Background

After publication of the MAGIC trial results, preoperative chemotherapy is increasingly used to treat advanced gastric cancer before resection. Tools for measuring response must be assessed.

Methods

We identified all patients with gastric cancer treated with neoadjuvant chemotherapy and R0 resection between 1991 and 2005 from a prospective database. Patients receiving preoperative radiation were excluded. Histologic response to treatment was graded from 0% to 100% by a single pathologist. Kaplan-Meier survival analysis was performed to identify the relationship between response and outcome and to identify factors predictive of disease-specific survival (DSS). Multivariate analysis was performed to identify independent predictors.

Results

A total of 168 patients underwent R0 resection after receiving neoadjuvant chemotherapy. Thirty-three percent of tumors were at the gastroesophageal junction. Cisplatin-based therapy was used for 68% of patients. Twenty-two percent of patients had a >50% pathologic response to treatment. Median follow-up after resection for all patients was 25 months. Median DSS for all patients was 33 months. Three-year DSS improved from 44% to 69% with at least a 50% histologic response (P = .01). Factors associated with decreased DSS included positive nodes at resection, pT3 tumor or greater, high grade, perineural or vascular invasion, and <50% response. Multivariate analysis identified nodal status and perineural or vascular invasion as independent predictors of survival.

Conclusions

Posttreatment nodal status and perineural or vascular invasion at resection, but not graded histologic response, independently predict DSS after neoadjuvant chemotherapy and surgical resection of gastric cancer.