Identifying Sex-Specific Sub-phenotypes of Alzheimer's Disease Progression Using Longitudinal Electronic Health...

Area of Science:

• Neuroscience
• Genetics
• Data Science

Background:

• Alzheimer's Disease (AD) disproportionately affects women, highlighting the need for sex-specific research.
• Understanding sex-based differences in AD progression is critical for developing targeted interventions.

Purpose of the Study:

• To uncover sex-specific Alzheimer's Disease (AD) sub-phenotypes and progression trajectories.
• To leverage longitudinal electronic health record (EHR) data for personalized AD risk stratification.

Main Methods:

• Developed a long short-term memory (LSTM)-based autoencoder for temporal patient representation.
• Utilized hierarchical agglomerative clustering on learned embeddings to identify sub-phenotypes.
• Analyzed sex-stratified AD cohort from the OneFlorida+ Clinical Research Consortium.

Main Results:

• Successfully identified five distinct sex-based AD sub-phenotypes with high confidence.
• Demonstrated unique AD progression patterns for males and females.
• Revealed differences in clinical characteristics and comorbidities between sexes in AD development.

Conclusions:

• The identified sex-specific sub-phenotypes offer valuable insights into AD heterogeneity.
• Findings support the advancement of personalized AD intervention and treatment strategies.
• Highlights the importance of sex as a biological variable in neurodegenerative disease research.