A Proposal for Modification of the Barcelona Clinic Liver Cancer Staging System Considering the Prognostic Implication of Performance Status

Abstract

BACKGROUND/AIMS: Barcelona Clinic Liver Cancer (BCLC) C stage demonstrates considerable heterogeneity because it includes patients with either symptomatic tumors (performance status [PS], 1-2) or with an invasive tumoral pattern reflected by the presence of vascular invasion (VI) or extrahepatic spread (EHS). This study aimed to derive a more relevant staging system by modification of the BCLC system considering the prognostic implication of PS. METHODS: A total of 7,501 subjects who were registered in the Korean multicenter hepatocellular carcinoma (HCC) registry database from 2008 to 2013 were analyzed. The relative goodness-of-fit between staging systems was compared using the Akaike information criterion (AIC) and integrated area under the curve (IAUC). Three modified BCLC (m-BCLC) systems (#1, #2, and #3) were devised by reducing the role of PS. RESULTS: As a result, the BCLC C stage, which includes patients with PS 1-2 without VI/EHS, was reassigned to stage 0, A, or B according to their tumor burden in the m-BCLC #2 model. This model was identified as the most explanatory and desirable model for HCC staging by demonstrating the smallest AIC (AIC=70,088.01) and the largest IAUC (IAUC=0.722), while the original BCLC showed the largest AIC (AIC=70,697.17) and the smallest IAUC (IAUC=0.705). The m-BCLC #2 stage C was further subclassified into C1, C2, C3, and C4 according to the Child-Pugh score, PS, presence of EHS, and tumor extent. The C1 to C4 subgroups showed significantly different overall survival distribution between groups (p<0.001). CONCLUSIONS: An accurate and relevant staging system for patients with HCC was derived though modification of the BCLC system based on PS.