Machine learning identifies three Parkinson’s subtypes to better track disease progression

The use of machine learning models identified three distinct subtypes of Parkinson’s disease, which “could have immediate implications” in detecting clinical outcomes, researchers reported in NPJ Parkinson’s Disease.

“Prediction of disease and disease course is a critical challenge in the patient counseling, care, treatment and research of complex heterogeneous diseases. Within PD, meeting this challenge would allow appropriate planning for patients and symptom-specific care,” Anant Dadu, of the University of Illinois at Urbana-Champaign, and colleagues wrote.

Dadu and colleagues used unsupervised and supervised machine learning models on 294 cases from the Parkinson’s Disease Progression Marker Initiative to identify patient subtypes and predict disease progression.

A total of 263 cases were validated in an independent cohort. The authors distinguished three distinct disease subtypes with highly predictable progression rates that corresponded to slow, moderate and fast disease progression.

The authors reported projections of disease progression 5 years after initial diagnosis with an average area under the curve of 0.92 (95% CI, 0.95 + 0.01) for the slower progression group, 0.87 + 0.03 for the moderate group and 0.95 + 0.02 for the fast group.

In addition, the authors identified serum neurofilament light as a significant indicator of fast disease progression, among other key biomarkers of interest, they wrote.

“Our data-driven study provides insights to deconstruct PD heterogeneity,” Dadu and colleagues wrote. “This approach could have immediate implications for clinical trials by improving the detection of significant clinical outcomes. We anticipate that machine learning models will improve patient counseling, clinical trial design and ultimately individualized patient care.”