Interim decisions in oncology trials.

Abstract

e12530 Background: Interim analyses are commonly performed in oncology clinical trials. For a phase III trial, interim monitoring rules are in general pre-specified for an independent data monitoring committee (DMC) to allow for the possibility of early stopping. Challenges and controversies exist when interim monitoring and subsequent decisions alter the course of the trial. Methods: Abstracts and primary publications of ASCO plenary session presentations for the last 10 years (2003-2012) were reviewed. Interim analyses related to the study design, conduct, endpoints, DMC decision, and reporting were reviewed. Mid-trial adaptations not limited to interim results were also summarized. Results: Fifty-two presentations were identified. Forty-one were randomized controlled trials; almost all were phase III trials. Interim monitoring and analyses were discussed in the primary manuscripts of 31 trials. The objectives of these interim analyses were safety monitoring, inference for superior efficacy, or futility. Fourteen trials crossed interim boundaries and were declared by the DMCs to have met the study objectives on efficacy endpoints such as progression-free survival and overall survival (OS). Mid-trial adaptations included dropping an arm or selecting a dose based on interim safety results. Other trial adaptations, motivated by issues such as slower accrual rate, change in treatment paradigm, or unexpected longer trial duration, included modifications of trial objectives, primary endpoints, sample sizes, and unplanned analyses. Conclusions: Interim monitoring is an important component of clinical trials but is not consistently reported. Decisions made after the primary endpoint has crossed the interim boundary have an impact on subsequent inferences of other endpoints (e.g. OS). Greater transparency and practical guidelines are needed to ensure continued trial integrity and valid inferences post interim.