Clustering and prediction of disease progression trajectories in Huntington's disease: An analysis of Enroll-HD data...

Area of Science:

• Neuroscience
• Genetics
• Computational Biology

Background:

• Huntington's disease (HD) is a progressive neurodegenerative disorder with variable symptom severity and progression rates.
• Diagnosis of manifest HD typically relies on genetic confirmation and/or motor symptoms, despite earlier cognitive and behavioral signs.
• Understanding factors influencing HD progression is crucial for effective patient management.

Purpose of the Study:

• To model the longitudinal natural history of disease progression in individuals with manifest Huntington's disease.
• To identify key features predicting disease trajectory and group patients based on progression rates.
• To inform the development of prognostic models for individualized clinical care in HD.

Main Methods:

• Retrospective analysis of the global, observational Enroll-HD study data (NCT01574053).
• Unsupervised machine learning (k-means clustering) to group 4,961 individuals with manifest HD into rapid, moderate, and slow progressors.
• Supervised machine learning (XGBoost) to identify predictive features of disease trajectory.

Main Results:

• Individuals with manifest HD were successfully clustered into three distinct progression groups: rapid (25.3%), moderate (45.5%), and slow (29.2%).
• The top predictor for cluster assignment was the cytosine adenine guanine-age product score (CAG repeat length x age).
• Other significant predictors included years since symptom onset, history of apathy, body mass index, and age at enrollment.

Conclusions:

• The CAG-age product score is a key factor influencing Huntington's disease progression rate.
• Identifying distinct progression trajectories and predictive factors can enhance understanding of HD.
• Prognostic models based on these findings could significantly improve individualized clinical care planning and disease management for HD patients.