Using Minimization for a Complex Trial Design
A multi-center phase II trial was conducted in just over 100 patients to compare metabolic changes in the prostate after administration of two treatments for locally advanced prostate cancer. The changes of interest were assessed by a biopsy taken at different randomized times and by magnetic resonance spectroscopy/magnetic resonance imaging (MRS/MRI) in about one third of randomly selected patients. In order to protect against chance imbalances, patients were additionally stratified by their level of prostate-specific antigen (PSA) at baseline. The methodology adopted to implement this complex trial design was minimization for the various randomized factors.