A1786
Title: Characterizing longitudinal sMRI biomarker trajectories in Alzheimer's disease using the DSMM
Authors: Kaidi Kang - Wake Forest University School of Medicine (United States) [presenting]
Abstract: Characterizing longitudinal trajectories of sMRI-derived biomarkers across Alzheimer's Disease (AD) progression is critical for understanding Disease dynamics, yet remains challenging due to the fragmentary nature of available data: existing datasets capture only brief windows of an individual's decades-long Disease course. This is compounded by limited availability of repeated sMRI measurements. Together, these restrict standard analyses to participants with sufficient imaging follow-up and observed clinical events, and introduce selection bias and reduced statistical power. To address these limitations, the double anchoring events-based sigmoidal mixed model (DSMM) was extended to characterize longitudinal sMRI-derived biomarkers in AD. The DSMM employs Disease-relative time scales anchored to incident AD diagnosis and enables inclusion of participants without an observed conversion event, broadening the analyzable study population. Using harmonized multi-cohort data from the Alzheimer's Disease Sequencing Project Phenotype Harmonization Consortium (ADSP-PHC), imaging trajectories are aligned relative to AD onset across participants regardless of conversion status. This strategy mitigates selection bias, enhances statistical power, and yields clinically interpretable biomarker estimates along the Disease continuum, enabling formal comparison of the temporal ordering of multimodal biomarker changes in AD.
