Verification of T-plus staging system for colorectal cancer based on nomogram analysis of single center’s 25-year follow-up.

516 Background: We proposed T-plus staging system which abandon the criterion to discriminate stage II/III by lymph node status and strengthens weighting of the T stage according to cluster analysis of the summary survival data of SEER. In this study, this principle of the T-plus staging system was verified with single center data by nomogram analysis. Methods: The 1,099 patients with colorectal cancer diagnosed before 2005 were analyzed to build a novel staging system based on nomogram (nomo-staging system). The remaining 981 patients diagnosed after 2005 were used to test the performance of nomo-staging system, T-plus staging system and the 7th edition TNM staging system by Akaike information criteria (AIC), Harrell’s c-index and the area under the curve (AUC) of Receiver Operating Characteristic to predict 5-year overall survival. A smaller AIC, higher c-index and higher AUC indicated a better staging system. Results: The nomo-staging system and T-plus staging system were listed in Table 1. The validation found that the 7th edition TNM staging system showed the weakest performance. For AIC, both nomo-staging system and T-plus staging system showed smaller value than the 7th edition TNM staging system (3214.912 vs. 3224.643 vs. 3229.810). For Harrell’s c-index, both the T-plus staging system and nomo-staging system showed higher value than the 7th edition TNM staging system (0.6814 vs. 0.6787 vs. 0.6778). For AUC to predict 5-year OS, the T-plus staging system and nomo-staging system showed slightly higher value than the 7th edition TNM staging system (0.6944 vs. 0.6924 vs. 0.6913). Conclusions: We propose replacement of lymph node status as the criterion to discriminate colorectal cancer stage II/III with greater weighting of the T stage. [Table: see text]