A nomogram for predicting recurrence after complete resection for thymic epithelial tumors based on the TNM classification: A multi-institutional retrospective analysis.

ABSTRACT

OBJECTIVES: This study aimed to compare the predictive ability between the Masaoka-Koga (M-K) staging system and the 8th TNM staging system for the recurrence of thymic epithelial tumors (TETs). In addition, a nomogram was developed on the basis of the proposed TNM classification to predict individual recurrence rate. METHODS: A retrospective study was performed on 445 patients who underwent complete resection (R0) of TETs between January 2000 and February 2013. Concordance index (C-index) was used as a statistical indicator to quantify the prediction power of the prediction models. RESULTS: In multivariate analysis, tumor stage and WHO classification were independent recurrence factors in a predictive model on the basis of M-K and TNM stage. The TNM model showed higher C-index than the M-K model (0.837 vs 0.817). The nomogram, on the basis of the TNM model, revealed a highly predictive performance, with a bootstrap-corrected C-index of 0.85 (95% CI, 0.76 to 0.93). CONCLUSIONS: A predictive model based on the 8th TNM stage was slightly better than that based on M-K stage with respect to recurrence after R0 of TETs. The proposed nomogram could be applied to estimate the individual recurrence rate and make decisions for proper surveillance.