Prognostic Nomograms for Prediction of Recurrence and Survival After Curative Liver Resection for Hepatocellular Carcinoma

In Brief Objective: To develop clinical predictive nomograms generating per-patient numerical probabilities of postoperative recurrence-free and overall survival at specific times. Background: The prognosis after surgical resection is diverse in patients with early-stage hepatocellular carcinoma (HCC). Methods: In a retrospective review, we evaluated data from 1085 mostly early-stage patients newly diagnosed with HCC who were subsequently treated by curative resection. We randomly divided the subjects into derivation (n = 760) and validation (n = 325) samples. Multivariate Cox proportional hazards models were developed and separately validated on the basis of pre- and postoperative clinical and pathological covariates assessed for association with 2-year recurrence and 5-year HCC-specific death. The discriminatory accuracy of the models was compared with traditional tools by analyzing receiver operating characteristic curves. Results: The statistical nomograms built on the basis of sex, serum albumin, platelet count, microvascular invasion, and calculated tumor volume had good calibration and discriminatory abilities, with c-indices of 0.69 (2-year recurrence) and 0.66 (5-year survival), respectively. These models showed satisfactory goodness-of-fit and discrimination abilities in the independent validation cohort (c-index, 0.66 for 2-year recurrence; and 0.67 for 5-year survival). The areas under the receiver operating characteristic curve using our methods were greater than those of conventional staging systems in the validation patients, indicating better discriminatory capability (corresponding c-indices, 0.55–0.56; and 0.55–0.61, respectively). Conclusions: Our simple user-friendly calculators, which present graphically postsurgical prognostic models for recurrence and survival outcomes in patients with curatively resectable HCC, offer useful guidance to clinicians and patients for individually planning recurrence surveillance and adjuvant therapy. We developed new postoperative prognostic nomograms by which the individual magnitude of the potential risk of recurrence and cancer death at specific times in patients with hepatocellular carcinoma can be numerically scored. Our nomograms could be useful as adjuncts in decision making regarding postoperative management in patients undergoing curative resection for early hepatocellular carcinoma.