Staging and treatment of hepatocellular carcinoma

Transcription

1 Multimodal Image-Guided Tailored Therapy of Early and Intermediate Hepatocellular Carcinoma: Long-Term Survival in the Experience of a Single Radiologic Referral Center Tito Livraghi, 1 Franca Meloni, 1 Alberto Morabito, 2 * and Claudio Vettori 1 Abbreviations: AFP, serum alpha-fetoprotein; BCLC, Liver Unit of Barcelona classification; CT, computed tomography; DCP, serum des-gamma-carboxy-prothrombin; EASL, European Association for the Study of the Liver; HCC, hepatocellular carcinoma; MIGTT, multimodal image-guided tailored therapy; PEI, percutaneous ethanol injection; RFA, radiofrequency ablation; stace, selective transcatheter arterial chemoembolization; TACE, transcatheter arterial chemoembolization; TNM, tumor-node-metastasis. From the 1 Radiology Department, Ospedale Civile, Vimercate, Milano, Italy, and 2 S. Paolo Hospital, Milano, Italy; * 2 Chair of Medical Statistics. Address reprint requests to Tito Livraghi, M.D., Radiology Department Via Cereda 23, Vimercate, Italy. Telephone: ; FAX: ; tito.livraghi@aovimercate.org. Copyright 2004 by the American Association for the Study of Liver Diseases Published online in Wiley InterScience ( DOI /lt The best treatment policy for some patients with hepatocellular carcinoma (HCC) and compensated cirrhosis is still controversial. The aim of this study was to evaluate the long-term survival and related prognostic factors of patients with early and intermediate HCC (Liver Unit of Barcelona classification) treated in a radiologic referral center by a multimodal image-guided tailored therapy (MIGTT), applied over time, choosing the procedure patient by patient according to the presentation of the disease. Between May 1996 and May 2003, 374 patients (210 with early and 164 with intermediate HCC) were treated with MIGTT. Radiofrequency ablation was considered the first choice; ethanol injection was preferred for nodules at risk for radiofrequency; and selective chemoembolization was preferred for nodules not recognizable at ultrasound examination, those not retreatable after an unsuccessful ablation technique, or for satellites. The rate of perioperative mortality and major complications was 0.2% and 4.5%, respectively. The 1-, 3-, and 5-year survival rates were 90%, 69%, and 49% and 83%, 43%, and 28% for patients with early and intermediate HCC, respectively. At the univariate analysis, the indicators of a poor prognosis were Child s class B, portal hypertension, abnormal bilirubin, infiltrating tumor, and abnormal serum alpha-fetoprotein (AFP) level. In conclusion, within the limits of historical comparisons, in early HCC, MIGTT should be considered an appropriate option for unresectable patients or for resectable patients presenting with adverse prognostic factors. In intermediate HCC, the 3-year survival rate obtained with MIGTT was better than the best survival rate reported with conventional chemoembolization; however, the benefit for patients presenting with poor prognostic factors remains unclear. (Liver Transpl 2004;10:S98 S106.) Staging and treatment of hepatocellular carcinoma (HCC), even though fifth in frequency among all malignancies, 1 have not yet achieved a general consensus. The tumor-node-metastasis (TNM) staging system has been criticized, and other staging and prognostic systems that assess additional and different parameters have been proposed. 2 5 One of them, proposed in 1999 by the Liver Unit of Barcelona (BCLC classification), 5 was accepted as a guideline for treatment at the Consensus Conference of the European Association for the Study of the Liver (EASL). 6 This staging classification, although rather intentionally simple, is easily applicable and provides useful information for clinical and therapeutic purposes. According to this staging classification, HCC patients are stratified into 4 categories (early, intermediate, advanced, and terminal), each linked to certain treatment indications. Patients at an early stage are those with chronic hepatitis or compensated cirrhosis who present asymptomatic single HCC or up to 3 nodules 3 cm or less. Patients exceeding these neoplastic parameters, that is, (1) with larger multinodular tumors (and also with an inoperable single tumor 5cm according to a simplified model of stratification), 7 (2) without cancer-related symptoms (performance status test 0), and (3) without vascular invasion or extrahepatic spread, fit into the intermediate stage. According with the EASL guidelines, patients with early HCC can benefit from curative or effective therapies such as liver transplantation, partial resection, or percutaneous ablation treatments; patients with intermediate HCC are generally candidates to receive only noncurative regional intraarterial therapies. However, no unequivocal evidence exists to establish the best treatment for several patients with compensated cirrhosis. Accordingly, each referral center, taking into account the studies available, follows a personal algorithm for borderline patients. Following the staging proposed by EASL, the current study evaluated the long-term survival of patients S98 Liver Transplantation, Vol 10, No 2, Suppl 1 (February), 2004: pp S98 S106

2 MIGTT in Early and Intermediate HCC S99 Table 1. Single-Variable Analysis of Possible Prognostic Factors For Cumulative Survival in 210 Patients With Early HCC Undergoing Multimodal Image-Guided Tailored Treatment Variable Patients No. 1y 3y 5y P Value Age NS Number single NS multiple Type nodular infiltrating Bilirubin normal abnormal Portal absent hypertension present AFP DCP NS Child A B Abbreviations: NS, not significantly different. with early and intermediate HCC treated in a single radiologic referral center with a multimodal imageguided tailored therapy (MIGTT). The MIGTT consisted of radiofrequency ablation (RFA), percutaneous ethanol injection (PEI), and selective transarterial chemoembolization (stace), applied over time, choosing the technique patient by patient according to the presentation of the disease. An additional outcome of the study was the assessment of the clinical variables that can influence the prognosis and the therapeutic strategy. Material and Methods Patients The current study considered only patients treated since RFA was routinely used in our department. Between May 1996 and May 2002, 374 consecutive patients were treated using an MIGTT: 210 patients (mean age, 68.7 years; age range, years) with early HCC [considered inoperable for one of the following factors: (1) Child s class B, (2) transaminases 3 times the normal values, (3) 75 years of age, (4) serum bilirubin 1.5 mg/dl, (5) anatomical resection unfeasible, (6) multiple tumors, (7) refusal of surgery, or also considered operable with tumor till to 2.0 cm, the rational being the certainty to achieve a complete necrosis using RFA and the low rate of perilesional neoplastic microinvasion at such size], and 164 patients (mean age, 69.3 years; age range, years) with intermediate HCC (excluding those with more than approximately 30% of the liver occupied by the tumor). Four additional patients with early HCC treated with RFA as a bridge to liver transplantation were excluded. Two patients with severe coagulation disorders (prothrombin activity 40% or platelet count 30,000/mL), and patients with severe cirrhosis (Child-Pugh class C) or encephalopathy were excluded from the study. The clinical data of patients with early HCC are summarized in Table 1, and the clinical data of those patients with intermediate HCC are summarized in Table 2. The tumor was considered nodular when it presented as a capsule or with well-demarcated margins; it was considered infiltrating when it presented as at least 1 nodule with irregular margins or the presence of satellites. Serum alphafetoprotein (AFP), serum des-gamma-carboxy-prothrombin (DCP), and bilirubin concentration were considered normal or abnormal according to the following levels, respectively: 100 ng/ml and 100 ng/ml, 2 ng/ml and 2 ng/ml, 1.2 mg/dl and 1.2 mg/dl. Portal hypertension was considered present when the patient had splenomegaly and platelet count less than 70,000/mL. 7 Of the patients with intermediate HCC, 94 were enrolled presenting this stage at the first detection, and 70 had previously been treated in other centers for early HCC (6 with surgery and 64 with PEI or RFA). Their survival was measured from the date of enrollment at our center. The diagnosis of HCC was confirmed by 2 coincident imaging techniques, by 1 imaging technique associated with AFP levels 400 ng/ml, or by biopsy in doubtful cases. 6 Table 2. Single-Variable Analysis of Possible Prognostic Factors for Cumulative Survival in 164 Patients With Intermediate HCC Undergoing Multimodal Image-Guided Tailored Treatment Variable Patients No. 1y 3y 5y P Value Age NS Number single NS multiple Type nodular NS infiltrating Bilirubin normal NS abnormal Portal absent hypertension present AFP DCP Child A B Abbreviations: NS, not significantly different.

3 S100 Livraghi et al. Statistical Analysis We investigated the overall 5-year survival (Kaplan Meyer method) and the relationship of 8 pretreatment clinicopathological variables, divided in 2 subsets, that could affect survival: age ( 65 or 65 years), number of tumors (single or multiple), tumor pattern (nodular or infiltrating), serum AFP ( 100 ng/ml or 100 ng/ml), serum DCP ( 2 ng/ml or 2 ng/ml), bilirubin level ( 1.2 mg/dl or 1.2 mg/dl), portal hypertension (no or yes), Child s class (A or B). Univariate analysis was applied according to the Cox proportional hazard risk model. Choice of Treatment The 3 therapeutic procedures were considered complementary and were used according to the features of the disease, especially location, presence of satellites, and response to the initial procedure. The approach meant a tailored strategy for each patient. For instance, the same lesion could be treated with only one or a combination of different techniques when the first result was unsatisfactory. Otherwise, a multifocal HCC could be treated with only one or with all the techniques, during a single hospital stay or over the years. RFA was considered the first choice treatment. Such a choice reflected the results of previous studies demonstrating the advantages of RFA in terms of local efficacy and side effects. In these studies, RFA obtained a higher rate of necrosis than PEI in small tumors and in infiltrating lesions of any size and avoided the side effects (particularly transient increase in portal hypertension) occurring after single-session PEI when a large amount of ethanol was required According to our previous experience, RFA also obtained greater local efficacy than whole-liver TACE without its side effects and without impairment of liver function (Fig. 1). 13 PEI was preferred in (1) lesions at risk with RFA, 14 that is, adjacent to main biliary ducts or to intestinal loops (above all when fibrotic adhesions between the hepatic capsule and intestinal wall were suspected because of the risk of perforation, Fig. 2), (2) small lesions with a location difficult to reach where the use of the fine needle for PEI ensured a less traumatic approach in case of repeated attempts (Fig. 3), and (3) portions of tumor close to large vessels (because of the heat sink effect ). Selective TACE was used in (1) lesions not recognizable at ultrasound examination, (2) lesions not completely necrotized and presenting the remnant vital tissue scattered or not recognizable at ultrasound examination for an additional 3 Figure 1. (A, B, C, D). Multinodular bilobar hepatocellular carcinoma (HCC) treated with radiofrequency ablation (RFA). Portal phase computed tomography (CT) scan obtained 18 months after therapy (1 insertion per each nodule) presented high local efficacy for every tumor not achievable with conventional whole-body transcatheter arterial chemoembolization (TACE). In this case, RFA also avoided the side effects and the damage of nonneoplastic tissue usually obtained after conventional TACE.

4 MIGTT in Early and Intermediate HCC S101 internally cooled electrodes (Radionics Inc., Burlington, MA, USA) were used. For PEI we used a 21-gauge needle with a closed conical tip and 3 terminal side holes (PEIT needle; Hakko, Tokyo, Japan). For stace, a 3.0 F microcatheter (FasTracker-325; Boston Scientific, Cork, Ireland) was inserted into the feeding segmental or subsegmental artery through a 5-F catheter indwelling in the proximal hepatic artery. An emulsion of 1 5 ml iodized poppyseed oil (Lipiodol Ultra-Fluid; Guerbet, Paris, France) and mg epirubicin hydrochloride (Farmorubicina; Pharmacia & Upjohn, Milan, Italy) was introduced, followed by gelatin sponge particles (Gelfoam; Upjohn, Kalamazoo, USA). More detailed technical information about the procedures has been published already. 8 13,15 Results Treatment No patient was excluded from treatment for technical reasons. All but 3 nodules in 2 patients with multiple HCC (because of the impossibility to treat them with Figure 2. Single, inoperable hepatocellular carcinoma (HCC), 5.5 cm in diameter, located in segment 6, surrounded by intestinal loops. Owing to the risk of perforation, one-session percutaneous ethanol injection (PEI) was preferred to radiofrequency ablation (RFA). Arterial phase computed tomography (CT) obtained the day after therapy demonstrated no enhancement inside the tumor, suggesting its complete necrosis (confirmed during follow-up). Gas caused by necrosis (a normal pattern in the days immediately after PEI) is also recognizable. treatment with RFA or PEI (Fig. 4), and (3) the presence of satellite nodules after the achievement of complete necrosis of the main tumor after RFA or PEI (Fig. 5). In conformity with our experience, the uptake of such small volumes of neoplastic tissue is given easily by the absence of arterial theft from the rich supply of the main tumor. Pretreatment and Posttreatment Work-Up For all patients, pretreatment and posttreatment work-up included ultrasound examination, more recently using second-generation echo enhancers (Sonovue, Bracco, Milan, Italy), as well as unenhanced and dual phase, contrast-enhanced helical computed tomography (CT). The work-up was performed before treatment, 1 month after treatment, and subsequently every 4 6 months after treatment. In some patients, usually at the beginning of the RFA experience or in cases at risk, CT examination was performed the day after treatment. Additionally, AFP and DCP levels were obtained using the same timing in all patients. Technique Percutaneous ablation techniques were performed under realtime ultrasound guidance. For RFA, 20-cm-long, 18-gauge Figure 3. Multinodular hepatocellular carcinoma (HCC), consisting of 6 small tumors located in the upper portion of the liver. The patient was previously successfully treated with radiofrequency ablation (RFA) or percutaneous ethanol injection (PEI), in different sessions, for other 16 nodules. Owing to the difficult approach, one-session PEI, less traumatic than RFA, was preferred for such presentation. Portal phase computed tomography (CT) scan, obtained 18 months after therapy, demonstrated no enhancement (also during the arterial phase), suggesting complete necrosis. After 5-year follow-up, the patient is apparently free of disease.

5 S102 Livraghi et al. Figure 4. (A, B, C, D). Single, inoperable, infiltrating hepatocellular carcinoma (HCC), 6.7 cm in diameter, treated with radiofrequency ablation (RFA) and segmental transcatheter arterial chemoembolization (TACE) during 1 hospital stay. (a) Computed tomography (CT) scan during the arterial phase demonstrating the tumor before therapy. (b) CT scan during the arterial phase performed the day after RFA (5 insertions), showing the persistence of a small hyperdense vital neoplastic area (arrow) at the tumor periphery. (c) Because this area was unrecognizable at ultrasound examination, segmental TACE was performed the following day. CT scan obtained immediately after treatment showed good distribution of Lipiodol around the tumor. (d) CT scan during the arterial phase performed 2 years after therapy demonstrated no enhancement inside the large neoplastic area treated with RFA and complete uptake of Lipiodol inside the minute neoplastic area treated with TACE.

6 MIGTT in Early and Intermediate HCC S103 stace for anatomical difficulties) were treated. With 1 or more sessions, 97 patients with early HCC received only RFA; 10 patients received only PEI; 4 patients received only stace; 49 patients received PEI and RFA; 31 patients received RFA and stace; 2 patients received PEI and stace; and 17 patients received PEI, RFA, and stace. In total, RFA was used for 194 patients, PEI was used for 78 patients, and stace was used for 54 patients. A total of 199 (94.7%) patients received more than 1 treatment session during the follow-up. With 1 or more sessions, 79 patients with intermediate HCC received only RFA; 8 patients received only PEI; 3 patients received only stace; 34 patients received RFA and stace; 26 patients received RFA and PEI; and 12 patients received RFA, PEI, and stace. In total, RFA was used in 151 patients, stace was used in 49 patients, and PEI was used in 46 patients. A total of 126 (76.8%) patients received more than 1 treatment session during the follow-up. Complications No deaths occurred in patients with early HCC. Seven cases (3.3%) of major complications were reported, all after RFA but 1 after PEI. Four patients had neoplastic seeding, 1 patient had pleural hemorrhage, and 2 patients had peritoneal hemorrhage. One death (0.6%) occurred in a patient with intermediate HCC (septic shock and multiorgan failure). This patient s case has been reported in previous studies. 11,14 Ten major complications (6.0%) occurred, all after RFA: 2 patients had neoplastic seeding, 2 patients had peritoneal hemorrhages, 1 patient with a subglissonian tumor adjacent to intestinal loops had peritonitis caused by perforation of the colonic wall, 1 patient had formation of arterioportal shunt with intrahepatic hematoma and peritoneal hemorrhage, 1 patient had septicemia, 1 patient had diaphragmatic perforation with bile pleural effusion, 1 patient had gastric perforation, and 1 patient had transient liver impairment. In total, the mortality 4 Figure 5. (A, B). Multinodular hepatocellular carcinoma (HCC) consisting of 3 nodules treated with radiofrequency ablation (RFA). (a) During the follow-up, all nodules showed complete necrosis, but the largest nodule (5 cm in diameter at baseline) presented the growth of same peripheral satellites on computed tomography (CT) scan performed during the arterial phase. (b) After selective transcatheter arterial chemoembolization (TACE), CT scan performed during the portal phase demonstrated the complete uptake of Lipiodol inside the satellites.

7 S104 Livraghi et al. rate was 0.2%, and the major complications rate was 4.5%. Survival Rates and Analysis of Prognostic Factors Follow-up ranged from 12 to 84 months. At the end of the follow-up (May 2003), the overall 1-, 3-, and 5-year calculated survival rate of the 210 patients with early HCC was 90%, 69%, and 49%, and that of the 164 patients with intermediate HCC was 83%, 43%, and 28%, respectively. According to the different parameters evaluated, 1-, 3-, and 5-year survival rates are reported in Tables 1 and 2. Therefore, according to the statistical analysis, the survival of patients with early HCC was prevalently affected by bilirubin levels, Child s class, portal hypertension, AFP levels, and type of tumor, whereas that of patients with intermediate HCC was prevalently affected by Child s class, AFP and DCP levels, and portal hypertension. At the end of the study, the rate of patients free of disease with early and intermediate HCC was 25.2% and 17.6% respectively. Discussion HCC usually coexists with an underlying hepatic chronic disease. According to the stage, one disease will prevail over the other. For such reason, therapies should not worsen liver function. HCC is an organ pathology; therefore, the first nodule detected is only a prelude to others. A study on resected patients demonstrated that multicentricity is already present in 50% of early stages and that 93% of patients with a single, minute HCC presented other nodules within 5 years. 16 Being multicentric over time, HCC needs multistep treatments. Early HCC Liver transplantation is the only option able to offer a definitive cure. Partial resection or percutaneous ablation treatments can offer a palliative cure, achieving a definitive cure only locally. In fact, according to a Japanese nationwide survey, only 1.6% of all resected patients presenting with intrahepatic recurrence was reresected. 17 Although it is understood that partial resection ensures the highest possibility to completely ablate the tumor, different comparative studies based on historical results 18,19 and the only prospective trial comparing surgery and PEI demonstrated roughly equivalent results. 20 In the prospective study, comparable patients with early HCC (single or multiple 3 cm) presented a 3- to 5-year survival of 82% and 59% and 84% and 61%, respectively. The explanation is probably because of a balance among advantages and disadvantages of the two therapies, the most important advantages of PEI (and RFA) being repeatability, no loss or damage of nonneoplastic tissue, and lower complication rate. Generally, the overall results of both therapies were hampered by an incorrect selection of the patients, part of them being treated even though they had adverse prognostic factors. For instance, the Liver Unit of Barcelona reported the usual the mean rate reported by most studies 5-year overall survival rate around 50% after resection. 19,21 23 However, when the patients were divided according to two simple adverse prognostic factors portal hypertension and abnormal bilirubin a rate of 74% was obtained (the best so far reported) in patients with normal values and a rate of only 25% in the worst candidates. 21 The fact that the survival of the worst candidates was comparable with two recently reported survival rates of untreated patients (20% and 16% respectively), even with a more adverse profile, 5,24 questions the indication for surgery in such patients. In our study, 5-year survival of patients with portal hypertension and abnormal bilirubin was 38%. Therefore, within the limits of historical comparisons, MIGTT could be considered an appropriate option for resectable patients with the aforementioned adverse prognostic factors. In any case, even after many years and the development of new procedures and technical improvements, most survival curves for patients treated with resection as well for patients treated with percutaneous therapies have not shown significant increases The explanation could be that such survival curves were probably penalized and flattened by the presence of a number of patients who presented with a particularly aggressive tumor, refractory to any therapy, whose characteristics are still to be identified. Research in this direction is ongoing. 24 Nevertheless, it is plausible that the patients of the current study, being treated with therapies complementary to one another, may present an intention-to-treat survival better than that obtained with a single therapy. Intermediate HCC Every patient with the early stage not treated with liver transplantation is destined to present, in due time, the intermediate stage, unless portal thrombosis appears contemporaneously, rendering the stage advanced. Thus, intermediate HCC is often an obligatory phase during the natural course of the disease in patients cured with resection or other therapies for early HCC

8 MIGTT in Early and Intermediate HCC S105 or the phase at which the disease is first detected. For instance, in the current study, 42.7% of the patients were enrolled for progression of disease (generally because of new nodules) after an initial curative treatment of early HCC, and 57.3% of the patients were enrolled presenting in the intermediate stage at first detection. There is no standard therapy for patients at this stage. For these patients, the usual therapeutic option is whole-liver or lobar TACE, even though no general consensus guarantees such a therapeutic approach. In fact, the most randomized studies, however weak for some bias, were not able to demonstrate a clear improvement of survival, and the lack of benefit was related to a counterbalance between local tumor control and damage to nonneoplastic tissue, which can hasten liver insufficiency. However, two more recent trials were able to demonstrate a benefit after TACE, one reporting a 2- and 3-year survival of 63% and 29% for treated patients stringently selected (the best survival so far reported) versus the 27% and 17% for the control group (P.009). 25,26 One purpose of our study was to report the longterm survival of patients, traditionally treated with conventional TACE, who were treated with MIGTT. The rational for such a therapeutic strategy was based on the results obtained after a trial in which we compared only RFA and conventional TACE in patients with intermediate HCC. The trial demonstrated that RFA obtained greater local efficacy without the side effects and without the impairment of liver function of TACE. 13 In such a trial, although comparison of the control of tumor growth (50% in the RFA group vs. 30% in the TACE group) was not statistically significant, the major complication rate after treatment and the 2-year survival rate were significantly in favor of the RFA group (P.07 and P.05, respectively). After these results, we decided to abandon the treatment with conventional TACE in favor of the MIGTT reported here and to use TACE only selectively. Although a possible comparison is based only on historical data, the 3-year survival obtained with MIGTT was better than the best survival obtained with conventional TACE (i.e., 43% vs. 29%). Analysis of Prognostic Factors As reported in other studies on treated or untreated HCC in cirrhosis, statistically adverse prognostic factors often were related to the grade of liver function (i.e., Child s class B), abnormal bilirubin and portal hypertension, and the infiltrating type of tumor or abnormal neoplastic markers. Our results confirmed such data. Considering the survival of patients with early HCC reported in Table 1, treatment of patients presenting several of these adverse prognostic factors remains questionable. Considering the survival of patients with intermediate HCC reported in Table 2, the Child s class B even alone questions the usefulness of any kind of treatment. With regard to the comparison according to the stage, only in early HCC bilirubin concentration and type of tumor results were significant. References 1. Parkin DM, Bray F, Ferlay J, Pisani P. Estimating the world cancer burden: GLOBOCAN Int J Cancer 2001; 94: Makuuchi M, Belghiti, J Belli G, et al. IHPBA concordant classification of primary liver cancer: working group report. J Hepatobiliary Pancreat Surg 2003;10: The Cancer of the Italian Program (CLIP) investigators. A new prognostic system for hepatocellular carcinoma: a retrospective study of 435 patients. Hepatology 1998;28: Chevret S, Trinchet JC, Mathieu D, Rached AA, Beaugrand M, Chastang C. A new prognostic classification for predicting survival in patients with hepatocellular carcinoma. J Hepatol 1999; 31: Llovet JM, Bru C, Bruix J. Prognosis of hepatocellular carcinoma: the BCLC staging classification. Semin Liver Dis 1999; 19: Bruix J, Sherman M, Llovet JM, et al. Clinical management of hepatocellular carcinoma. Conclusions of the Barcelona-2000 EASL Conference. J Hepatol 2001;35: Bruix J, Llovet JM. Prognostic prediction and treatment strategy in hepatocellular carcinoma. Hepatology 2002;35: Livraghi T, Goldberg SN, Lazzaroni S, Meloni F, Solbiati L, Gazelle GS. Small hepatocellular carcinoma: treatment with radio-frequency ablation versus ethanol injection. Radiology 1999;210: Ikeda M, Okada S, Ueno H, Okusaka T, Kuryama H. Radiofrequency ablation and percutaneous ethanol injection in patients with small hepatocellular carcinoma: a comparative study. Jpn J Clin Oncol 2001;31: Lencioni R, Allgaier HP, Cioni D, et al. Small hepatocellular carcinoma in cirrhosis: randomized comparison of radiofrequency thermal ablation versus percutaneous ethanol injection. Radiology 2003;228: Livraghi T, Goldberg SN, Lazzaroni S, et al. Hepatocellular carcinoma: radio-frequency ablation of medium and large lesions. Radiology 2000;214: Livraghi T, Benedini V, Lazzaroni S, Meloni F, Torzilli G, Vettori C. Long term results of single session PEI in patients with large hepatocellular carcinoma. Cancer 1998;83: Livraghi T, Meloni F, Corso R, Rampoldi A. Comparison between RF ablation and TACE in the treatment of multifocal hepatocellular carcinoma: preliminary experience. Radiology 2000;217(P): Livraghi T, Solbiati L, Meloni F, Gazelle GS, Halpren EF, Goldberg SN. Percutaneous radio-frequency ablation: complications encountered in a multicenter study of the treatment of focal liver tumors. Radiology 2003;226:

9 S106 Livraghi et al. 15. Matsui O, Kadoya M, Yoshikawa J, Gabata T, Takashima T, Demachi H. Subsegmental transcatheter arterial embolization for small hepatocellular carcinoma: local therapeutic effect and 5-year survival rate. Cancer Chemother Pharmacol 1994;33:S 84 S Nakashima O, Kojiro M. Recurrence of hepatocellular carcinoma: multicentric occurrence or intrahepatic metastases? A viewpoint in terms of pathology. J Hepatobiliary Pancreat Surg 2001;8: Arii S, Teramoto K, Kawamura T, et al. Characteristics of recurrent hepatocellular carcinoma in Japan and our surgical experience. J Hepatobiliary Pancreat Surg 2001;8: Livraghi T, Bolondi L, Buscarini L, et al. No treatment, resection and ethanol injection in hepatocellular carcinoma: a retrospective analysis of survival in 391 patients with cirrhosis. J Hepatol 1995;22: Ryu M, Shimamura Y, Kinoshita T, et al. Therapeutic results of resection, transcatheter arterial embolization and ethanol injection in 3225 patients with hepatocellular carcinoma: a retrospective multicenter study. Jpn J Clin Oncol 1997;27: Yamamoto J, Okada S, Shimada K, et al. Treatment strategy for small hepatocellular carcinoma: comparison of long-term results after ethanol injection and surgical resection. Hepatology 2001; 34: Llovet JM, Funster J, Bruix J. Intention to treat analysis of surgical treatment for early hepatocellular carcinoma: resection versus transplantation. Hepatology 1999;30: Wayne JD, Lauwers GY, Ikai I, et al. Preoperative predictors of survival after resection of small hepatocellular carcinoma. Ann Surg 2002;235: Makuuchi M. Remodeling the surgical approach to hepatocellular carcinoma. Hepatogastroenterology 2002;49: Villa E, Moles A, Ferretti I, et al. Natural history of inoperable hepatocellular carcinoma: estrogen receptors status in the tumor is the strongest prognostic factor for survival. Hepatology 2000; 32: Llovet JM, Real MI, Montana X, et al. Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial. Lancet 2002;359: Lo CM, Ngan H, Tso WK, et al. Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma. Hepatology 2002;35: