Nomogram to Predict Survival of Patients With Recurrence of Hepatocellular Carcinoma After Surgery

BACKGROUND & AIMS: We aimed to establish and validate a nomogram to predict survival at 2 and 5 years after recurrence of hepatocellular carcinoma (HCC) in patients who have undergone curative resection. METHODS: We developed a nomogram using data from a training cohort of 638 patients (most with hepatitis B virus infection) with recurrence of HCC after curative resection at Sun Yat-sen University Cancer Center, in Guangzhou, China from 2007 through 2013. The median follow-up time was 39.7 months. Patients were evaluated every 3-4 months for the first 2 years after resection and every 3-6 months thereafter. The nomogram was based on variables independently associated with survival after HCC recurrence, including antiviral treatment; albumin-bilirubin grade and alpha-fetoprotein level at recurrence; time from primary resection to recurrence; size, site, number of recurrences; and treatment for recurrence. We validated the nomogram using data from an independent internal cohort of 213 patients treated at the same institution and an external cohort of 127 patients treated at 2 other centers in China, from 2002 through 2009. The predictive accuracy of the nomogram was measured using Harrell's concordance index (C index) and compared with the Barcelona Clinic Liver Cancer staging system of recurrence. RESULTS: Our nomogram predicted survival of patients in the training cohort with a C-index of 0.797 (95% CI, 0.765-0.830)-greater than that of the Barcelona Clinic Liver Cancer staging system for recurrence (C-index score, 0.713; 95% CI, 0.680-0.745) (P < .001). This nomogram accurately stratified patients into subgroups with predicted long, medium, and short survival times: the proportions of patients in each group who survived 2 years after HCC recurrence were 91.2%, 67.6%, and 23.8%; the proportions of patients in each group who survived 5 years after HCC recurrence were 74.9%, 53.3%, and 9.1%. Our nomogram predicted patient survival times with C-index scores of 0.756 (95% CI, 0.703-0.808) in the internal validation cohort and 0.747 (95% CI, 0.701-0.794) in the external validation cohorts. CONCLUSIONS: We developed a nomogram to determine the probability of survival, at different time points, of patients with recurrence of HCC (most with hepatitis B virus infection), after curative resection and validated it internally and externally.