A groundbreaking study conducted by Weill Cornell Medicine has utilized machine learning to categorize Parkinson’s disease into three distinct subgroups, potentially revolutionizing patient-specific treatment approaches.
Researchers at Weill Cornell Medicine analyzed data from an existing study and identified three subtypes of Parkinson’s disease: Rapid Pace, Inching Pace, and Moderate Pace. This classification acknowledges the diverse progression of the disease and aims to tailor treatments to the specific needs of patients.
Experts have lauded the findings as logical and promising, but they caution that larger population studies are necessary to refine these models for greater accuracy.
This development follows recent research from Boston University, where an artificial intelligence model predicted the likelihood of developing Alzheimer’s disease. The Weill Cornell study’s findings, published in npj Digital Medicine, are anticipated to aid researchers and clinicians in providing targeted treatments for Parkinson’s disease subtypes.
The Three Subtypes of Parkinson’s Disease
The study analyzed data from 406 participants in the Parkinson’s Progression Markers Initiative (PPMI), an international observational study collecting comprehensive clinical, biospecimen, multi-omics, and brain imaging data. Using a deep-learning model called deep phenotypic progression embedding (DPPE), researchers modeled the multidimensional, longitudinal progression data of the participants.
The classification revealed three subtypes based on the disease’s progression pace:
Rapid Pace (PD-R): Marked by rapid symptom progression, 54 participants (13.3%) fell into this category.
Inching Pace (PD-I): Characterized by mild baseline symptoms and relatively mild progression, this subtype included 145 participants (35.7%).
Moderate Pace (PD-M): Exhibiting mild baseline symptoms with moderate progression, this was the largest group, comprising 207 participants (50.9%).
The authors highlighted the necessity of treating these subtypes as unique sub-disorders within clinical practice, suggesting that such classifications could inform patient stratification and management.
Implications for Targeted Treatment
Identifying specific Parkinson’s disease subtypes allows for more precise clinical approaches. For example, patients with the Rapid Pace subtype might benefit from more aggressive therapeutic strategies and closer monitoring, while those with the Inching Pace subtype may require less intensive management. This knowledge could guide the selection of medications, including repurposing existing drugs like metformin, which the study suggests might be particularly beneficial for the PD-R subtype.
Dr. Clemens Scherzer, a physician-scientist and professor of neurology at Yale School of Medicine, emphasized the need for larger populations to develop and validate such classifiers. He explained that precision medicine aims to predict disease progression and therapeutically intervene ahead of time to prevent complications. Identifying the disease driver in each patient and developing targeted therapeutics are crucial steps in this process.
Dr. Daniel Truong, a neurologist and medical director of the Truong Neuroscience Institute, noted that subtyping Parkinson’s disease is a logical approach. It allows for predictive and preventive healthcare tailored to each subtype, enabling early intervention for rapid progressive patients and more focused clinical trials for new treatments.
Challenges and Future Directions
Steven Allder, a consultant neurologist at ReHealth, agreed that identifying different subgroups would allow for specific treatment plans. He outlined potential treatments for each subtype:
Inching Pace (PD-I): Focus on maintaining quality of life and preventing symptom progression through lifestyle modifications, physical therapy, and possibly neuroprotective drugs.
Moderate Pace (PD-M): Benefit from a combination of pharmacological treatments to manage symptoms and slow progression, such as dopamine agonists, MAO-B inhibitors, or other disease-modifying therapies.
Rapid Pace (PD-R): Early intervention with metformin and other neuroprotective agents could be crucial for managing this rapidly progressing subtype.
Allder also raised concerns about the accessibility of AI models for predicting diseases like Parkinson’s. He emphasized the importance of ensuring that advanced diagnostic tools and treatments derived from AI research are accessible to all patients, especially in under-resourced settings. Additionally, he pointed out potential issues related to data privacy and security, stressing the need for AI models to be validated across diverse populations.
Dr. Scherzer reiterated the significance of large, high-quality, longitudinal data sets of Parkinson’s patients for training and validating AI models. He emphasized that the success of AI in predicting outcomes depends on the size and quality of the input data, underscoring the necessity of further research and trials.
The Weill Cornell study marks a significant step forward in the classification and treatment of Parkinson’s disease, offering hope for more personalized and effective care for patients.