Attempted salvage resection for recurrent gastric or gastroesophageal cancer.

The purpose of this study was to determine the outcome of surgery for patients with recurrent gastric or gastroesophageal cancer. We queried records from 7,459 patients who presented with gastric or gastroesophageal cancer to our institution from 1973 through 2005 to identify those for whom resection of recurrent disease had been attempted. We assessed the associations between various clinicopathologic factors and resectability with logistic regression analysis and between clinicopathologic factors and overall survival (OS) with the Cox proportional hazards model. Sixty patients underwent attempted resection for recurrent cancer. In 31 cases (52%), recurrent disease proved unresectable at laparotomy. Factors associated with the ability to undergo re-resection included neoadjuvant treatment prior to initial resection [odds ratio (OR) 12.2, 95% confidence interval (CI) 1.9-75.6] and having an isolated local recurrence (OR 5.1, 95% CI 1.3-20.5). Of the 29 patients who underwent re-resection, 14 required adjacent organ resection, and 6 required interposition grafting. Three- and 5-year OS rates for all 60 patients were 21% and 12%, respectively; median follow-up time was 23 months. Median OS for patients undergoing resection was 25.8 months (95% CI 17.1-49.8) versus 6.0 months (95% CI 4.0-10.5) for unresectable patients (P < 0.001). Initial tumor location at the gastroesophageal junction was associated with diminished OS [hazard ratio (HR) 2.8, 95% CI 1.2-6.5] and ability to undergo resection of recurrence was associated with improved OS (HR 0.2, 95% CI 0.1-0.6). We conclude that surgical resection of select patients with recurrent gastric or gastroesophageal cancer can result in improved OS but often requires adjacent organ resection or interposition graft placement.