Long term outcomes of immune checkpoint inhibition in oesophageal cancer

Abstract Background Phase III studies have shown a survival benefit for the addition of immune checkpoint inhibitors (ICIs) to conventional chemotherapy for patients with gastro-oesophageal cancer. Large scale studies of chemo-immunotherapy prior to and after surgery are under way. We studied the clinical and biological effects of adding the PD-L1 inhibitor durvalumab (D) to peri-operative chemotherapy or chemoradiotherapy in a multi-centre phase II trial and used multiple phenotyping approaches to identify features associated with benefit. Methods Patients suitable for surgery were given D for 4 weeks, then one of Oxaliplatin and Capecitabine (CapOx), fluorouracil, leucovorin, oxaliplatin and taxotere (FLOT), or chemoradiotherapy (CXRT, CROSS regimen) were given with D before surgery, and D was continued for up to 12 further infusions postoperatively. Toxicity, response, and survival were recorded until death, for 3 years or until June 2022 when the study ended. Comparisons were made with propensity score matched patients treated over the same period in the one of the participating sites. Tumour phenotyping was with single cell and bulk sequencing approaches. Results 11 pts received D-CapOx, 9 D-FLOT and 15 D-CXRT. Treatment side effects were as expected. Only 2 out of 35 patients did not have surgery, due to progressive disease on D-CXRT. Two (5.7%) and 6 (17.1%) patients had complete or partial pre-op responses (RECIST) and 3 (8.6%) had progression prior to surgery. Tumour regression at surgery (Mandard 1-3) occurred in 24/33 pts. 2-year survival was 81.8%, 77.8% and 77.8% respectively in the 3 arms. After propensity score matching, there was a survival advantage after addition of ICI (p=0.047). In selected patients we also demonstrate molecular features associated with response. Conclusions Adding D to standard neo-adjuvant regimens for gastro-oesophageal cancer is well tolerated. Survival times and response rates exceeded those expected for the regimens used (2-year survivals with FLOT 68% and CROSS 67%) in this non-randomised multi-centre trial. There was a survival advantage from the addition of ICI when compared to a matched cohort of patients treated in one of the study centres over the same period. Specific pre-treatment tumour phenotypes were associated with durable cure.