A novel prognostic nomogram is more accurate than conventional staging systems for predicting survival after resection of hepatocellular carcinoma

BACKGROUND: Prediction of survival after resection of hepatocellular carcinoma (HCC) remains difficult. Numerous staging systems have been devised for purposes of risk classification; we sought to identify the optimal staging system to predict postoperative survival. STUDY DESIGN: One hundred eighty-four patients who underwent primary complete resection of HCC at our institution between 1989 and 2002 were classified according to 8 contemporary staging systems. The ability of these systems to predict relative survival for randomly selected pairs of patients was quantified using the Harrel's concordance index. A novel prognostic nomogram was constructed using prognostically relevant variables. RESULTS: After a median followup of 46 months for surviving patients, the median overall survival was 38 months. The concordance indices for the existing staging systems ranged from 0.54 to 0.59. Only the 2002 American Joint Commission on Cancer system demonstrated a concordance index with a 95% confidence interval exceeding 0.5, indicating that the ability of conventional systems to predict relative survival of randomly selected pairs of patients was generally no better than chance. We developed a novel nomogram based on patient age, serum alpha-fetoprotein level, operative blood loss, resection margin status, tumor size, satellite lesions, and vascular invasion. The nomogram demonstrated a markedly superior concordance index of 0.74 (95% CI, 0.68 to 0.80). A separate nomogram for prediction of recurrence-free survival was also generated. CONCLUSIONS: Contemporary staging systems for HCC do not accurately predict postoperative outcomes. Our prognostic nomogram provides a mechanism for accurate prediction of survival and risk stratification and will require validation at other hepatobiliary centers.