Prognostic Performance Of Lymph Node Staging Systems In Patients With Resectable Gastric Cancer: Comparison Of Population-Based Nomograms.

Ruihua Xu,Zi-Xian Wang,Miao-zhen Qiu

DOI: https://doi.org/10.1200/JCO.2016.34.15_suppl.e15533

IF: 45.3

2016-01-01

Journal of Clinical Oncology

Abstract:e15533 Background: Two ratio-based lymph node staging systems, the lymph node ratio (LNR) and log odds of metastatic lymph nodes (LODDS), were proposed to refine the prediction of survival, but the use of these systems instead of the metastatic lymph node (MLN) number-based staging for survival prediction among patients with resectable gastric cancer (GC) is not widely accepted. We compared the prognostic performances of the nomograms based on MLN, LNR, and LODDS to predict the 5-year overall survival (OS) in patients with resectable GC. Methods: We analyzed 15320 patients with resectable GC in the Surveillance, Epidemiology, and End Results (SEER) database between 1988 and 2010 for nomogram development. Missing data were handled using multiple imputation. Factors affecting OS were identified by multivariate Cox analysis. When assessed as a continuous covariate, each MLN, LNR, and LODDS variable was incorporated into a nomogram with other significant prognosticators to predict the 5-year OS. The discriminatory abilities of the three nomograms were compared using the concordance indices (C-indices). An independent Chinese dataset (1595 cases) was used as an external validation set. Results: Negative lymph node count was included in the MLN-based nomogram as an independent favorable prognosticator (P < 0.001). The discriminatory abilities of the three nomograms were comparable in the SEER and Chinese datasets (C-indices for the MLN-, LNR-, and LODDS-based nomograms: 0.744, 0.741, and 0.744, respectively in the SEER set; 0.715, 0.712, and 0.713, respectively in the Chinese set; P > 0.152 for all multiple comparisons). The nomograms all showed significant predictive superiorities to the 7th American Joint Committee on Cancer TNM staging (C-indices: 0.713 in the SEER set, P < 0.001 for all; 0.693 in the Chinese set, P < 0.040 for all) and good calibration. The discriminatory abilities of the nomograms maintained comparable even within patients with node-negative GC or insufficient nodes harvested ( < 15 nodes) in both datasets. Conclusions: The MLN-, LNR-, and LODDS-based nomograms are equivalently effective to refine survival prediction for patients with resectable GC.