Objective. The conditions upon which endoscopic resection (ER) can be considered curative for submucosal invasive gastric cancer remain controversial; thus, unnecessary surgery is sometimes performed after ER. Our purpose is to evaluate the significance of several clinicopathological factors for predicting the absence of lymph node (LN) metastasis of submucosal invasive gastric cancer and thus determining cases in which ER can be considered curative.

Patients and methods. The study group comprised 220 patients with submucosal invasive gastric cancer that was resected surgically or endoscopically. Patients treated by ER underwent additional surgical resection. The presence of LN metastasis was evaluated in all patients, retrospectively.

Results. LN metastasis was detected in 37 (16.8%) of the 220 patients. Independent risk factors for LN metastasis were width of submucosal invasion >6000 mu m, lymphatic involvement, undifferentiated type at the deepest invasive portion, depth of submucosal invasion >1000 mu m, and tumor diameter >30 mm. The group of 36 patients with submucosal invasion to a depth of <= 1000 mu m, tumor diameter <= 30 mm, differentiated type as the dominant histologic type, and absence of vessel involvement was entirely free of LN metastasis (95% confidence interval, 0-8.0%).

Conclusions. Taken together, the five independent risk factors may allow expansion of the criteria for determining whether ER for submucosal invasive gastric cancer has been curative.