Background 
Parkinson’s disease and Alzheimer’s Disease and Related Dementia (ADRD) are categories of aging disorders with complex and heterogeneous symptomatology. The clinical manifestations of these neurological disorders are characterized by heterogeneity in age at onset, disease duration, rate of progression, and constellation of cognitive, motor, and non-motor features. Due to these variable presentations, counseling of patients about their individual risks and prognosis is limited. There is an unmet need for predictive tests that facilitate early detection and characterization of distinct disease subtypes as well as improved, individualized predictions of the disease course. The emergence of machine learning to detect hidden patterns in complex, multi-dimensional datasets provides unparalleled opportunities to address this critical need. We have developed predictive models for Alzheimer’s and Parkinson’s related disorders. Our efforts have focused on: (1) developing predictive models and identification of biomarkers which are multi-modal (clinical, biological, genetic, and imaging data), and (2) relying on unsupervised machine learning techniques to identify disease subtypes that are multidimensional. 

Methods and findings  
We used unsupervised and supervised machine learning approaches for subtype identification and prediction. We used machine learning methods on comprehensive, longitudinal clinical data from the Parkinson Disease Progression Marker Initiative (PPMI) (n=328 cases) and Alzheimer's Disease Neuroimaging Initiative (ADNI) (n= 248) to identify patient subtypes and to predict disease progression. The resulting models were validated in an independent, clinically well-characterized cohort from the Parkinson Disease Biomarker Program (PDBP) (n=112 cases). Our analysis distinguished three distinct disease subtypes with highly predictable progression rates, corresponding to slow, moderate and fast disease progressors, after 24, 48, and 72 months from baseline. We achieved highly accurate projections of disease progression four years after initial diagnosis with an average Area Under the Curve of 0.93 (95% CI: 0.96 ± 0.01 for PDvec1, 0.87 ± 0.03 for PDvec2, and 0.96 ± 0.02 for PDvec3). We have demonstrated robust replication of these findings in the independent validation cohort.

Conclusion 
These data-driven results enable clinicians to deconstruct the heterogeneity within their patient cohorts. This knowledge could have immediate implications for clinical trials by improving the detection of significant clinical outcomes that might have been masked by cohort heterogeneity. We anticipate that machine learning models will improve patient counseling, clinical trial design, allocation of healthcare resources and ultimately individualized clinical care.

10 - Predicting Onset, Progression, and Clinical Subtypes of Alzheimer’s and Parkinson’s Disease from Genomic and Longitudinal Clinical Data