Alzheimer?s disease (AD) is characterized by heterogeneity in sex, race, APOE ?4 status, and tau. APOE ?4 status has been used to stratify AD patients in clinical trials (e.g., the aducanumab trial). Women with AD have faster cognitive decline compared to men with AD from longitudinal studies. African Americans decline faster than Whites on memory tests and visuospatial functioning. Furthermore, we know that the changes in cognitive functions are negatively associated with tau levels. In addition to the heterogeneity, another challenge facing AD trials is the follow-up time in study designs. In the recent aducanumab trial, the dose of aducanumab was increased during the course of the study for APOE ?4 positive patients, but their follow-up times remained unchanged. The proposed project will respond to PAR-19-070: Research on Current Topics in Alzheimer?s Disease and Its Related Dementias. We will develop adaptive designs to allow the modi?cation of follow-up time for patients with dose change in Aim 1. The published results from the aducanumab trial will be used in simulation studies to compare the statistical performance of the proposed adaptive designs with the existing designs without follow-up time change. Our simulation results indicated that our proposed adaptive designs guarantee the type I error rate and power, while the existing designs do not.
In Aim 2, we will develop new optimal composite scores for each subpopulation by using the baseline data and the rate of change over time to better understand the differences in cognition and trajectories of decline. We will develop one optimal composite score for each subpopulation strati?ed by sex, race, APOE ?4 status, and tau, based on data from the ADNI study. We will demonstrate how best to manage statistically the effects of demographic, genetic, and biomarker factors on cognitive ability of AD patients. The optimal composite scores are expected to be more sensitive to detect cognition change compared to the measures we traditionally use. In 2019, the Food and Drug Administration released a ?nal guidance on developing enrichment strategies in clinical investigations to promote innovation in drug development. Treatment effect heterogeneity exists among patients with different characteristics. Identifying subpopulations who are more likely to respond to a new treatment at a given dose would signi?cantly increase the success rate of AD trials and avoid the types of issues that occurred in the aducanumab trial. Adaptive enrichment designs for AD trials will measure the treatment effectiveness of each subpopulation at the interim analysis for futility based stopping, and can save sample sizes compared to the existing designs. We will add a constraint on the probability of ?wrong? stopping for futility to avoid stopping the enrollment for a possible effective treatment on a subpopulation. This project will provide new statistical tools for AD research to ef?ciently identify individuals at risk of AD and quickly detect disease progression for AD patients with different characteristics.
Treatment effect heterogeneity exists among patients with different characteristics including demographic, ge- netic, and biomarker risk factors. The proposed optimal composite scores will increase our ability to measure outcomes in AD and can be applied directly to personalized AD treatments. The proposed adaptive designs will improve the ?exibility of trial designs which would provide greater ef?ciency in AD drug development.
