The prevalence of RCTs in PARP and immune-checkpoint blockade combination trials: An appraisal of clinical evidence generation in oncology.

Abstract

e23006 Background: PARP-inhibitors were developed to promote lethal genomic instability in cancers with homologous recombination repair (HRR) defects, ultimately receiving approval for patients with BRCA gene mutations. At the tumor microenvironment, DNA instability is linked to higher expression of immune activating signals such as PDL1. PDL1 is one of the primary targets for immune-checkpoint inhibitors (ICIs), receiving approval for numerous cancer types. In-vitro studies support mechanisms of possible synergy for PARP and ICIs in combination. Despite both therapy classes being available to oncologists for nearly a decade, the clinical benefit for these agents together remains unclear. Our study aims to appraise the global clinical trial landscape for trials testing combination PARP and PD1/PDL1 regimens. Methods: Search terms for all known approved or unapproved PD1/PDL1 and PARP inhibitors associated with trials registered on the National Library of Medicine’s website, clinicaltrials.gov, were included in the cohort as long as one interventional arm administered both agents. Qualitative and quantitative trial data features were collected from the Cancer Research Institute (CRI) Immuno-Oncology Intelligence Database and complemented with manually sourced information from trial registries and publications. Contribution of Effect (COE) was defined as trials with two or more arms that compare our combination regimen of interest to PARP or an ICI alone. The presence of randomization, eligibility for randomized-controlled trial (RCT) status as well as COE was determined by three reviewers separately with disagreements settled by committee decision. Results: In total, we found 188 clinical trials that enrolled 39,268 patients. A minority of 17 trials were phase I not considered as RCTs, leaving 37,827 patients across 171 trials for enrollment in phase II trials and beyond. Of these 171 trials, 80 (47 %) were single-arm studies without active comparator. Forty-nine (29%) used randomized assortment, 40 (23%) met criteria for an RCT and 33 (19%) met criteria for RCT with COE. In total, these 33 trials enrolled 14,803 patients. Of the 33, 17 (10%) enrolled patients with cancers that are approved for PARP blockade by the FDA and 11 (6%) employed genetic testing for HRR defects. Sponsorship ranged from 16 trials by industry, 6 by cooperative oncology group, 6 by academic institutions and 5 by the National Cancer Institute. Conclusions: Our study results highlight that the global clinical trial apparatus directs significant resources and patients towards trials with lower standards of evidence generation. Across sponsor types, there was a considerable number of trials that met the rigor of an RCT with COE. These findings show both promise and areas of improvement within the trial landscape to generate evidence relevant to the use of PARP and ICIs in combination.