Design and Analysis considerations due to delayed treatment effects observed with Immuno-Oncology agents
*Brent McHenry, Bristol-Myers Squibb 

Keywords: Delayed separation, Kaplan-Meier curve, immuno-oncology agent

Delayed separation in Kaplan-Meier curves for progression free survival has been observed recently for several Immuno-Oncology agents. These observations have impacts on designs and analyses of future studies, including potential reduction in power for standard tests, consideration of alternative testing procedures and timing of interim and/or final analyses. The impact on power of standard log-rank tests for PFS analyses, as a function of the length of the delayed treatment effect will be assessed and compared with the power for alternative methods of comparison, e.g. weighted log-rank tests. The timing of interim and final analyses will also be discussed.