Prognosis of hepatocellular carcinoma associated with Child class B and C cirrhosis in relation to treatment: a multivariate analysis of 411 patients at a single center.

Background/purpose: Given that the prognosis of patients with hepatocellular carcinoma (HCC) complicating severe cirrhosis remains uncertain, particularly with regard to various therapeutic strategies, we have evaluated the prognosis in a series of patients with homogeneous diagnostic and therapeutic histories.

Methods: From 1990 to 1998, 411 consecutive HCC patients associated with Child class B and class C cirrhosis who did not have lymph node or distant metastasis were treated by partial hepatectomy (PH; n = 48), percutaneous ethanol injection (PEI; n = 105), transcatheter arterial chemoembolization (TACE; n = 189), chemotherapy, or supportive care (chemo/supportive; n = 69). Univariate survival curves were estimated. The Cox model, stratified by the treatment groups, was used for multivariate analysis.

Results: As of January 1999, 305 patients (74.2%) had died. Overall median survival was 23.4 months. There were statistically significant differences between the survival times of patients receiving PH or PEI and TACE, as compared with those receiving chemo/supportive care. According to multivariate analysis, the independent predictive survival factors were: albumin level (> or = 3.0 g/dl), esophageal varices (i.e., absence), tumor size (< or = 3.0 cm), tumor number (solitary), and alpha-fetoprotein (AFP) level (<400 ng/ml). According to the total number of risk factors and the median survival, all patients were divided into four subgroups. For the score 0 group (no risk factor group), 3- and 5-year survival rates were 83.1% and 68.0% for PH, and 87.5% and 62.3% for PEI, respectively. In the score 1-2 group (one or two risk factors), survival rates at 3 and 5 years were 53.1% and 40.3% for PH, 54.8% and 33.2% for PEI, and 35.4% and 22.8% for TACE, respectively. For patients with a score of 3 or more, there were no differences among the treatment groups, excluding those with chemo/supportive care.

Conclusions: These findings indicate that, in HCC patients with complicating Child B and C cirrhosis, PEI and PH should be considered first for subgroups of patients with scores (risk factors) of 0-2, as an acceptable survival rate was obtained in such patients. Therefore, the advantages and disadvantages of these therapies regarding tumor size and location should be counterbalanced. In patients with a score of 3 or more, TACE, when possible, could be a first choice because of its applicability and its adjuvant nature with respect to other therapies such as liver transplantation.