Risk stratification and predictive risk-scoring model for lymph node metastasis in early gastric cancer

We are increasingly experiencing difficulty in deciding whether to perform gastrectomy after noncurative endoscopic resection of early gastric cancer (EGC) for patients at high risk for surgery. If the differences in risk for lymph node metastasis (LNM) on the basis of noncurative status are understood, the decision becomes easier. The present study aimed to stratify the LNM risk and develop and validate a risk-scoring model for predicting LNM.

By retrospectively reviewing 3131 patients with solitary EGC who underwent gastrectomy with lymphadenectomy at our institution between July 1997 and May 2013, LNM risk was stratified and a risk-scoring model was developed on the basis of the identified independent risk factors for LNM. The scoring was validated using 352 other surgically resected EGC cases. The discriminatory accuracy of the scoring was measured by area under receiver operating characteristic curve (AUROC).

LNM was detected in 386 of 3131 cases. LNM risk in each subgroup, stratified by the identified independent risk factors, such as tumor size, depth, histological type, ulcerative findings, and lymphovascular involvement, considerably varied from 0 % to >50 % even among the current guidelines' noncurative subgroups. An 11-point scoring model was built, and AUROCs were 0.84 (95 % confidence interval, 0.82-0.86) and 0.82 (0.75-0.88) in the development and validation sets, respectively.

The present study revealed detailed LNM risk stratification data, and developed and validated an 11-point scoring model.