A simple prognostic scoring system for patients with unresectable hepatocellular carcinoma treated by chemo-embolization

BACKGROUND: Hepatocellular carcinoma (HCC) comprises heterogeneous groups of patients with differing outcomes. METHODS: In order to attempt to identify patient sub-sets, we retrospectively examined the records of 750 patients with biopsy-proven unresectable HCC, who were treated with hepatic artery chemo-embolization and were followed till death. We used the Cox proportional hazard model, as it was neutral with respect to prejudged cut-off between short and long survival. RESULTS: On univariate analysis, we found a shorter survival to be associated with male gender, presence of cirrhosis, portal hypertension, portal vein thrombosis and elevations of bilirubin, GGTP, ALKP, AFP, DCP, PT and albumin. Five factors were found to be statistically significant (p<0.05) on multivariate analysis, namely presence of cirrhosis or ascites and elevations of AFP, ALKP or GGTP. We developed a simplified scoring system based upon the sum of the hazard ratios of each of these five factors. By combining the two factors with the heaviest hazard ratios from our multivariable analysis, namely AFP (+ = >100 ng/mL) and ALKP (+ = >100 IU/mL), we found a simple parsimonious prognostic tool, which segregated the patients into survival groups, namely AFP- ALKP-; either AFP+ or ALKP+; and AFP+ ALKP+; these three groups corresponded to a 24-month survival of 70%, 32% and 12%, respectively. CONCLUSION: we found that only two lab functions, AFP and ALKP levels, in our large HCC patient cohort undergoing hepatic artery chemo-embolization, had prognostic significance.