. Another reason is as follows: The acoustic power is rather high at the mechanical index of 0.2, and the same plane is continuously insonated for 1 minute. Therefore, microbubbles in the surrounding liver as well as the targeted focal liver lesion start to be destroyed, which interferes with a comparison of echogenicity between them. After an approximately 15-minute delay after the vascular phase, ultrasonic observation was commenced again in the harmonic mode to examine the entire liver; this period was defined as the Kupffer phase. During the pause after the vascular phase, surgical procedures that did not influence later contrast-enhanced intraoperative ultrasound were allowed. Hypoechoic focal liver lesions were searched for in the hyperechoic surrounding liver with accumulated microbubbles (Fig. 1B). Repeated scans were available; the contrast effect has been reported to be maintained at least 30 minutes after injection [34]. The focus point was set at the bottom of the liver. One additional injection of Sonazoid at the same dose was allowed to assess the vascularity of another focal liver lesion after surveillance in the Kupffer phase. Therefore, a maximum of two focal liver B 1316 AJR:196, June 2011

Vascular phase examined (n = 234) Detected in preoperative imaging (n = 329) Kupffer phase examined (n = 343) Resection (n = 352) Study subject (n = 374) Vascular phase not examined (n = 109) lesions per patient was examined in the vascular phase. The observation in the Kupffer phase was not repeated after the second injection because all focal liver lesions were theoretically examined by a thorough scan in the first Kupffer phase. A flowchart of all subject tumors, which summarizes examination procedures, is shown in Figure 3. Not all lesions examined during the vascular phase were included in the study subjects because some were not resected and others were resected but histologically diagnosed as benign. The 22 tumors found in resected specimens during histologic examination were considered isoechoic during the Found during Kupffer phase (n = 9) Vascular phase examined (n = 5) Found during intraoperative ultrasound (n = 14) Vascular phase not examined (n = 4) Found in resected specimen (n = 22) Fig. 3 Flow chart summarizes examination profile of all subject tumors. In total, 239 tumors were examined during vascular phase of contrast-enhanced intraoperative ultrasound. During Kupffer phase, all possible focal liver lesions were theoretically examined (n = 374). Tumors found in resected specimen (n = 22) were categorized as isoechoic. Kupffer phase because they had not been differentiated from the surrounding liver. For precise localization of a focal liver lesion, relation with up to the fourth order portal branch, i.e., next branching of Couinaud segment, was described, or the liver capsule was marked with electrocautery referring to intraoperative ultrasound. All contrast-enhanced intraoperative ultrasound categorization was performed prospectively during the operation in consensus between two of the four physicians. Tumor characterization by an offsite observer was not performed in this study because the surgical strategies were decided partially referring to the categorization. Statistics All statistical analyses were performed with the help of the commercially available software (PASW Statistics 17.0.2, SPSS) for Microsoft Windows. The Fisher exact test was applied for comparing proportions. Statistical significance was assessed using a two-tailed test at p < 0.05. Results A total of 407 histologically proven HCCs were detected among all 239 enrolled patients. Because 33 extensively necrotic HCCs were excluded, the subjects of this study included 374 HCCs, which consisted of 71 well-differentiated, 239 moderately differentiated, and 64 poorly differentiated HCCs. Ultrasonic observation during the vascular phase was done in all 239 patients; vascularity of one or two focal liver lesions, whether benign or malignant, was examined in each patient. Among those lesions for which vascularity was examined, 239 were included in the subjects of this study, which corresponded to 64% of all 374 HCCs. Of the 239 lesions, 187 were observed after the first injection of Sonazoid and 52 after the second injection. During the Kupffer phase, all 374 HCCs were theoretically examined because the entire liver was explored through repeated scans. Findings of contrast-enhanced intraoperative ultrasound among HCCs classified by histologic grade are shown in Table 2. Of the 239 HCCs examined during the vascular phase of contrast-enhanced intraoperative ultrasound, 186 (78%) were hy- TABLE 2: Findings of Contrast-Enhanced Intraoperative Ultrasound Among Hepatocellular Carcinoma (HCC) Classified by Histologic Grade Contrast-Enhanced Intraoperative Ultrasound Findings Hypervascular Well-Differentiated HCC Moderately Differentiated HCC Poorly Differentiated HCC (n = 64) Total (n = 374) Kupffer hypoechoic 22 120 36 178 Kupffer isoechoic 3 4 1 8 Isovascular Kupffer hypoechoic 5 13 5 23 Kupffer isoechoic 7 4 0 11 Hypovascular Kupffer hypoechoic 1 15 3 19 Kupffer isoechoic 0 0 0 0 Vascular not examined Kupffer hypoechoic 10 73 15 98 Kupffer isoechoic 23 10 4 37 AJR:196, June 2011 1317

Arita et al. 250 Not examined Hypovascular Isovascular Hypervascular 250 Isoechoic Hypoechoic 18 No. 150 100 50 0 33 1 5 12 25 Well 124 pervascular, 34 (14%) were isovascular, and 19 (8%) were hypovascular (Table 2). The vascular phase findings of each histologic grade are summarized in Figure 4. The proportion of hypervascular tumors among well-differentiated HCCs was 66% (25/38) and tended to be lower than that of 80% (161/201) among moderately and poorly differentiated HCCs, although the difference failed to reach statistical significance (p = 0.058, Fisher exact test), whereas the proportion of hypervascular tumors among moderately differentiated HCCs was 79% (124/156) and was similar to that of 82% (37/45) among poorly differentiated HCCs (p = 0.83, Fisher s exact test). During the Kupffer phase of contrast-enhanced intraoperative ultrasound, 318 (85%) of all 374 HCCs were hypoechoic (Table 2). The Kupffer phase findings among each histologic grade are summarized in Figure 5. The proportion of hypoechoic tumors among well-differentiated HCCs was 54% (38/71) and significantly 83 15 17 Moderately Differentiated 19 37 Poorly (n = 64) Fig. 4 Graph shows proportions of tumors with each finding during vascular phase of contrast-enhanced ultrasound among each histologic grade of hepatocellular carcinoma (HCC). Hypervascular tumors were significantly less frequent among well-differentiated than moderately and poorly differentiated HCC. 3 No. 150 100 lower than that of 92% (280/ 303) among moderately and poorly differentiated HCCs (p < 0.0001, Fisher exact test), whereas the proportion of hypoechoic tumors among moderately differentiated HCCs was 92% (221/239) and similar to that of 92% (59/64) among poorly differentiated HCCs (p = 1.00, Fisher exact test). Because 10 patients who had an allergy to iodinated contrast agents were excluded, 229 patients underwent dynamic CT. Dynamic CT findings during the early phase are summarized in Figure 6. The proportion of hypervascular tumors among well-differentiated HCCs was 51% (36/71) and significantly lower than that of 87% (253/291) among moderately and poorly differentiated HCCs (p < 0.0001). Dynamic CT findings during the late phase are summarized in Figure 7. The proportion of hypodense tumors among well-differentiated HCCs was 59% (42/71) and significantly lower than that of 85% (247/291) among moderately and poorly differentiated HCCs (p < 0.0001, Fisher exact test). 50 0 33 38 Well 221 Moderately Differentiated Of the 239 HCCs that were examined in both phases of contrast-enhanced intraoperative ultrasound, 11 were neither hypervascular in the vascular phase nor hypoechoic in the Kupffer phase, i.e., diagnostic findings for HCC were absent during both phases. The characteristics of the 11 tumors are summarized in Table 3. The median size of the 11 tumors was 1.5 cm. Among the 11 tumors, seven (64%) were well-differentiated HCC and the other four tumors were well- or moderately differentiated HCC, of which a major component was well-differentiated tumors. The classic pattern for HCC in dynamic CT, i.e., concomitantly hypervascular during early phase and hypodense during late phase, was seen only in three of 11 tumors. Discussion This study showed that findings during both the vascular and Kupffer phases of contrast-enhanced intraoperative ultrasound using Sonazoid are correlated with the histologic grade of HCC. In particular, such correlation was 5 59 Poorly (n = 64) Fig. 5 Graph shows proportions of tumors with each finding during Kupffer phase of contrast-enhanced ultrasound among each histologic grade of hepatocellular carcinoma (HCC). Hypoechoic tumors were significantly less frequent among well-differentiated than moderately and poorly differentiated HCC. 1318 AJR:196, June 2011

No. 250 150 100 50 0 Not examined Hypovascular Isovascular Hypervascular 11 24 36 Well 12 8 19 Moderately Differentiated found stronger for the Kupffer phase. Moderately and poorly differentiated HCCs more frequently showed hypervascular findings during the vascular phase and hypoechoic findings during the Kupffer phase than welldifferentiated HCCs did. On the other hand, moderately and poorly differentiated HCCs showed similar findings during both the vascular and Kupffer phases of contrast-enhanced intraoperative ultrasound. These results could be applicable to contrast-enhanced ultrasound in clinical settings. Estimating the histologic grade is helpful for clinical management of patients with HCC. Well-differentiated HCC less frequently induces intrahepatic metastasis through vascular invasion than do moderately and poorly differentiated HCC [5 8], thus surgeons could select a nonanatomic resection when a tumor is estimated to be well differentiated, and conversely, an anatomic resection would be selected to eradicate possible intrahepatic metastasis. For a similar reason, when a patient has solitary HCC with a marginal size, 9 53 Poorly (n = 64) Fig. 6 Graph shows proportions of tumors with each finding during early phase of dynamic CT among each histologic grade of hepatocellular carcinoma (HCC). Hypervascular tumors were significantly less frequent among well-differentiated than moderately and poorly differentiated HCC. 2 No. 250 150 100 Not examined Hyperdense Isodense Hypodense a physician could confidently select either hepatic resection or local ablation using the information regarding histologic grade. A hypervascular finding during the vascular phase of contrast-enhanced intraoperative ultrasound was more frequent among moderately and poorly differentiated HCCs (80%) than among well-differentiated HCCs (66%) in this study, although the difference barely failed to reach statistical significance, perhaps because of the small sample size. A study of contrast-enhanced ultrasound using Definity reported similar findings (well-differentiated, 61%; moderately and poorly differentiated, 94%) [28], whereas three studies using SonoVue reported that the proportion of hypervascular findings among well-differentiated HCCs (82 100%) was higher than that in this study, and the proportion was not significantly different from moderately and poorly differentiated HCCs (96 100%) [10, 29, 35]. Such a discrepancy may be caused partly by the difference of the tumor size. The median size of the subject tumors in the 50 0 29 42 Well 12 35 191 Moderately Differentiated current study was 1.5 cm and rather smaller compared with those of the three studies using SonoVue, which were 2.7 cm [10], 5.88 cm [29], and 3.8 cm [35]. Other possible causes included the difference of contrast medium or type of harmonic mode; however, an explanatory mechanism is yet to be found. The hypervascular findings of dynamic CT were also more frequent among moderately and poorly differentiated HCCs than among well-differentiated HCCs. This result was similar to the reports that claimed early phase findings of dynamic CT are well correlated with histologic grade [36 39]. This finding would indicate that the results of the study subjects of the current study would be worth comparing with other studies, although the median diameter, about 1.8 cm, was relatively small. A hypoechoic finding during the Kupffer phase of contrast-enhanced intraoperative ultrasound was more frequent among moderately and poorly differentiated HCCs than among well-differentiated HCCs in this study. Similar 1 8 56 Poorly (n = 64) Fig. 7 Graph shows proportions of tumors with each finding during late phase of dynamic CT among each histologic grade of hepatocellular carcinoma (HCC). Hypodense tumors were significantly less frequent among well-differentiated than moderately and poorly differentiated HCC. AJR:196, June 2011 1319

Arita et al. TABLE 3: Hepatocellular Carcinoma (HCC) Without Diagnostic Findings During Both Phases of Contrast-Enhanced Intraoperative Ultrasound Tumor Liver Size (cm) Depth (cm) CT Early CT Late Histologic Grade 1 Liver cirrhosis 2.7 0.0 Hypointense Hypointense Well-differentiated 2 a Chronic hepatitis 1.6 3.0 Isointense Isointense Well-differentiated 3 a Chronic hepatitis 1.0 1.0 Isointense Isointense Well-differentiated 4 a Chronic hepatitis 1.0 2.0 Hypointense Hypointense Well-differentiated 5 a Chronic hepatitis 0.5 2.0 Hypointense Hypointense Well-differentiated 6 Liver cirrhosis 1.5 0.2 Hypointense Hypointense Well-differentiated 7 Chronic hepatitis 2.0 1.0 Hypointense Hypointense Well to moderately differentiated 8 Liver cirrhosis 1.3 2.0 Hyperintense Hypointense Well to moderately differentiated 9 Liver cirrhosis 1.2 1.5 Hyperintense Isointense Well-differentiated 10 Chronic hepatitis 1.0 2.0 Hyperintense Hypointense Well to moderately differentiated 11 Chronic hepatitis 1.3 2.5 Hyperintense Hypointense Well to moderately differentiated a Tumors 2 and 3 were found in the same patient, and tumors 4 and 5 were found in the same patient. results were reported by Korenaga et al. [21], who assessed Kupffer phase images of contrast-enhanced ultrasound using Sonazoid for 48 histologically proven HCCs. The echogenicity ratio of tumor against liver parenchyma was calculated in the study; the mean ratio for well-differentiated HCCs was 0.84 and was significantly higher than that for moderately differentiated HCCs (0.43) and poorly differentiated HCCs (0.30). The current study has corroborated the difference of findings during the Kupffer phase between well-differentiated HCCs and moderately and poorly differentiated HCCs, with a larger number of tumors, including as many as 71 well-differentiated HCCs and exclusively using resected specimens for histologic grading. The reason for the discrepancy among histologic grades may be that well-differentiated HCCs possess a comparable number of Kupffer cells to liver parenchyma, whereas moderately and poorly differentiated HCCs possess fewer or none [17, 18, 30]. In this study, a hypodense finding during the late phase of dynamic CT was significantly more frequent among moderately and poorly differentiated HCCs than among welldifferentiated HCCs, whereas the proportions were similar between moderately differentiated HCCs and poorly differentiated HCCs. Two studies reported similar results [37, 39] and another reported that proportions of hypodense tumors were equivalently high among three histologic grades [38]. This discrepancy may be explained by the difference of scan timing; the current study and the former two adopted 120 180 seconds for the interval, whereas the latter study used 300 seconds. Therefore, the results of the current study seemed acceptable. Well-differentiated HCCs were dominant among the 11 HCCs that were neither hypervascular in the vascular phase or hypoechoic in the Kupffer phase (Table 2). As many as nine of 39 well-differentiated HCCs (32%) examined in both phases corresponded to such cases. The median size of the 11 tumors was 1.3 cm and similar to that of all tumors (1.5 cm). Therefore, the lack of diagnostic findings could not be attributed to the small size of tumors. Given that most of the 11 tumors also lacked the classic pattern in dynamic CT, such tumors would be at an early stage of the cellular dedifferentiation pathway. One of the drawbacks of this study is that all HCCs were not examined during the vascular phase of contrast-enhanced intraoperative ultrasound. This stems from the limitation of administering Sonazoid; all focal liver lesions could not be examined during the vascular phase when a patient had three or more focal liver lesions. More tumors can be examined during the vascular phase of contrast-enhanced intraoperative ultrasound if the dose of Sonazoid used in one injection is reduced, although this aspect necessitates further validation. Nonetheless, the assessment of the diagnostic ability of this modality seemed appropriate because the differences of proportion of tumors reached statistical significance regarding the findings during the vascular phase. Another drawback may be the specific setting of the study. Only patients who were scheduled to undergo surgery for HCC were the subjects. Thus, the results concerning findings in both the vascular and Kupffer phase may not be applicable to all patients with focal liver lesions. In conclusion, contrast-enhanced intraoperative ultrasound using Sonazoid is useful for estimating the histologic grade of HCC, and, in particular, findings during the Kupffer phase strongly correlate with histologic grade. References 1. Parkin DM, Bray F, Ferlay J, Pisani P. Estimating the world cancer burden: Globocan 0. Int J Cancer 1; 94:153 156 2. Ikeguchi M, Sato N, Hirooka Y, Kaibara N. Computerized nuclear morphometry of hepatocellular carcinoma and its relation to proliferative activity. J Surg Oncol 1998; 68:225 230 3. Shimada M, Takenaka K, Taguchi K, et al. Prognostic factors after repeat hepatectomy for recurrent hepatocellular carcinoma. Ann Surg 1998; 227:80 85 4. Wayne JD, Lauwers GY, Ikai I, et al. Preoperative predictors of survival after resection of small hepatocellular carcinomas. Ann Surg 2; 235:722 730 5. Sato M, Watanabe Y, Lee T, et al. Well-differentiated hepatocellular carcinoma: clinicopathological features and results of hepatic resection. Am J Gastroenterol 1995; 90:112 116 6. Takayama T, Makuuchi M, Hirohashi S, et al. Early hepatocellular carcinoma as an entity with a high rate of surgical cure. Hepatology 1998; 28:1241 1246 7. Esnaola NF, Lauwers GY, Mirza NQ, et al. Predictors of microvascular invasion in patients with hepatocellular carcinoma who are candidates for orthotopic liver transplantation. J Gastrointest Surg 2; 6:224 232 8. Kim BK, Han KH, Park YN, et al. Prediction of microvascular invasion before curative resection of hepatocellular carcinoma. J Surg Oncol 8; 97:246 252 9. Gaiani S, Celli N, Piscaglia F, et al. Usefulness of contrast-enhanced perfusional sonography in the assessment of hepatocellular carcinoma hypervascular at spiral computed tomography. J Hepatol 4; 41:421 426 1320 AJR:196, June 2011

10. Nicolau C, Catala V, Vilana R, et al. Evaluation of hepatocellular carcinoma using SonoVue, a second generation ultrasound contrast agent: correlation with cellular differentiation. Eur Radiol 4; 14:1092 1099 11. Quaia E, Calliada F, Bertolotto M, et al. Characterization of focal liver lesions with contrast-specific US modes and a sulfur hexafluoride-filled microbubble contrast agent: diagnostic performance and confidence. Radiology 4; 232:420 430 12. Catala V, Nicolau C, Vilana R, et al. Characterization of focal liver lesions: comparative study of contrast-enhanced ultrasound versus spiral computed tomography. Eur Radiol 7; 17:1066 1073 13. Yanagisawa K, Moriyasu F, Miyahara T, Yuki M, Iijima H. Phagocytosis of ultrasound contrast agent microbubbles by Kupffer cells. Ultrasound Med Biol 7; 33:318 325 14. Kindberg GM, Tolleshaug H, Roos N, Skotland T. Hepatic clearance of Sonazoid perfluorobutane microbubbles by Kupffer cells does not reduce the ability of liver to phagocytose or degrade albumin microspheres. Cell Tissue Res 3; 312:49 54 15. Watanabe R, Matsumura M, Munemasa T, Fujimaki M, Suematsu M. Mechanism of hepatic parenchyma-specific contrast of microbubble-based contrast agent for ultrasonography: microscopic studies in rat liver. Invest Radiol 7; 42:643 651 16. Hatanaka K, Kudo M, Minami Y, et al. Differential diagnosis of hepatic tumors: value of contrastenhanced harmonic sonography using the newly developed contrast agent, Sonazoid. Intervirology 8; 51[suppl 1]:61 69 17. Imai Y, Murakami T, Yoshida S, et al. Superparamagnetic iron oxide-enhanced magnetic resonance images of hepatocellular carcinoma: correlation with histological grading. Hepatology 0; 32:205 212 18. Lim JH, Choi D, Cho SK, et al. Conspicuity of hepatocellular nodular lesions in cirrhotic livers at ferumoxides-enhanced MR imaging: importance of Kupffer cell number. Radiology 1; 220:669 676 19. Kitamura H, Kawasaki S, Nakajima K, Ota H. Correlation between microbubble contrast-enhanced color Doppler sonography and immunostaining for Kupffer cells in assessing the histopathologic grade of hepatocellular carcinoma: preliminary results. J Clin Ultrasound 2; 30: 465 471 20. Inoue T, Kudo M, Hatanaka K, et al. Imaging of hepatocellular carcinoma: qualitative and quantitative analysis of postvascular phase contrast-enhanced ultrasonography with Sonazoid comparison with superparamagnetic iron oxide magnetic resonance images. Oncology 8; 75[suppl 1]:48 54 21. Korenaga K, Korenaga M, Furukawa M, Yamasaki T, Sakaida I. Usefulness of Sonazoid contrastenhanced ultrasonography for hepatocellular carcinoma: comparison with pathological diagnosis and superparamagnetic iron oxide magnetic resonance images. J Gastroenterol 9; 44:733 741 22. Bru C, Maroto A, Bruix J, et al. Diagnostic accuracy of fine-needle aspiration biopsy in patients with hepatocellular carcinoma. Dig Dis Sci 1989; 34:1765 1769 23. Caturelli E, Solmi L, Anti M, et al. Ultrasound guided fine needle biopsy of early hepatocellular carcinoma complicating liver cirrhosis: a multicentre study. Gut 4; 53:1356 1362 24. Forner A, Vilana R, Ayuso C, et al. Diagnosis of hepatic nodules 20 mm or smaller in cirrhosis: prospective validation of the noninvasive diagnostic criteria for hepatocellular carcinoma. Hepatology 8; 47:97 104 25. Pawlik TM, Gleisner AL, Anders RA, Assumpcao L, Maley W, Choti MA. Preoperative assessment of hepatocellular carcinoma tumor grade using needle biopsy: implications for transplant eligibility. Ann Surg 7; 245:435 442 26. [No authors listed]. Terminology of nodular hepatocellular lesions: International Working Party. Hepatology 1995; 22:983 993 27. Kokudo N, Bandai Y, Imanishi H, et al. Management of new hepatic nodules detected by intraoperative ultrasonography during hepatic resection for hepatocellular carcinoma. Surgery 1996; 119:634 640 28. Jang HJ, Kim TK, Burns PN, Wilson SR. Enhancement patterns of hepatocellular carcinoma at contrast-enhanced US: comparison with histologic differentiation. Radiology 7; 244:898 906 29. Liu GJ, Xu HX, Lu MD, et al. Correlation between enhancement pattern of hepatocellular carcinoma on real-time contrast-enhanced ultrasound and tumour cellular differentiation on histopathology. Br J Radiol 7; 80:321 330 30. Tanaka M, Nakashima O, Wada Y, Kage M, Kojiro M. Pathomorphological study of Kupffer cells in hepatocellular carcinoma and hyperplastic nodular lesions in the liver. Hepatology 1996; 24:807 812 31. Inoue T, Kudo M, Watai R, et al. Differential diagnosis of nodular lesions in cirrhotic liver by post-vascular phase contrast-enhanced US with Levovist: comparison with superparamagnetic iron oxide magnetic resonance images. J Gastroenterol 5; 40:1139 1147 32. Torzilli G, Palmisano A, Del Fabbro D, et al. Contrast-enhanced intraoperative ultrasonography during surgery for hepatocellular carcinoma in liver cirrhosis: is it useful or useless? A prospective cohort study of our experience. Ann Surg Oncol 7; 14:1347 1355 33. Yang H, Liu GJ, Lu MD, Xu HX, Xie XY. Evaluation of the vascular architecture of hepatocellular carcinoma by micro flow imaging: pathologic correlation. J Ultrasound Med 7; 26:461 467 34. Watanabe R, Matsumura M, Chen CJ, Kaneda Y, Ishihara M, Fujimaki M. Gray-scale liver enhancement with Sonazoid (NC100100), a novel ultrasound contrast agent detection of hepatic tumors in a rabbit model. Biol Pharm Bull 3; 26:1272 1277 35. Fan ZH, Chen MH, Dai Y, et al. Evaluation of primary malignancies of the liver using contrast-enhanced sonography: correlation with pathology. AJR 6; 186:1512 1519 36. Amano S, Ebara M, Yajima T, et al. Assessment of cancer cell differentiation in small hepatocellular carcinoma by computed tomography and magnetic resonance imaging. J Gastroenterol Hepatol 3; 18:273 279 37. Sakabe K, Yamamoto T, Kubo S, et al. Correlation between dynamic computed tomographic and histopathological findings in the diagnosis of small hepatocellular carcinoma. Dig Surg 4; 21:413 420 38. Monzawa S, Ichikawa T, Nakajima H, Kitanaka Y, Omata K, Araki T. Dynamic CT for detecting small hepatocellular carcinoma: usefulness of delayed phase imaging. AJR 7; 188:147 153 39. Yoon SH, Lee JM, So YH, et al. Multiphasic MDCT enhancement pattern of hepatocellular carcinoma smaller than 3 cm in diameter: tumor size and cellular differentiation. AJR 9; 193: 1581; [web]w482 W489 AJR:196, June 2011 1321