Statistical Consideration for Event-Free Survival With Cure Rate in Acute Myeloid Leukemia Studies

Area of Science:

• Hematology
• Clinical Trials
• Biostatistics

Background:

• Event-free survival (EFS) is a primary endpoint in acute myeloid leukemia (AML) studies.
• Induction treatment failure (ITF) is defined as not achieving complete remission (CR).
• Recent FDA guidance (2022) specifies ITF as an event at randomization, acknowledging variable induction periods.

Purpose of the Study:

• To address limitations in conventional sample size calculations for AML trials.
• To develop a new sample size calculation method that accounts for both ITF and a cure fraction.
• To validate the proposed method against existing approaches and simulation data.

Main Methods:

• Decomposition of the log-rank test statistic into ITF and Non-ITF components, similar to Xu et al. (2021).
• Development of a novel sample size calculation formula incorporating ITF and cure fraction.
• Comparison of analytically calculated power with empirical power from simulations across various sample settings.

Main Results:

• The proposed sample size calculation method demonstrated analytical power closely matching empirical power from simulations.
• The new method effectively accounts for the presence of both induction treatment failure and cured patients.
• The approach helps prevent overestimation or underestimation of required sample sizes in AML studies with cure fractions.

Conclusions:

• The developed sample size calculation method provides a more accurate approach for AML clinical trials.
• Accounting for ITF and cure fractions is crucial for robust statistical power and appropriate sample size determination.
• This methodology enhances the reliability of EFS as an endpoint in AML research.