Treatment and prognostic factors in patients with hepatocellular carcinoma.

INTRODUCTION: Hepatocellular carcinoma is a leading cause of death from cancer worldwide. Survival of patients depends on tumor extension and liver function, but yet there is no consensual prognostic model. AIMS: To evaluate the influence on survival of pretreatment parameters (clinico-laboratorial, liver function, tumor extension, Okuda and Cancer of the Liver Italian program (CLIP) staging) and treatment modalities. METHODS: We retrospectively analyzed 207 patients, diagnosed between 1993 and 2003. The initial treatment was: surgery--six patients; radiofrequency ablation--21; percutaneous ethanol injection--29; transarterial chemoembolization--49; tamoxifen--49; supportive care alone--53. Factors determining survival were assessed by Kaplan-Meier method and Cox regression models. RESULTS: Median survival was 24 months. In univariate analysis, Child-Pugh classification and Model for end-stage liver disease (MELD) score, portal vein thrombosis (PVT), tumor size, number of lesions, Okuda and CLIP scores were all associated with prognosis (P < 0.001). Alpha-fetoprotein levels were not predictive of survival. Independent predictors of survival were ascites, bilirubin, PVT and therapeutic modalities (P < 0.001). In early stage hepatocellular carcinoma (HCC), survival was similar for both percutaneous ablation modalities, either radiofrequency or ethanol injection (P = NS). In advanced HCC, survival was better in patients receiving tamoxifen than supportive care alone (P < 0.001). CONCLUSION: This study reinforces the importance of baseline liver function (Child-Pugh classification and MELD score) in the survival of patients with HCC, although staging systems allowed the stratification of patients in different prognostic groups. Ascites, bilirubin and PVT were independent pretreatment predictors of survival. All treatments influenced the patient's outcome, whether in early or advanced stages.