ORR is commonly used as the primary endpoint in Ph2t. ICI have different mechanisms of action to chemotherapy or molecular targeted agents (MTA). The validity of ORR as a surrogate for progression-free survival (PFS) and overall survival (OS) with ICI is uncertain and may differ by tumor type. We performed a meta-analysis of randomized controlled trials (RCTs) in advanced solid cancers that compared ICI to chemotherapy, MTA or placebo to address this question.
We performed a literature search to determine the current Ph2t designs used in ICI trials. Efficacy data from single-arm trials and RCTs were extracted. Amongst the RCTs, correlations between ORR odds ratio (OR) with PFS hazard ratio (HR) and OS HR were examined for between randomized arms comparisons. Correlations within ICI treatment arms of the RCTs between ORR with PFS and OS rates were also studied. Using data from the RCTs, multivariable models that examined the relationships between ORR, 6-month PFS and 12-month OS rates were developed and their predictive performances validated in the single-arm trials.
Of 87 Ph2ts identified, most were single arm design (68%), and only 10% were RCTs with concurrent standard of care arms. ORR was the most common (60%) primary endpoint and PFS was uncommon (8%). A total of 20 RCTs (4 Ph2t and 16 phase 3 trials) with mature data were examined. There were 25 treatment comparisons in 8 different tumors (non-small cell lung cancer 44%, melanoma 24%). For RCTs in all tumors, the correlations (r) between ORR OR with PFS HR, ORR OR with OS HR, and PFS HR with OS HR were 0.63, 0.57 and 0.42 respectively. Within the ICI arms, r between ORR with 6-month PFS, ORR with 12-month OS, and 6-month PFS with 12-month OS were 0.37, 0.08 and 0.74 respectively. In the single-arm trials dataset, we were able to accurately predict 12-month OS using the actual 6-month PFS with the multivariate model developed from our RCTs dataset. Conversely, when ORR was used to predict 6-month PFS or 12-month OS, there was poor agreement between actual and predicted results.
These data do not support the use of ORR as a surrogate for OS in ICI trials. In future ICI Ph2t, 6-month PFS should be the primary endpoint rather than ORR.
Clinical trial identification
Legal entity responsible for the study
M. Friedlander: Receives personal fees and grants with an advisory board position with AstraZeneca. Does not have any patents planned, pending or issued, broadly relevant to the work. All other authors have declared no conflicts of interest.