|
Abstract:
|
Cancer biomarker discoveries typically involve utilizing patient specimens collected as a previously completed clinical trial. In practice, there is often desire to preserve biospecimens for studies that are most likely to yield useful information. We propose a two-stage adaptive design for time-to-event endpoints which terminates the biomarker study in a futility interim if the model performance is unsatisfactory at an early stage. Stage one involves testing whether the measure of discrimination for survival models (C-index) exceeds a pre-specified threshold. We describe the computation of cross-validated C-index and evaluation of the statistical significance using re-sampling techniques. Stage two involves an independent model validation. Our simulation studies show that when the signature performance is undesirable (under the null), the proposed design maintains type I error at the nominal level and has high probabilities of terminating the study early. Under the alternative hypothesis, power of the design depends on the true event proportion, sample size, and targeted improvement in the discriminant measure. Some practical aspects of the proposed method will be discussed.
|
