Leveraging digital twins for sample size reduction, powering secondary endpoints or subgroups in a simulated TRAILBLAZER‐ALZ 2 trial

Due to the reduced variability in endpoints when using digital twins as covariates in statistical analyses, the TRAILBLAZER-ALZ 2 trial could have enrolled up to 24% fewer control arm participants to draw the same conclusions about treatment efficacy as previously reported for the study's primary outcome, the integrated Alzheimer Disease Rating Scale (iADRS) at 18 months. Alternatively, digital twins could be incorporated to increase power up to 15% for secondary endpoints. Potential sample size reductions and power gains are reported in Table 1. Similarly, subgroup analyses would gain power [Table 2] CONCLUSION: By including digital twins as a supercovariate in the analysis of a simulated donanemab trial, we demonstrate digital twins ability to reduce overall sample size or add power to secondaries and subgroups not factored into initial sample size calculations.