Phase I and Phase II Objective Response Rates are Correlated in Pediatric Cancer Trials: An Argument for Better Clinical Trial Efficiency.

Although many phase I trials report tumor response, formal analysis of efficacy is deferred to phase II. We reviewed paired phase I and II pediatric oncology trials to ascertain the relationship between phase I and II objective response rate (OR%). Single-agent phase I trials were paired with corresponding phase II trials (comparable study drug, dosing schedule, and population). Phase I trials without efficacy data or a matching phase II trial were excluded. OR% was tabulated for all trials, and phase II authors' subjective conclusions regarding efficacy were documented; 35 pairs of trials were analyzed. The correlation between phase I and II OR% was 0.93. Between phase II studies with a "positive" conclusion versus a "negative" one, there was a statistically significant difference in mean phase I OR% (32.0% vs. 4.5%, P<0.001). Thirteen phase II studies were undertaken despite phase I OR% of 0%; only 1 had a "positive" conclusion, and none exceeded OR% of 15%. OR% are highly correlated between phase I and II pediatric oncology trials. Although not a formal measure of drug efficacy, phase I OR% may provide an estimate of phase II response, inform phase II study design, and should be given greater consideration.