Introduction Patients clinically staged with > T2 or node positive (>cT2/cN+) oesophageal or OGJ cancer are routinely considered for neoadjuvant chemotherapy prior to attempted curative resection (NA). Patients with cardiovascular comorbidity are often precluded from NA and the effect on survival in a contemporary setting is not clear.

Method Single centre retrospective study of all patients undergoing attempted curative therapy for >cT2/cN+ mid or lower oesophageal or OGJ adenocarcinoma or squamous cell carcinoma diagnosed between 2001 and 2013. Uni and multivariable survival analyses were performed using a prospectively-maintained audit database.

Results A total of 371 patients were included and 289 (78%) commenced NA comprising predominantly 2 cycles of cisplatin and 5-fluorouracil (n = 264; 91%) with 274 (95%) undergoing subsequent resection. Patients undergoing surgery alone (S) were significantly older (mean = 8 years, p < 0.001) but were well matched with patients undergoing NA for clinical stage, gender and histology.

Based on intention to treat, NA was associated with a significantly improved median overall survival compared to S (29 months compared to 21 months; P = 0.008). Recurrence-free survival did not, however, differ between NA and S (p = 0.585). The 90 day in-hospital mortality rate was higher in S compared to NA (10% vs. 3%; P = 0.011), likely reflecting the increased comorbidity of these patients. When patients dying in hospital within 90 post-operative days were excluded, the overall survival benefit of NA over S was reduced to a median of 2.7 months (p = 0.04).

Pathological tumour size, differentiation, pT, pN and R stage were associated with survival (all P < 0.01) but did not differ significantly between patients undergoing NA or S suggesting a limited down-staging effect of NA.

To assess histological response to chemotherapy, the Mandard tumour regression grade (TRG) was determined, blinded to outcome, in a subset of 111 patients undergoing NA. Only 21 patients (19%) exhibited a histological response to NA (TRG I-III) and TRG was not associated with survival (P = 0.08).

Conclusion Resection carries a significantly higher perioperative mortality risk in patients with >cT2/cN+ oesophageal or OGJ cancer and comorbidity precluding neoadjuvant therapy. In clinically stage-matched patients, neoadjuvant chemotherapy was associated with a significant survival benefit over surgery alone.

A low histological response rate to chemotherapy was noted and in those patients surviving to discharge the overall survival benefit of NA was limited to a median of <3 months. More effective therapies are needed.

Disclosure of interest None Declared.