Challenges in the design and conduct of phase III brain tumor therapy trials

Twenty years ago a phase III trial by the Brain Tumor Study Group demonstrated that surgery and radiation were better treatment for malignant glioma than surgery alone.¹ Maximal surgical resection and conventional radiotherapy quickly became the standard of care for patients with high-grade glioma. Since then many other treatments, most often cytotoxic drugs,² have been added to surgery and radiation, and compared with standard therapy in randomized controlled trials(RCTs). None has unequivocally prolonged tumor control or patient survival.³ While these failures speak largely to the resilient nature of malignant glioma, certain issues surrounding the initiation and conduct of phase III trials of brain tumor treatment have surfaced in recent years and warrant discussion before the next generation of antiglioma therapies are tested. Some RCTs were based on inconclusive phase II data and may have been initiated prematurely, while others had divergent entry criteria, inadequate control for prognostic factors, and different outcome measures. Moreover, few studies evaluated quality of life (QOL), toxicity, or cost in a comprehensive way and none considered patient preferences for treatment. In this report we describe methods of assessing the results of phase II trials as a means of selecting treatments that most warrant phase III evaluation, and discuss challenges in the design and conduct of RCTs of new therapies for malignant glioma.

When is an RCT warranted? When are the results of a phase II study sufficiently compelling to warrant a randomized trial? Randomized controlled trials are initiated when new treatments are found to have antiglioma activity in phase II evaluation. Phase II trials are of two types:(1) those that measure tumor response and (2) those that measure tumor control. In studies of response, treatment is given to patients with visible tumor, and the rate and duration of the response is recorded. …