Application of value framework to phase III trials of immune checkpoint inhibitors in esophageal and gastric cancer.

Abstract

e18817 Background: Recent clinical trials testing use of immune-checkpoint inhibitors in esophageal and gastric carcinomas have shown both positive and negative results. Given these mixed signals, we aimed to assess outcomes and tolerability of key subgroups using the ASCO Net Health Benefit Score (NHBS) and ESMO Magnitude of Clinical Benefit Scale (MCBS). Methods: A detailed search for phase III trials investigating use of FDA-approved anti PD-1 or anti PD-L1 drugs in esophageal and gastric cancer trials was completed using www.clinicaltrials.gov followed by primary literature identification on PubMed. These studies were then independently selected, reviewed, and scored using the ASCO NHBS and ESMO MCBS by two investigators, and any discrepancy was mutually resolved. The ASCO NGBS scores were compared by primary site of cancer (esophageal vs gastric) and PD-L1 expression using the Mann-Whitney test. The ESMO-MCBS grading were compared using Fisher's Exact test. Results: Of the 45 records identified, 15 clinical trials were included. Of them, six were primarily esophageal cancer trials, and nine were primarily gastric cancer trials. Of all 15, 10 stratified their analysis based on PD-L1 expression. The ASCO NHBS score was significantly higher for esophageal cancer than gastric cancer (mean 40, range 20 – 56.6 vs. mean 12, range -1.1 – 18.4, p < .01). No difference, however, was observed in survival and response rate outcome endpoints between the two groups. Similarly, the ESMO MCBS was also higher for esophageal cancer group than gastric cancer (p < .05). Additionally, the median scores were higher in the patients with high PD-L1 expression vs low PD-L1 expression (mean 37, range 11.2 – 66.6 vs. mean 14, range -19.5 – 43.6, p < .05). Conclusions: In this cohort of drug registration trials of immune checkpoint inhibitors for upper GI cancers, we showed the ASCO NHB scores and ESMO scores were consistently higher among esophageal cancer trials than gastric cancer trials and in those with high PD-L1 expression than those with low PD-L1 expression. Oncologists should carefully discuss overall clinical value and expectations of immunotherapy benefit with their patients per cancer histology and PDL1 expression, exercising caution in cases of gastric adenocarcinoma with low PD-L1 expression.[Table: see text]