Introduction The advantages of neoadjuvant therapy (NAT) in oesophago-gastric (OG) cancer have been proven in randomised trials. Benefits are modest and likely to be restricted to patients who respond well to therapy. The risks may outweigh benefits in earlier stages of disease. This study explores the risks associated with NAT, analyses which patients benefit and investigates whether there is a need for a change in NAT treatment strategy.

Method All patients planned for surgical resection of OG cancer in this unit between 01/2010 and 12/2014 were identified.

Part a) patients were divided according to pre-operative stage. Within each stage, survival was compared in patients undergoing NAT and those undergoing surgery alone. Similar comparisons were made with the NAT group further divided into histological responders and non-responders.

Part b) from the cohort, patients undergoing NAT were identified. Adverse effects (AE) were recorded and graded according to the Common Terminology Criteria for Adverse Events v3.0. The relationships between the presence of AE and a range of variables were investigated along with survival.

Survival analyses were undertaken by plotting Kaplan-Meier curves. Groups were compared using the Log Rank (Mantel-Cox) test.

Results 587 patients were identified from the database.

Part a) only stage III patients had improved survival with NAT vs. surgery alone (P = 0.02). Those in stage II with a histological response to treatment showed a trend towards better survival vs. surgery alone (P = 0.058). In all patients above the threshold for offering NAT (≥stage II), survival was no different in non-responders vs. those having surgery alone. Age and performance status were lower in the NAT group vs. the surgery only group and were both associated inversely with survival. This could mean NAT effects are overestimated and true survival in non-responders to NAT may be poorer than in patients having surgery alone.

Part b) 376 patients received NAT and 88 (23.4%) suffered AE. 69% of these were severe, life-threatening or fatal. NAT course completion rate was lower in patients suffering AE vs. those not (P = 0.003). 20.5% patients with AE did not proceed to resection vs. 6.4% without AE (P = 0.001). Survival was better in patients without AE vs. with AE (P = 0.02) and was better in those undergoing resection (P < 0.001) but was no different in patients completing the course vs. not completing (P = 0.861).

Conclusion Efficacy of NAT is dependent on cancer stage and histological response to NAT. AE are common, often severe and associated with a reduced resection rate and survival. Identifying a means of predicting response to NAT is of great importance and should be used to guide treatment alongside/instead of disease stage.

Disclosure of interest None Declared.