Can recurrence patterns after curative resection for gastric adenocarcinoma (GCa) inform the selection of adjuvant treatment?

Abstract

136 Background: Despite improvements in multidisciplinary care and surgical technique, ≈40% of patients recur after curative resection for GCa. Adjuvant treatment decisions should be informed by the expected timing and site of failure, but accurate data are limited. The aim of our study was to describe and analyze the patterns of recurrence at our center. Methods: Patients who underwent curative intent margin-negative resection from 2006-16 were identified from a prospective GCa database. Date of first detection and site of recurrence were determined. Survival curves were calculated by the KM method. Univariate and multivariable analyses (MVA) were performed to identify predictive variables. Significance was set at p < 0.05. Results: Of 123 patients who met inclusion criteria, median age was 70 yrs, 37% were female, and median follow-up after resection was 24 mos (IQR 11-40). Five-year OS was 72% and RFS 69%. In 34 of 36 patients who recurred, recurrence was detected within 2 yrs. Receipt of adjuvant therapy (Macdonald or MAGIC) was associated with better 5-year OS and RFS than surgery alone (79% vs. 60%, 77% vs. 60%) and delayed time to recurrence (median 10.4 vs. 7.1 mos). Receipt of CRT by the Macdonald protocol appeared to lower the risk of locoregional recurrence vs. surgery alone (see Table) despite routine D2 lymphadenectomy. On MVA, recurrence at any site was predicted by age £ 60y (OR 3.8), male gender (4.0), and AJCC stage > II (4.6). Peritoneal recurrence was predicted only by age £ 60y (3.2). Conclusions: Recurrence after curative resection occurred within 2 yrs of resection. Locoregional recurrence was uncommon in patients who received postoperative CRT. Peritoneal recurrence, which might be abrogated by prophylactic HIPEC at the time of resection, is not well explained by clinicopathologic variables, and novel molecular predictors are required. [Table: see text]