Assessment of objective response rate (ORR) by investigator versus blinded independent central review in pivotal trials of drugs approved for solid tumor indications.

Abstract

e13570 Background: Objective response rate (ORR), defined as the proportion of patients with a complete response or partial response to treatment according to Response Evaluation Criteria in Solid Tumors (RECIST), is the most common endpoint used in pivotal trials supporting FDA approval of cancer drugs for solid tumor indications. Blinded independent central review (BICR) is frequently employed in clinical trials to minimize bias in evaluation of response rate, as historically, assessment of response by investigators (INV) has been shown to overestimate treatment effect. In this study, we analyzed the variability in assessment of ORR between INV and BICR in trials supporting recent Food and Drug Administration (FDA) approvals of drugs for solid tumor indications. Methods: The FDA’s novel drug approvals (2015-2019) were reviewed to identify drugs receiving primary approval for solid tumor indications. Drug approval packages accessed via the Drugs@FDA database and primary publications for the pivotal trials accessed via PubMed were reviewed for investigator-assessed and BICR-assessed ORR. For trials reporting both assessments, the difference between INV and BICR ORR was determined across all study arms. Data are presented using descriptive statistics. Results: A total of 36 drugs received primary approval for the treatment of solid tumors between 2015 and 2019. Of the 40 supporting trials, ORR was the primary endpoint for 21 (52.5%), progression-free survival for 13 (32.5%), and overall survival for 2 (5.0%). ORR was evaluated in 35 of the 40 trials (87.5%). Eight (22.9%) of the 35 trials evaluated INV ORR only, 5 (14.3%) evaluated BICR ORR only, and 22 (62.9%) evaluated both INV and BICR ORR. Among the 22 trials (29 arms in total), the mean difference between BICR- and INV-assessed ORR was -4.3% (95% CI: -6.4, -2.3); the range was -13.1 to 5. INV-assessed ORR was greater than BICR-assessed ORR in 22 of 29 arms (75.9%). The mean difference between BICR- and INV-assessed ORR among the 6 arms representing placebo or active control was -6.0 (95% CI: -11.0, -0.9), compared with -3.9 (95% CI: -6.3, -1.5) among the 23 experimental arms. Conclusions: Compared with BICR, INV overestimated ORR in three-quarters of the trial arms, including those representing control and experimental treatments. Despite this variability, for one fifth of the trials supporting approval of drugs to treat solid tumors, INV was the only method used to assess ORR. For consistency, and the ability to make relative cross-trial comparisons of ORR between agents, BICR should be considered for evaluation of tumor response in all registrational trials.