CT Evaluation of the Progression of Hypoattenuating Nodular Lesions in Virus-Related Chronic Liver Disease

Takayasu et al. CT in Liver Disease Hepatobiliary Imaging Original Research A C M E D E N T U R I C A L I M A G I N G Kenichi Takayasu 1 Yukio Muramatsu 2 Yasunori Mizuguchi 1 Takuji Okusaka 3 Kazuaki Shimada 4 Tadatoshi Takayama 4,5 Michiie Sakamoto 6,7 AJR 2006; 187:454 463 0361 803X/06/1872 454 American Roentgen Ray Society Y O F Takayasu K, Muramatsu Y, Mizuguchi Y, et al. Keywords: dynamic CT, hepatocarcinogenesis, liver disease, oncologic imaging DOI:10.2214/AJR.05.0705 Received April 25, 2005; accepted after revision June 24, 2005. 1 Department of Diagnostic Radiology, National Cancer Center Hospital, 5-1-1, Tsukiji, Chuo-ku, Tokyo 104-0045, Japan. Address correspondence to K. Takayasu (ktakayas@ncc.go.jp). 2 Cancer Screening Division, Research Center for Cancer Prevention and Screening, National Cancer Center, Tsukiji, Chuo-ku, Tokyo 104-0045, Japan. 3 Division of Hepatobiliary and Pancreatic Oncology, National Cancer Center Hospital, Tokyo 104-0045, Japan. 4 Division of Hepatobiliary Surgery, National Cancer Center Hospital, Tokyo 104-0045, Japan. 5 Present address: Third Department of Surgery, Nihon University School of Medicine, Tokyo, Japan. 6 Division of Pathology, National Cancer Center Research Institute, Tokyo 104-0045, Japan. 7 Present address: Department of Pathology, Keio University School of Medicine, Tokyo 160-8582, Japan. CT Evaluation of the Progression of Hypoattenuating Nodular Lesions in Virus-Related Chronic Liver Disease OBJECTIVE. The purpose of this study was to clarify the natural outcomes of hypoattenuating nodular lesions in patients with virus-related chronic liver disease depicted on dynamic CT. MATERIALS AND METHODS. Sixty lesions (mean size, 1.3 cm) exhibiting hypoattenuation or isoattenuation in the arterial and delayed phases of dynamic CT were retrospectively evaluated with additional CT (mean, six examinations) for a mean period of 838 days. The primary end point was emergence of hyperattenuating areas within hypoattenuating lesions, a phenomenon called attenuation conversion. Cumulative attenuation conversion rates suggesting rates of malignant transformation were calculated with the Kaplan-Meier method, and factors affecting attenuation conversion rate were analyzed with the Cox proportional hazard model. RESULTS. Thirty-six (60%) of 60 hypoattenuating lesions developed to hyperattenuating lesions, 21 were unchanged, and three disappeared spontaneously. The 36 lesions that became hyperattenuating were divided into two subgroups according to lesion enhancement pattern: hyper-in-hypoattenuating (n = 25) and entirely hyperattenuating (n = 11). The cumulative attenuation conversion rates for the 60 hypoattenuating lesions were 15.8%, 44.3%, and 58.7% at 1, 2, and 3 years. The hyper-in-hypoattenuating lesions showed more rapid progression to entirely enhanced lesions. Positive results for hepatitis C viral antibody (p = 0.028) and initial lesion size (p = 0.007) showed a positive correlation with attenuation conversion rate. CONCLUSION. Hypoattenuating hepatic nodular lesions in chronic liver disease depicted on dynamic CT have high malignant potential and should be followed with special attention to conversion from hypoattenuation to hyperattenuation to determine the optimal timing of treatment. epatocellular carcinoma (HCC) is H the fifth most common cancer in the world and the third leading cause of cancer-related death [1]. HCC remains one of the malignancies with the poorest prognosis because of the advanced stage of the cancer, associated liver cirrhosis, and a high recurrence rate after treatment. Early diagnosis of HCC is imperative to enable patients to undergo curative therapy, such as surgical resection [2], percutaneous ethanol injection (PEI) [3], radiofrequency ablation [4], or liver transplantation [5]. The introduction of advanced imaging techniques, especially helical CT, has made it possible to detect small HCCs and to differentiate HCC from other lesions with relative ease. The characteristic enhancement pattern of HCC includes hyperattenuation in the arterial phase and hypoattenuation or isoattenuation in the delayed or equilibrium phase of dynamic CT [6]. In our daily practice, however, we frequently encounter small hypoattenuating or unenhancing nodular lesions with hypoattenuation or isoattenuation in the arterial and delayed phases of dynamic CT [7 10]. Until the mid 1990s in Japan, most hypoattenuating nodular lesions of chronic liver disease were aggressively treated with surgical resection [11] because of the malignant potential for development to HCC through multistep progression of hepatocarcinogenesis [12]. These lesions develop through stages from adenomatous hyperplasia (corresponding to low-grade dysplastic nodule proposed by the International Working Party [13, 14]) through atypical adenomatous hyperplasia, early HCC, and nodule-in-nodule HCC to, finally, overt HCC [15, 16], even though de novo developments are presumed as other pathways leading to hepatocarcinogenesis. Resected hypoattenuating nodular lesions are histopathologically graded as early HCC, adenomatous hyperplasia, or atypical adenomatous hyperplasia [16]. At our hospital, early HCC accounted for 14% of 980 surgically 454 AJR:187, August 2006

CT in Liver Disease TABLE 1: Characteristics of 53 Patients with 60 Hypoattenuating Lesions and Results of Univariate Analysis of Factors Affecting Attenuation Conversion Variable Value p Mean age (y) 65 < 65 vs 65 0.061 Sex 0.725 Male 37 Female 16 Hepatitis B surface antigen 0.074 Positive 6 Negative 47 Hepatitis C virus antibody 0.012 Positive 45 Negative 8 Child-Pugh classification 0.208 Grade A 48 Grade B 5 Grade C 0 Mean α-fetoprotein value (ng/ml) 48 20 vs > 20 0.017 History of hepatic surgery for HCC 0.606 Yes 25 No 28 No. of patients with hypoattenuating lesions 0.559 1 lesion 47 2 lesions 5 3 lesions 1 Mean tumor size (cm) 1.3 1.0 vs > 1.0 0.002 CT attenuation pattern in arterial and delayed phases (no. of lesions) 0.025 Hypoattenuation/hypoattenuation 44 Isoattenuation/hypoattenuation 15 Hypoattenuation/isoattenuation 1 Note Values are numbers of patients unless otherwise indicated. resected HCC nodules in 664 patients [17]. We have encountered hypoattenuating nodular lesions that have developed to partial hyperattenuation within hypoattenuating lesions (hyper-in-hypoattenuating type). These nodules correlate histopathologically to the nodulein-nodule type of HCC (formerly called early advanced HCC) [18] or to entirely hyperattenuating lesions (entirely enhanced type), corresponding to overt HCC on follow-up CT. In the management of hypoattenuating lesions, PEI and radiofrequency ablation have replaced surgery because these procedures are less invasive than surgical treatment [19]. Questions have arisen, however. The first is whether local ablation therapy for hypoattenuating lesions is indispensable for prolonging the survival of patients with chronic liver disease, because lifethreatening overt HCC emerges frequently in an area of the liver different from that of the lesion being followed. The second is whether ablation therapies cause deterioration of the residual liver and complications such as needle track seeding [20] and radiofrequency ablation related death [21, 22]. The critical time for management of hypoattenuating lesions is unknown. Materials and Methods For 8.5 years beginning in April 1993, the records of 129 consecutive patients with hypoattenuating nodular lesions diagnosed with helical CT were retrospectively selected from the CT database at our insti- tution. The aim of the study was to survey HCC in high-risk patients with positive results for hepatitis B virus, hepatitis C virus, or both; chronic liver disease; and a history of hepatectomy for HCC. The institutional review board at our institution did not require approval or informed consent for medical records or imaging examinations. Inclusion criteria were the presence of hypoattenuating lesions exhibiting as one of three CT attenuation patterns in the arterial and delayed or equilibrium phases of dynamic CT, that is, a hypo-hypoattenuating pattern relative to the surrounding liver parenchyma, a hypo-isoattenuating pattern, or an iso-hypoattenuating pattern. The number of hypoattenuating lesions was limited to no more than three, no association with overt HCC at initial CT, and follow-up period of more than 3 months after the initial CT examination. A total of 76 patients were excluded: 31 patients because they had undergone local ablation therapy with PEI or radiofrequency ablation within the initial follow-up period of 3 months; 23 patients because they had four or more lesions; 14 patients because they were lost to follow-up; and eight patients because they had active second primary cancers in other organs during the follow-up period. Fifty-three patients with 60 hypoattenuating lesions were included in the study (Table 1). The mean age of the patients was 65 years (range, 47 80 years), and the male to female ratio was 37:16. Six patients had positive results for hepatitis B virus surface antigen, 45 for hepatitis C virus antibody, and two for both viral markers. Four patients had negative results for both markers. According to the Child-Pugh classification, 48 patients had grade A disease, and five had grade B disease. Mean α-fetoprotein level (normal value, < 20 ng/ml) was 48 ng/ml (range, 3 299 ng/ml). Twenty-five patients had a history of partial hepatectomy for HCC a mean of 1,030 days (range, 121 3,299 days) before the initial CT examination, and 28 (52.8%) of the patients had undergone no previous therapy for HCC. Fortyseven patients had one lesion, five had two lesions, and one patient had three lesions. Follow-up helical CT was performed at least every 6 months (mean, 6 times; range, 1 23 times). The primary end point was the time when attenuation conversion from hypoattentuation to hyperattenuation was recognized. This point was emergence of an enhanced area within a hypoattenuating lesion, that is, a hyper-in-hypoattenuating type corresponding to nodule-in-nodule type HCC [18] or an entirely hyperattenuating type consistent with overt HCC. The primary end point for unchanged hypoattenuating lesions was the date of final follow-up CT examination. The mean interval between initial CT and end point was 838 days (range, 103 3,100 days). A lesion that appeared anew in a portion of the liver different from that of the lesion being followed was designated a recurrent lesion. AJR:187, August 2006 455

Takayasu et al. Examinations combining CT hepatic arteriography (CTHA) and CT arterial portography (CTAP) were performed on 18 patients with 19 hypoattenuating lesions before the emergence of attenuation conversion on follow-up CT. The mean interval between initial CT examination and combination study was 254 days A C Fig. 1 66-year-old woman with hypoattenuating lesion that progressed to hyper-in-hypoattenuating type (nodule-in-nodule hepatocellular carcinoma [HCC]) (group B1) and advanced to complete hyperattenuation (overt HCC) (group B2). A, Arterial phase CT scan shows 1.0-cm hypoattenuating lesion (arrow) in segment II of liver. B, Delayed phase of A shows hypoattenuating lesion (arrow). C, Follow-up arterial phase CT scan 2 years 8 months after A shows hyperattenuating area measuring 1.0 cm within hypoattenuating lesion (arrow) measuring 2.5 cm. D, Delayed phase of C shows entire shape as hypoattenuating lesion (arrow). (Fig. 1 continues on next page) B D 456 AJR:187, August 2006

CT in Liver Disease (range, 0 1,400 days). Fine-needle (21-gauge) biopsy under sonographic monitoring (Sonopsy, Hakko) was performed on 13 lesions (n = 13 patients) immediately after recognition of attenuation conversion. Biopsy also was performed on two other lesions that remained hypoattenuating, but definite diagnosis was not obtained because of sampling error. The mean interval between initial CT examination and needle biopsy was 696 days (range, 39 1,740 days). All 36 hypoattenuating lesions that became hyperattenuating were managed with various interventions. All patients underwent helical CT in two phases (arterial and delayed phases; TCT900S and X- vigor, Toshiba Medical Systems) or three phases (arterial, portal, and delayed phases; Aquilion MDCT, Toshiba). A dose of 120 ml of iopamidol (Iopamiron 300 mg I/mL, Schering) was injected into an antecubital vein at a rate of 3 ml/s. CT was started 35 40 seconds (arterial phase), 70 seconds (portal phase), and 3 minutes (delayed phase) after the start of injection of contrast medium. Entire livers were scanned within one breath-hold for approximately 8 20 seconds, depending on liver size. Scanning parameters were as follows: axial singleor four-slice mode, 5 10 mm beam collimation, 0.5 1.0 s/rotation, 0.7 3.0 of helical pitch (0.7 1.0 of pitch factor), 120 150 kvp, and 200 250 mas. Image reconstructions were 5 7 mm thick. E Fig. 1 (continued) 66-year-old woman with hypoattenuating lesion that progressed to hyper-in-hypoattenuating type (nodule-in-nodule hepatocellular carcinoma [HCC]) (group B1) and advanced to complete hyperattenuation (overt HCC) (group B2). E, Arterial phase CT scan 8 months after C shows 3.0-cm lesion (arrow) almost entirely occupied by hyperattenuating component. F, Delayed phase of E shows hypoattenuating component (arrow) with ring enhancement around lesion. CTAP was started 20 seconds after injection of 90 ml of ioversol (Optiray 350 mg I/mL, Yamanouchi) diluted with saline solution (1:3 ratio; iodine, 87.5 mg I/mL) at a speed of 3 ml/s into the superior mesenteric artery. CTHA was started 10 seconds after injection of 60 ml of ioversol into the proper, right, or left hepatic artery at a speed of 1.3 2 ml/s [8, 10]. Beam collimation and image reconstructions were 7 mm. Image Analysis Hard copies of dynamic CT scans obtained from the CT database were independently reviewed by two abdominal radiologists who had 18 and 21 years of experience, respectively. The location and size of a hypoattenuating lesion with no features to suggest cyst, hemangioma, or abscess according to Couinaud segmentation were recorded. Follow-up CT scans were studied with special attention to attenuation conversion: partial or total enhancement within a lesion. For hyper-in-hypoattenuating type lesions, the size of the internal hyperattenuating portion was measured. Lesions developing anew in a liver segment different from the initial one were considered recurrent (second primary) lesions, and the size and characteristics of those lesions were recorded. The diagnosis of overt HCC was based on the following features: hyperattenuation in the arterial phase and hypoattenuation or isoattenuation with or without ring enhancement in the delayed phase of dynamic CT. For CTHA and CTAP, intratumoral attenuation such as hyperattenuation, isoattenuation, and hypoattenuation within the corresponding lesion was recorded. To calculate the ratio of the chronologic change in tumor size at final CT examination to that at initial CT examination, the final diameter of the lesion was divided by the initial diameter. When the interpretations of the two radiologists differed, a third radiologist, who had 12 years of experience, joined the discussion until consensus was reached. To elucidate factors affecting attenuation transforming from hypoattenuation to hyperattenuation, the following 10 factors were analyzed with univariate analysis: age, sex, presence of hepatitis B virus surface antigen, presence of hepatitis C virus antibodies, Child-Pugh classification, α-fetoprotein level, previous hepatic surgery for HCC, number of hypoattenuating lesions in each patient, initial lesion size, and CT patterns of lesions in arterial and delayed phases (Table 1). Multivariate analysis was performed. Statistical Analysis Univariate analysis was performed with the Kaplan- Meier method, and the significant differences were evaluated with generalized Wilcoxon s and log-rank tests. F AJR:187, August 2006 457

Takayasu et al. Fig. 2 57-year-old man with hypoattenuating lesion and no change in attenuation even with increase in tumor size (group A, unchanged type). A, Arterial phase of initial CT scan shows simple cyst (arrowhead), but lesion is not visible. B, Delayed phase of A shows simple cyst (arrowhead) and subtle 1.0-cm hypoattenuating lesion (arrow) in segment IV of liver. C, Arterial phase of last CT scan 3 years 10 months after A shows hypoattenuating lesion (arrow) has grown to 2.5 cm. D, Delayed phase of C. (Fig. 2 continues on next page) A C B D Multivariate analysis was performed with the Cox proportional hazard model with a backward stepwise procedure for selection of covariates. Student s t tests were used to compare two subgroups of lesions. Values of p < 0.05 in both tails were considered significant differences. All analyses were done with SPSS 11.0 software. Results For the 60 hypoattenuating lesions, the patterns in the arterial and delayed phases on the 458 AJR:187, August 2006

CT in Liver Disease Fig. 2 (continued) 57-year-old man with hypoattenuating lesion and no change in attenuation even with increase in tumor size (group A, unchanged type). E, CT hepatic arteriogram obtained 1 day after C shows hypoattenuating lesion (arrow) that appeared isoattenuating on CT scan during arterial portography (not shown). Hypoattenuation (n = 21) (Group A, unchanged type) Hyper-in-hypo type (n = 25) (Group B1, nodule-in-nodule HCC) Entirely hyperattenuating type (n = 6) (Group B3, overt HCC) Hypoattenuation (n = 60 lesions) Hyperattenuation (n = 36) (Group B) Fig. 3 Outcome of 60 hypoattenuating nodular lesions in 53 patients on dynamic CT chronologically divided by different attenuation pattern. HCC = hepatocellular carcinoma. initial dynamic CT scans were as follows: hypo-hypoattenuating pattern (Figs. 1A and 1B), 44 lesions; iso-hypoattenuating pattern (Figs. 2A and 2B), 15 lesions; and hypo-isoattenuating pattern, one lesion (Table 1). The portal phase CT scans of 12 lesions showed isoattenuation and hypoattenuation in six lesions each. On CTHA and CTAP of 19 hypoattenuating lesions, nine lesions were hypoisoattenuating (Fig. 2E); eight, hypo-hypoattenuating; and two, iso-hypoattenuating. Isoattenuation (n = 3) (Group C, spontaneous regression) Entirely hyperattenuating type (n = 11) (Group B2, overt HCC) On the basis of chronologic changes in lesion attenuation, the 60 hypoattenuating lesions in 53 patients were categorized into one of three groups (Fig. 3). Group A was composed of 21 (35%) of 60 hypoattenuating lesions that did not change (Fig. 2); group B, of 36 (60%) hypoattenuating lesions that became hyperattenuating; and group C, of three isovascular lesions that regressed spontaneously. The 36 hyperattenuating lesions in group B were divided into two subgroups according to enhancement pattern: 25 hy- E per-in-hypoattenuating lesions (group B1) (Figs. 1C and 1D) and 11 entirely hyperattenuating lesions that did not show the midterm stage of the hyper-in-hypoattenuating type (group B2) (Fig. 4). Six of 25 hyper-in-hypoattenuating type lesions followed further became entirely hyperattenuating (group B3) (Figs. 1E and 1F). Cumulative attenuation conversion rates for the 60 hypoattenuating lesions were 15.8% at 1 year, 44.3% at 2 years, 58.7% at 3 years, and 77.2% at both 4 and 5 years. For the 53 patients, the cumulative attenuation conversion rates were 17.7% at 1 year, 48.9% at 2 years, 61.9% at 3 years, and 80.5% at both 4 and 5 years. In regard to differences in lesion enhancement pattern, cumulative attenuation conversion rates at 1 year, 2 years, and 3 years (Fig. 5) were as follows: 24.0%, 60.0%, and 80.0% for 25 hyper-inhypoattenuating lesions (group B1) and 27.3%, 72.7%, and 91.0% for 11 entirely hyperattenuating lesions (group B2). There was no statistically significant difference between hyper-in-hypoattenuating lesions (group B1) and entirely hyperattenuating lesions (group B2). The six hyper-inhypoattenuating lesions followed further (group B3) that transformed to entirely hyperattenuating lesions had an 83.3% cumulative conversion rate at 1 year, which was significantly higher than the rates in groups B1 and B2 (p < 0.001) (Fig. 5). The mean intervals between initial CT and final CT were 1,024 days (range, 153 3,100 days) in group A, 690 days (range, 162 1,783 days) in group B1, 741 days (range, 92 2,528 days) in group B2, and 1,120 days (range, 417 2,448 days) in group C. For group B3, the mean interval between hyper-in-hypoattenuating and entirely hyperattenuating change was 226 days (range, 148 406 days). Thirteen lesions that converted from hypoattenuation to hyperattenuation during follow-up were subjected to needle biopsy for confirmation of malignancy. The histopathologic results for the nine hyper-in-hypoattenuating lesions were well-differentiated HCC in seven cases and moderately differentiated HCC in two cases. The results for the four entirely enhanced lesions were well-differentiated HCC in one case and moderately differentiated HCC in three cases. Relation Between Chronologic Change in Lesion Size and Attenuation Conversion Pattern The mean sizes of 60 hypoattenuating lesions at initial and last CT examinations were 1.3 cm (range, 0.5 3.2 cm) and 2.0 cm (range, 0 4.8 cm) (Table 2). Chronologic changes in lesion size were as follows: 49 lesions in- AJR:187, August 2006 459

Takayasu et al. creased in size, eight were stable, and three disappeared. The mean sizes of lesions in groups A, B, B1, B2, and C at initial and final CT examinations are shown in Table 2. There was no statistically significant difference in ratio of mean tumor size on final CTto size on initial A C Fig. 4 56-year-old man with hypoattenuating lesion that progressed to entirely hyperattenuating lesion (overt HCC) (group B2) without detection of hyper-in-hypoattenuating type lesion (nodule-in-nodule HCC). A, Arterial phase CT scan shows hypoattenuating 0.8-cm lesion (arrow) in segment VII that was also hypoattenuating in delayed phase (not shown). B, 22nd follow-up arterial phase CT scan 6 years after A shows lesion is isoattenuating. C, Delayed phase of B shows hypoattenuating mass (arrow) measuring 2.5 cm. D, Arterial phase CT scan 6 months after B shows lesion (arrow) is entirely hyperattenuating and measures 3.0 cm. B D 460 AJR:187, August 2006

CT in Liver Disease Cumulative Attenuation Conversion (%) 100 80 60 40 20 0 Group B3 0 2 4 6 Observation Period (y) Fig. 5 Comparison of cumulative attenuation conversion rates for 57 hypoattenuating lesions (three lesions that regressed spontaneously were excluded). Group A = unchanged attenuation (n = 21), Group B1 = hyper-inhypoattenuating type (n = 25), Group B2 = entirely hyperattenuating type (n = 11), Group B3 = Group B1 lesions with additional follow-up (n = 6). Cumulative attenuation conversion rates of groups A, B1, and B2 were 0, 24.0%, and 27.3% at 1 year; 0, 60.0%, and 72.7% at 2 years; and 0, 80.0%, and 91.0% at 3 years. Cumulative attenuation conversion rate from hyper-in-hypoattenuating to entirely hyperattenuating type (group B3) was 83.3% at 1 year. Statistically significant difference was seen between groups B3 and B1 and between groups B3 and B2 (p < 0.001). No significant difference was seen between groups B1 and B2. TABLE 2: Relation Between Chronologic Change in Lesion Size and Attenuation Conversion Pattern No. of Mean Size of Lesion (cm) Lesion Size Ratio Group Lesions Initial CT Scan Final CT Scan Final/Initial CT a p b A 21 1.2 (0.5 2.5) 1.8 (1.0 4.8) 1.5 ± 0.6 NS B 36 1.4 (0.6 3.2) 2.2 (1.1 3.8) 1.8 ± 0.7 NS B1 25 1.4 (0.8 3.2) 2.3 (1.3 3.8) 1.8 ± 0.5 NS B2 11 1.3 (0.6 2.0) 2.1 (1.1 3.0) 1.9 ± 1.0 NS C 3 1.1 (1.0 1.3) 0 0 Overall 60 1.3 (0.5 3.2) 2.0 (0 4.8) 1.6 ± 0.7 Note Values in parentheses are ranges. NS = not significant. a Mean ± SD. b For mean ratio of lesion size, no statistically significant difference was seen between groups A and B, between groups A and B1, between groups A and B2, or between groups B1 and B2. CT among groups A, B, B1, and B2. The mean size of hyperattenuating portion within a hypoattenuating lesion (group B1) was 0.9 cm (range, 0.2 2.0 cm). Group B2 Group B1 Group A 8 10 Analyses of Factors Affecting Attenuation Conversion of Hypoattenuating Lesions Univariate analysis showed that hepatitis C virus antibody positivity, α-fetoprotein level, initial lesion size, and CT attenuation pattern were statistically significant factors (Table 1). Multivariate analysis revealed that hepatitis C virus antibody positivity (hazard ratio, 3.48; 95% confidence interval, 1.15 10.56; p = 0.028) and initial lesion size (hazard ratio, 2.21; 95% confidence interval, 1.24 3.94; p = 0.007) were independent factors for predicting attenuation transformation in hypoattenuating lesions. Emergence of Intrahepatic Recurrent Foci During Follow-up Period During the follow-up period, 19 recurrent foci developed in a total of 14 (26.4%) of 53 patients. All but two of these lesions derived from liver segments different from those of the primary lesion. Ten patients had 11 overt HCCs, two patients each had two hypoattenuating lesions, and two patients each had one overt HCC and one hypoattenuating lesion. The mean size of lesions was 1.9 cm (range, 1 2.4 cm), and the mean interval between initial CT and follow-up CT on which the recurrent foci were recognized was 823 days (range, 202 1,401 days). The cumulative recurrence rates at 1, 2, and 3 years were 0%, 18%, and 30%. Discussion Consensus has not been reached about the management of hypoattenuating nodular lesions in hepatitis virus related chronic liver disease. Several treatments such as surgery [11], PEI [19], and transarterial chemoembolization [23] have been used. However, the treatments have limitations; surgery is invasive, transarterial chemoembolization is not effective [23], and local ablation therapy can have severe complications, such as needle track implantation in PEI [20] and radiofrequency ablation related death [21]. Most surgically resected hypoattenuating lesions are pathologically diagnosed as earlystage HCC (high-grade dysplastic nodule) on the basis of characteristic macroscopic and microscopic findings [15, 16]. The discrepancy between findings of investigators inside [16, 24 26] and those outside Japan [13, 27, 28] regarding dysplastic nodule and early HCC has been reported to reflect differences in interpretation and application of nomenclature and not different biologic pathways [29, 30]. Histologic biopsy diagnosis of hypoattenuating lesions may not be easy because of factors such as sampling error for small lesions and intratumoral heterogeneity [12, 31 33]. Nevertheless, because of potential malignancy, hypoattenuating lesions have been treated without obtaining a malignant finding at all times by needle biopsy, which prevents elucidation of the true features of hypoattenuating nodules. In this study, CT attenuation patterns of 60 hypoattenuating lesions were hypo-hypoattenuating, iso-hypoattenuating, and hypo-isoattenuating in the arterial and delayed phases of dynamic CT. These findings were consistent with those of early HCC previously studied with dynamic CT [34]. Com- AJR:187, August 2006 461

Takayasu et al. bined CT attenuation patterns with CTHA and CTAP in 19 lesions showed compatible findings with those of early HCC [10], dysplastic nodule, and very well-differentiated HCC [35]. In a study at our hospital, Ueno et al. [19] found that 19 of 20 nodular lesions with no tumor stain on digital subtraction angiography or dynamic CT were histologically well-differentiated HCC. The other tumor was moderately differentiated. Our findings suggest the natural outcome in the majority of cases of early HCC and the minority of cases of adenomatous hyperplasia and atypical adenomatous hyperplasia. This study showed that 36 (60%) of 60 hypoattenuating lesions developed to hyperattenuating lesions. The overall cumulative attenuation conversion rates were 15.8% at 1 year, 44.3% at 2 years, and 58.7% at 3 years. The malignant transformation rate of adenomatous hyperplasia in the earlier study at our hospital was almost similar to that of the present study; 22% (4 of 18) of the lesions transformed within 1 year, 50% (5 of 10) within 2 years, and 80% (4 of 5) within 3 years [12]. Sakamoto and Hirohashi [31] reported that 13 (72%) and one (6%) of 18 nodules of adenomatous hyperplasia developed to early HCC and to overt HCC, respectively, and that four (33%) of 12 early HCCs progressed to overt HCC. Borzio et al. [33] reported that 28 (31%) of sonographically detected macronodules (including large regenerative nodules, low-grade dysplastic nodules, and high-grade dysplastic nodules) [13] in cirrhosis transformed into HCC within 3 years. The current study revealed that the cumulative attenuation conversion rate from hyper-inhypoattenuating (group B1) to entirely hyperattenuating type (group B3) was significantly higher than that for groups B1 and B2. These findings suggest that once a hypoattenuating lesion develops to the hyper-in-hypoattenuating type, the speed of progression to overt HCC may be accelerated. Similar findings were reported in studies conducted with MRI [36] and a combination of CTHA and CTAP [37]. Findings for all 13 lesions on which biopsy was performed after attenuation conversion revealed that hyper-in-hypoattenuating lesions had a higher ratio of well-differentiated to moderately differentiated HCC than did entirely hyperattenuating lesions. These findings indicate that to reduce the complications of needle track seeding and intrahepatic metastasis by radiofrequency ablation, patients with hyper-in-hypoattenuating lesions are better candidates for treatment than those with entirely hyperattenuating lesions. When the ratio of lesion size on both the initial and final CT examinations was compared with attenuation conversion pattern, there was no significant difference between these two factors. This finding suggests that careful follow-up should be focused on transformation of CT attenuation rather than on change in lesion size. Analysis of factors influencing conversion to hyperattenuation showed two statistically significant independent factors: hepatitis C viral antibody positivity and initial size of the lesion. A recent study [17] showed that multistep progression of hepatocarcinogenesis was seen at a significantly higher rate in patients with positive results for hepatitis C virus antibody than in those with positive results for hepatitis B virus surface antigen. Chronic persistent inflammation of the liver in the presence of hepatitis C viral antibody [38] may result in progression of hepatocarcinogenesis. Lim et al. [39] using helical CT drew the same conclusions as we did about initial tumor size, but Borzio et al. [33] did not come to the same conclusion. Recurrent foci were considered a multicentric occurrence rather than intrahepatic metastasis on the basis of results of clinical [11, 19] and pathologic [17] studies of early and overt HCC. The cumulative recurrence rates of 30% at 3 years and 60% at 5 years in our study were intermediate between those of surgical resection (15% and 53%) [11] and PEI (57% and 78%) [19]. There were some limitations to this study. No lesions were histopathologically proved at needle biopsy when the lesions were first diagnosed as hypoattenuating. The advantages and limitations of this factor in this study are mentioned earlier. Another limitation was changing CT technology and progression to use of thinner slice thickness over time because of the long period needed to elucidate the natural outcomes. Another limitation may have been the possibility of radiation hazard to patients undergoing repeated CT examinations. To avoid unnecessary risk, contrast-enhanced sonography or dynamic MRI may be alternatives [40, 41]. This study was retrospective; a prospective study with a larger number of patients is needed. We conclude that hypoattenuating nodular lesions of chronic liver disease progressed to hyperattenuating lesions (malignant transformation) on dynamic CT with frequencies of 15.8% and 58.7% at 1 and 3 years. Hyper-inhypoattenuating type lesions corresponding to nodule-in-nodule HCC developed significantly more rapidly than entirely hyperattenuating lesions (overt HCC). More attention should be paid to attenuation conversion than to change in lesion size to obtain optimal treatment, especially of patients with positive results for hepatitis C viral antibody and a relatively large lesion size on initial CT. Acknowledgments We thank Chise Sato and Hiroyo Ohchi for data collection and Kinuko Tajima for statistical analysis. References 1. Parkin DM, Bray F, Ferlay J, Pisani P. Estimating the world cancer burden: Globocan 2000. Int J Cancer 2001; 94:153 156 2. Arii S, Yamaoka Y, Futagawa S, et al. Results of surgical and nonsurgical treatment for small-sized hepatocellular carcinomas: a retrospective and nationwide survey in Japan. The Liver Cancer Study Group of Japan. Hepatology 2000; 32:1224 1229 3. Livraghi T, Giorgio A, Marin G, et al. Hepatocellular carcinoma and cirrhosis in 746 patients: longterm results of percutaneous ethanol injection. Radiology 1995; 197:101 108 4. Buscarini L, Buscarini E, Di Stasi M, Vallisa D, Quaretti P, Rocca A. Percutaneous radiofrequency ablation of small hepatocellular carcinoma: longterm results. Eur Radiol 2001; 11:914 921 5. Mazzaferro V, Regalia E, Doci R, et al. Liver transplantation for the treatment of small hepatocellular carcinomas in patients with cirrhosis. N Engl J Med 1996; 334:693 699 6. Murakami T, Kim T, Takamura M, et al. Hypervascular hepatocellular carcinoma: detection with double arterial phase multi-detector row helical CT. Radiology 2001; 218:763 767 7. Takayasu K, Makuuchi M, Hirohashi S, et al. Imaging of adenomatous hyperplastic lesions containing and not containing hepatocellular carcinoma in the liver [in Japanese]. Nippon Shokakibyo Gakkai Zasshi 1989; 86:2404 2412 8. Takayasu K, Moriyama N, Muramatsu Y, et al. The diagnosis of small hepatocellular carcinomas: efficacy of various imaging procedures in 100 patients. AJR 1990; 155:49 54 9. Matsui O, Kadoya M, Kameyama T, et al. Benign and malignant nodules in cirrhotic livers: distinction based on blood supply. Radiology 1991; 178:493 497 10. Takayasu K, Muramatsu Y, Furukawa H, et al. Early hepatocellular carcinoma: appearance at CT during arterial portography and CT arteriography with pathologic correlation. Radiology 1995; 194:101 105 11. Takayama T, Makuuchi M, Hirohashi S, et al. Early 462 AJR:187, August 2006

CT in Liver Disease hepatocellular carcinoma as an entity with a high rate of surgical cure. Hepatology 1998; 28:1241 1246 12. Takayama T, Makuuchi M, Hirohashi S, et al. Malignant transformation of adenomatous hyperplasia to hepatocellular carcinoma. Lancet 1990; 336:1150 1153 13. International Working Party. Terminology of nodular hepatocellular lesions. Hepatology 1995; 22:983 993 14. Krinsky G. Terminology of hepatocellular nodules in cirrhosis: plea for consistency. (editorial) Radiology 2002; 224:638 15. Kanai T, Hirohashi S, Upton MP, et al. Pathology of small hepatocellular carcinoma: a proposal for a new gross classification. Cancer 1987; 60:810 819 16. Sakamoto M, Hirohashi S, Shimosato Y. Early stages of multistep hepatocarcinogenesis: adenomatous hyperplasia and early hepatocellular carcinoma. Hum Pathol 1991; 22:172 178 17. Oikawa T, Ojima H, Yamasaki S, Takayama T, Hirohashi S, Sakamoto M. Multistep and multicentric development of hepatocellular carcinoma: histological analysis of 980 resected nodules. J Hepatol 2005; 42:225 229 18. Winter TC 3rd, Takayasu K, Muramatsu Y, et al. Early advanced hepatocellular carcinoma: evaluation of CT and MR appearance with pathologic correlation. Radiology 1994; 192:379 387 19. Ueno H, Okada S, Okusaka T, et al. Prognosis of hepatocellular carcinoma with no tumor stain treated by percutaneous ethanol injection. Hepatogastroenterology 2001; 48:1430 1434 20. Ishii H, Okada S, Okusaka T, et al. Needle tract implantation of hepatocellular carcinoma after percutaneous ethanol injection. Cancer 1998; 82:1638 1642 21. de Baere T, Risse O, Kuoch V, et al. Adverse events during radiofrequency treatment of 582 hepatic tumors. AJR 2003; 181:695 700 22. Curley SA, Marra P, Beaty K, et al. Early and late complications after radiofrequency ablation of malignant liver tumors in 608 patients. Ann Surg 2004; 239:450 458 23. Takayasu K, Wakao F, Moriyama N, et al. Response of early-stage hepatocellular carcinoma and borderline lesions to therapeutic arterial embolization. AJR 1993; 160:301 306 24. Arakawa M, Kage M, Sugihara S, Nakashima T, Suenaga M, Okuda K. Emergence of malignant lesions within an adenomatous hyperplastic nodule in a cirrhotic liver: observations in five cases. Gastroenterology 1986; 91:198 208 25. Wada K, Kondo F, Kondo Y. Large regenerative nodules and dysplastic nodules in cirrhotic livers: a histopathologic study. Hepatology 1988; 8:1684 1688 26. Terada T, Terasaki S, Nakanuma Y. A clinicopathologic study of adenomatous hyperplasia of the liver in 209 consecutive cirrhotic livers examined by autopsy. Cancer 1993; 72:1551 1556 27. Ferrell L, Wright T, Lake J, Roberts J, Ascher N. Incidence and diagnostic features of macroregenerative nodules vs small hepatocellular carcinoma in cirrhotic livers. Hepatology 1992; 16:1372 1381 28. Theise ND, Lapook JD, Thung SN. A macroregenerative nodule containing multiple foci of hepatocellular carcinoma in a noncirrhotic liver. Hepatology 1993; 17:993 996 29. Hirohashi S, Ishak K, Kojiro M, et al. Hepatocellular carcinoma. In: Hamilton S, Aaltonen L, eds. World Health Organization classification of tumours: pathology and genetics of tumours of the digestive system. Lyon, France: International Agency for Research on Cancer Press, 2000:159 172 30. Theise ND, Park YN, Kojiro M. Dysplastic nodules and hepatocarcinogenesis. Clin Liver Dis 2002; 6:497 512 31. Sakamoto M, Hirohashi S. Natural history and prognosis of adenomatous hyperplasia and early hepatocellular carcinoma: multi-institutional analysis of 53 nodules followed up for more than 6 months and 141 patients with single early hepatocellular carcinoma treated by surgical resection or percutaneous ethanol injection. Jpn J Clin Oncol 1998; 28:604 608 32. Seki S, Sakaguchi H, Kitada T, et al. Outcomes of dysplastic nodules in human cirrhotic liver: a clinicopathological study. Clin Cancer Res 2000; 6:3469 3473 33. Borzio M, Fargion S, Borzio F, et al. Impact of large regenerative, low grade and high grade dysplastic nodules in hepatocellular carcinoma development. J Hepatol 2003; 39:208 214 34. Takayasu K, Furukawa H, Wakao F, et al. CT diagnosis of early hepatocellular carcinoma: sensitivity, findings, and CT pathologic correlation. AJR 1995; 164:885 890 35. Hayashi M, Matsui O, Ueda K, et al. Correlation between the blood supply and grade of malignancy of hepatocellular nodules associated with liver cirrhosis: evaluation by CT during intraarterial injection of contrast medium. AJR 1999; 172:969 976 36. Sadek AG, Mitchell DG, Siegelman ES, Outwater EK, Matteucci T, Hann HW. Early hepatocellular carcinoma that develops within macroregenerative nodules: growth rate depicted at serial MR imaging. Radiology 1995; 195:753 756 37. Hayashi M, Matsui O, Ueda K, Kawamori Y, Gabata T, Kadoya M. Progression to hypervascular hepatocellular carcinoma: correlation with intranodular blood supply evaluated with CT during intraarterial injection of contrast material. Radiology 2002; 225:143 149 38. Takenaka K, Yamamoto K, Taketomi A, et al. A comparison of the surgical results in patients with hepatitis B versus hepatitis C related hepatocellular carcinoma. Hepatology 1995; 22:20 24 39. Lim JH, Seo JW, Choi D, Jang H, Lee W, Lim HK. Unidentified small low-attenuating hepatocellular nodules on helical CT in patients with chronic liver diseases: two-year follow-up. Chicago, IL: Radiological Society of North America, 2001:491 40. Ding H, Kudo M, Onda H, Suetomi Y, Minami Y, Maekawa K. Hepatocellular carcinoma: depiction of tumor parenchymal flow with intermittent harmonic power Doppler US during the early arterial phase in dual-display mode. Radiology 2001; 220:349 356 41. Choi BI, Kim TK, Han JK, Kim AY, Seong CK, Park SJ. Vascularity of hepatocellular carcinoma: assessment with contrast-enhanced second-harmonic versus conventional power Doppler US. Radiology 2000; 214:381 386 AJR:187, August 2006 463