Use of a novel index, the A-index, and its associated nomogram to predict overall survival rates after resection of primary hepatocellular carcinoma

Background: Several international staging or scoring systems don't accurately predict overall survival (OS) after resection of primary hepatocellular carcinoma (PHCC). Therefore, we attempted to overcome this limitation by constructing the A-index and its associated nomogram. Methods: We selected 672 patients who underwent curative resection of PHCC between January 2007 and February 2015 at the first affiliated hospital of the Wenzhou medical university. These subjects were randomly divided into the training (n = 470) and the validation group (n = 202) according to the ratio of 7:3. Results: We prepared the nomogram using eight independent risk factors including the A-index (calculated by 100 x aspartate transaminase /albumin /albumin) in the training cohort. The concordance index (C-index) of the nomogram for both training and validation set was similar in indicating the OS rate. The nomogram showed the strongest predictive power for the 1-year, 3-year, and 5-year OS, with the area under the ROC curve being 0.8182, 0.7892, and 0.7669, respectively. Correction curves showed consistent performance for both groups, stratification of the Kaplan-Meier curve was significant (P < 0.001), and decision curve analysis (DCA) showed the superiority of nomograms considering clinical effects. Conclusions: The predictive power of the nomogram integrating the A-index for OS was optimal.