Study Rationale: Parkinson’s disease (PD) has several subtypes. Identification of these subtypes in individuals with PD is important for understanding disease progression, developing new therapies, and determining optimal care. Changes in small structures or molecular signals in the brain are potential biomarkers that can be used to determine PD subtypes. We have developed image-enhancement techniques that increase resolution and contrast, while decreasing noise, in MRI scans used to identify such structures and signals in the brain. We will evaluate how the use of these image-enhancement techniques improves our ability to classify individuals with PD into clinically relevant subtypes.
Hypothesis: This study will evaluate the impact of image-enhancements techniques on the quality of measurements obtained from brain MRI scans and assess the use of those measurements to identify subtype in PD patients.
Study Design: We will utilize MRI data collected through the Parkinson’s Progression Markers Initiative (PPMI) study. Images will be analyzed and enhanced using 3D deep-learning super-resolution techniques. Measurements from these images will be combined with clinical information to produce models for use in classifying PD subjects into clinically relevant subtypes.
Impact on Diagnosis/Treatment of Parkinson’s disease: This project will increase our understanding of how to use MRI data to classify PD by subtype. This classification is important for understanding disease progression, developing new therapies and determining optimal patient care.
Next Steps for Development: Successful execution of this study would allow the construction of image-analysis pipelines and associated models that could be used to provide clinicians with predictions of subtype for individuals with PD.