Joint modeling of short-term and long-term outcomes for interim decision making in oncology phase II clinical trials
The goal of phase II clinical trial in oncology is to screen for promising drug candidates to be studied in phase III clinical trials. Poor designs of phase II oncology trials have contributed to the expansive failures of drug development in oncology. One way to address the problem is to use information more efficiently in early drug development. In particular, joint modeling of short-term tumor response and long-term progression free survival (PFS) time was proposed to make better decision in phase II oncology trials. In the setting of Bayesian adaptive randomization design for phase II oncology trial, a joint modeling approach was used to achieve more personalized treatment allocation for each new patient based on patient level data at baseline and predicted short-term tumor response. In the setting of interim analysis, early tumor response data was used in a joint modeling approach to provide additional information on the long-term PFS outcome. Simulation studies were conducted to demonstrate the improved operation characteristics over conventional modeling approaches.^
Lin, E, "Joint modeling of short-term and long-term outcomes for interim decision making in oncology phase II clinical trials" (2014). Texas Medical Center Dissertations (via ProQuest). AAI3689780.