The Prognostic Ability of Major Hepatocellular Carcinoma Staging Systems Is Improved by Including a Treatment Variable.

ABSTRACT

BACKGROUND AND AIMS: There has been significant debate regarding which hepatocellular carcinoma (HCC) staging system is best able to predict survival. We hypothesized that the prognostic ability of the Barcelona Clinic Liver Cancer (BCLC) and Hong Kong Liver Cancer (HKLC) systems would be improved with the addition of an explicit treatment variable. METHODS: We performed an analysis of a prospectively enrolled cohort of 292 patients undergoing 532 treatment episodes for HCC from 2006 to 2014. BCLC, standard nine-stage HKLC (HKLC9), and modified five-stage HKLC (HKLC5) for each treatment episode were assessed. Overall survival and time to disease progression were calculated for the initial treatment, re-treatment, and overall treatment cohorts. We compared the performance of various prognostic models including staging system alone, treatment alone, and staging system plus treatment using the corrected Akaike information criterion and Harrell's C statistic. RESULTS: The BCLC, HKLC5, and HKLC9 systems were significant predictors of survival and time to progression for all treatment cohorts (log rank test, p < 0.001). The addition of a treatment variable significantly improved (p < 0.01) the prognostic ability of the survival and time to progression models compared with those containing only the BCLC or HKLC stage across all treatment cohorts other than survival in re-treatment for BCLC (p = 0.094). CONCLUSIONS: Adding a treatment variable to major HCC staging systems improves their ability to predict survival and time to progression in initial treatment, re-treatment, and overall.