Modified Gastric Cancer AJCC Staging with a Classification Based on the Ratio of Regional Lymph Node Involvement: A Population-Based Cohort Study.

PURPOSE: This study evaluated the additional value of lymph node ratio (LNR) when used in combination with the 8th American Joint Committee on Cancer (AJCC) staging system for gastric cancer (GC) to establish a new LNR-based AJCC (rAJCC) staging system. METHODS: We searched the Surveillance, Epidemiology, and End Results database to identify patients who had undergone surgery to resect nonmetastatic GC during the period from 2004 to 2013. Recursive partitioning analysis was used to identify the optimal cutoff points for LNR, and an LNR-based N classification system was designed accordingly. The two staging systems were compared in terms of discriminatory ability, as measured by the concordance index. The likelihood ratio Chi square test was used to assess prognostic homogeneity. RESULTS: A total of 13,027 patients were included in the final analysis. All of the included patients, who belonged to the 8th AJCC IIIA category, could be further classified into the rIIB, rIIIA, rIIIB, and rIIIC subgroups using the proposed rAJCC classification system. A difference in 5-year overall survival rate was found between patients classified as having rIIA disease and those classified as having rIIIC disease (66.7% vs. 5.1%). The rAJCC staging system was superior to the 8th AJCC staging system in terms of discriminatory capacity and prognostic homogeneity. CONCLUSIONS: A new rAJCC staging system is proposed, with prognostic superiority to the 8th AJCC staging system for patients with GC. The rAJCC staging system may serve as a useful tool in clinical practice.