Metastasis remains one of the major challenges before hepatocellular carcinoma (HCC) is finally con-quered. This paper summarized a decade's studies on HCC metawstasis at the Liver Cancer Institute of Fudan University. We have established a sptepwise metastatic human HCC model system, which included a metastatic HCC model in nude mice (LCI-D20), a HCC cell line with hgh metastatic potential (MHCC97), a relatively low metastatic potential cell clone (MHCC97L) and several stepwise high metastatic potential cellclones (MHCC97H, HCCLM3, and HCCLM6) from their parent MHCC97 cell. Endeavors have been made for searching human HCC metastasis-related chromosomes/proteins/genes. Monogene-based studies revealed that HCC invasion/metastasis was similar to that of other solid tumorsa, and the biological characteristics of small HCC were only slightly better that that of large HCC. Using comparative genomic hybridization (CGH), fluorescence in situ hybridization (FISH), genotyping, cDNA microarray, and 2-dimensional gel electrophore-sis, we obtained some interesting results. In particular, in collaboration with the National Institute of Health (NIH) in the United Staters, we generated a molecular signature that can classify metastatic HCC patients, identified ostewopntin as a lead gene in the signature, and found that genes favoring metastasis progression were initiated in the primary tumors. We also found that chromosome 8p deletion, particularly in the region of 8p23, was associated with HCC metastasis. Cytokeratin 19 was identified as one of the proteins, which was found in MHCC97H, but not in MHCC97L cells. Experi-mental interventions using the high metastatic nude mice model have provided clues for the prevention of HCC metastasis. Translation from workbench to bedside demonstrated that serum VEGF, microvessel density, and p53 scoring may be of value for the prediction of postoperative metastatic recurrence. Interferon alpha proved effective for the prevention of recurrence both experimentally and clinically. In conclusion, HCC metastasis that probably initiated in the prim,ary tumor is a multigene-involved, multistep, and changing is a multigene-involved, multistep, and changing process. The further elucidation of the mechanism underlying HCC metastasis will provide a more solid basos for the prediction and prevention of the metastatic recurrence of HCC.

A highly metastatic LCI?D20 model of human hepatocellular carcinoma (HCC) was obtained via orthotopic implantation of a histologically preserved metastatic tumour selected from I of 30 surgical specimens, and has been maintained for 18 passages in nude mice. All mice with transplanted tumours exhibited 100% transplantability and metastatic ability, as well as various manifestations reminiscent of tumour behaviour in HCC pa-tients. These included: local growth, regional invasion, spontane-ous metastasis to liver, lymph nodes and lungs, and peritoneal seeding, Histological characteristics of the LCI-D20 tumour were similar to those of the original tumour. Karyotypic analysis revealed heteroploid cells. Immunohistochemically, expression of AFP and HBXAG was shown, Our nude mouse model with its high metastatic rate and short latency period could be an interesting tool for the study of human HCC.

The poor prognosis of hepatocellular carcinoma (HCC) was partly a result of the majority of unresectable HCCs in clinical patients. Fortunately, with the progress of regional cencer therapics and multimo-dality trealent, some of the licalized unresectable HCCs were converted to resectable ones. During the period 1960-1994, 72 of the 663 patients with surgically verified unresectable HCCs have been converted to resectable ones. During the period 1960-1994, 72 of the 663 patients with surgically verified unreseetable HCCs have been converted to rescctable. Sucessful cytoreduction with median diameter reduced from 10 cm to 5 cm was mainly a result of the triple or double combination treatment with hepatic artery ligation, hepatic artery cannulation with infusion, radioimmunotherapy. and fractionated regional radiotherapy. The interval hetween the first operation and the sequential reection was 5 months. The operative mortality was 1.4% for sequential resection, and the 5-year sursival was 62.1%. Analysis of factors influencing sequenlial resection rate revealed HCCs that were single nodule, well encapsulated, situated at right lobe or hepatic hilum, associated with micromodular cirrhosis, and treated with triple or double combination modalities had higher sequential resection rate as compared to their counterparts. Analysis of factors influencing sursival after sequential resection revealed that HCCs with a solitary lumor confined in one lobe, without tumor embolus, and without residual cancer in specimen of sequential resection, had longer survival. It is suggesled that localized unreseelable, solitary, well encapsulated, reght lobe or hilar HCC, associated with micronodular cirrhosis, will be good candidates for cytoreduction and sequential resection; and HCCs with unilateral involvement, without tumor embolus, and with complete necrosis of tumor after multimodality trcatment favored better prognosis.

Subclinical hepatocellular carcinoma (SCHCC) is defined as HCC without obvious HCC symptoms and signs. During 1958-1991, 391 patients with SCHCC were analyzed. In the entire series, 1) 67.3% was detected by natural population sereening using alpha-fetoprotein (AFP) serosurvey, while the others were discovered by high-risk population screening or regular health checkup using AFP and/or ultrasonography (US); 2) AFP>20 ug/L was found in 77.6% of patients; 3) serum hepatitis B surface antigen (HBsAg) was positive in 68.9; 4) associated liver cirrho-sis occurred in 89.1%; 5) the median tumor size was 5 cm, and limited resection was performed in the majority (71.3%); 7) re-resection for sub-clinical recurrence was done in 44 patients; and 8) cytoreduction and sequential resection was carried out in 13 patients with unresectable SCHCC. Comparison between SCHCC and clinical HCC (n=1, 251) revealed higher resectability(81.4% vs, 46.8%), lower operative mortality (1.9% vs. 6.0%), and higher 5-year survival (entire series; 50.7% vs. 20.6%; resection: 60.5% vs. 36.8%). It is concluded that the study of SCHCC has resulted in marked improvement of ultimate outcome of HCC; screening in high-risk populations using AFP and/or US, limited resection, and re-resection for subclinical recurrence are some of the key features.

Hepatocellular carcinoma(HCC) is one of the most common and aggressive human malignan-cies. Its high mortality rate is mainly a result of intra-hepatic metastases. We analyzed the ex-pression profiles of HCC samples without or with intra-hepatic metastases. Using a supervised machine-learning algorithm, we generated for the first time a molecular signature that can clas-sify metastatic HCC patients and identified genes that were relevant to metastasis and patient survival. We found that the gene expression signature of primary HCCS with accompanying metastasis was very similar to that of their corresponding metastases, implying that genes fa-voring metastasis progression were initiated in the primary tumors. Osteopontin, which was identified as a lead gene in the signature, was over-expressed in metastatic HCC; an osteopon-tin-specific antibody effectively blocked HCC cell invasion in vitro and inhibited pulmonary metastasis of HCC cells in nude mice. Thus, osteopontin acts as both a diagnostic marker and a potential therapeutic target for metastatic HCC.

Postoperative recurrence of human hepatocellular carcinoma (HCC) is the major issue that must be addressed to further improve prognosis. This study was undertaken to investigate the effects of interferon-alfa-1b (IFN-a-1b) on recurrent tumor and metastasis after curative resection in nude mice bearing an HCC xenograft with high metastatic potential. Tumor tissues from LCI-D20, a metastatic model of HCC in nude mice, were orthotopically implanted in 105 nude mice. Eleven days later, 64 mice underwent curative resection of liver tumors. IFN-a at different doses was administered subcutaneously to mice with or without resection. In mice without resection, when comparison was made among control, IFN 7.5 3 106 U/kg/day, 1.5 3 107 U/kg/day for treated groups, and 3 3 107 U/kg/day; tumor volume was 8,475 mm3 6 2,636 mm3, 7,963 mm3 6 3,214 mm3, 769 mm3 6 287 mm3, and 13 mm3 6 9 mm3; incidence of lung metastasis being 100%, 80%, 40%, and 0%; life span was 45 6 4 days, 53 6 8 days, 81 6 6 days, and 105 6 24 days, respectively. In mice with curative resection, when comparison was made among control, IFN 5 3 105 U/kg/day, 1 3 106 U/kg/day, 4 3 106 U/kg/day, 7.5 3 106 U/kg/day, 1.5 3 107 U/kg/day, and 3 3 107 U/kg/day for treated groups; incidence of recurrent tumor was 100%, 100%, 87.5%, 100%, 87.5%, 62.5%, and 12.5%; lung metastasis being 100%, 75%, 87.5%, 50%, 62.5%, 0%, and 0%, respectively. IFN-a inhibited neovascularization induced by LCI-D20 tumor specimens implanted into the micropocket of nude mice corneas. In conclusion, high-dose and longterm therapy with IFN-a dose-dependently inhibits tumor growth and recurrence after resection of HCC. The effect of IFN-a may be attributed to antiangiogenesis in this experiment. These results provide potential clinical implication, particularly for the prevention of recurrence after curative resection of HCC.