To determine whether metabolic or pathological response to preoperative chemotherapy can predict the relapse-free survival of gastroesophageal adenocarcinoma patients treated on a perioperative chemotherapy protocol. The prospectively collected data of a recently reported phase II trial of perioperative DCF chemotherapy (docetaxel/cisplatin/5-fluorouracil) were analyzed. Median relapse-free survival (RFS) was compared with the Wilcoxon rank-sum test between responders and non-responders according to defined metabolic (reduction in maximum standard uptake value of at least 35 %) and pathological (greater than 50 % tumor regression or ypN(0) status) criteria. A double-sided p value equal or inferior to 0.05 was considered significant. Patients were followed for a median of 807 days (95 % CI: 607-896). RFS was 576 days in metabolic non-responders versus not reached in metabolic responders (p 0.009) and 562 days in ypN+ versus not reached in ypN(0) patients (p 0.045). No statistically significant RFS difference was seen between low and high pathologic responders classified according to tumor regression criteria, although a trend was observed in favor of high pathologic responders. Simple metabolic and pathologic criteria used for the assessment of response to the preoperative part of perioperative chemotherapy can help to estimate the outcome of gastroesophageal adenocarcinoma patients.