Multicentric Assessment of the Hong Kong Liver Cancer Staging System in Chinese Patients Following Transarterial Chemoembolization.

Abstract

We aimed to validate the performance of the hepatitis B-based Hong Kong Liver Cancer (HKLC) staging system compared with the Barcelona Clinic Liver Cancer (BCLC) staging system in Chinese hepatocellular carcinoma (HCC) patients treated with conventional transarterial chemoembolization (TACE) as the initial treatment. The study was approved by the Institutional Review Boards at all participating centers. This retrospective study included 715 patients with HCC who underwent TACE as the initial treatment between January 2008 and December 2016 at three Chinese institutions. All of the patients calculated HCC stage using 5-substage HKLC (HKLC-5), 9-substage HKLC (HKLC-9), and the BCLC system. Based on overall survival (OS), these three staging systems' performance on treatment outcome prediction were compared using C statistic, Akaike information criterion (AIC), area under the receiver operating characteristic curve (AUC), linear trend Chi-square, likelihood ratio Chi-square, and calibration plots, respectively. The median OS was 10.1months. Compared with the BCLC system, the HKLC system, especially HKLC-9, showed better performance on survival prediction (HKLC-9: C = 0.689, AIC = 6646.162; HKLC-5: C = 0.683, AIC = 6662.663; BCLC: C = 0.680, AIC = 6654.146), homogeneity (likelihood ratio Chi-square: HKLC-9 = 232.38, HKLC-5 = 215.87, and BCLC = 224.39, P < 0.001), and calibration (R2: HKLC-9 = 0.923, HKLC-5 = 0.916, and BCLC = 0.914). HKLC-9 outperformed on AUC at 6-, 12-, and 24-month survival prediction than HKLC-5 and BCLC. BCLC showed better performance on monotonicity (linear trend Chi-square: HKLC-9 = 121.641, HKLC-5 = 117.389, and BCLC = 125.752; P < 0.001). Combining survival prediction, discrimination, and calibration, the HKLC, especially HKLC-9 system, performed better for Chinese patients treated with TACE than the BCLC system.