Journal of Clinical Oncology · 2008 · 267 citations · 15 references
MedicineOncologyPediatricsCancer Drug Toxicity ManagementPediatric Clinical PharmacologyPediatric OncologyClinical Trial DesignClinical Trial PhasesClinical TrialsClinical Trials In OncologyPediatric Clinical TrialsStudy DesignClinical Trial SimulationDiscrete Event Simulation ModelsDose-limiting Toxicity AssessmentPediatric Oncology TrialsPediatric Trial DesignPhase I Oncology TrialsSequential Clinical Trial DesignTrial Accrual
The study aims to shorten the conduct timeline of pediatric phase I oncology trials by employing a novel rolling six design, which will be tested prospectively in upcoming Children’s Oncology Group phase I trials. Using discrete event simulation of 1,000 trials based on historical data, the authors compared the traditional 3+3 design with a novel rolling six design that allows concurrent accrual of 2–6 patients per dose level depending on current enrollment, evaluable patients, DLTs, and risk, to assess study timelines, patient numbers, and safety. Simulations showed that the rolling six design reduced average study completion time to 294 ± 75 days versus 350 ± 84 days for the 3+3 design, while maintaining similar DLT rates (≈3.3 ± 1.1 vs 3.2 ± 1.1), indicating a significant decrease in duration without increased toxicity risk.
To shorten the study conduct timeline of pediatric phase I oncology trials by employing a novel trial design.A comparison of the traditional 3 + 3 patients per cohort, phase I trial design with a novel, rolling six design was performed by using discrete event simulation. The rolling six design allows for accrual of two to six patients concurrently onto a dose level based on the number of patients currently enrolled and evaluable, the number experiencing dose-limiting toxicity (DLT), and the number still at risk of developing a DLT. Clinical trial simulations (n = 1,000) were based on historical data and were performed using SAS 9.1.3 (SAS Institute, Cary, NC). Study timelines and patient numbers were determined for each design, and safety was assessed as a function of the number of DLTs observed.In twelve completed historical studies, the median time to study completion was 452 days (range, 220 to 606 days); number of evaluable participants enrolled was 22 (range, 11 to 33), and DLTs occurring per study was three (range, 0 to 5). In 1,000 study simulations, in which the average time to new patient accrual was 10 days, the average +/- standard deviation (SD) time to study completion was 294 +/- 75 days for the rolling six design versus 350 +/- 84 days for the 3 + 3 design, whereas the number of DLTs per study was the same (average +/- SD, 3.3 +/- 1.1 v 3.2 +/- 1.1 for the rolling six and 3 + 3 designs, respectively).The rolling six design may significantly decrease the duration of pediatric phase I studies without increasing the risk of toxicity. The design will be tested prospectively in upcoming Children's Oncology Group phase I trials.
15