Study Rationale: Medical imaging is a promising technique to help with the early diagnosis, understanding, and treatment of Parkinson's disease. Indeed, using advanced image acquisition techniques and software tools, researchers can derive insightful measurements from such images. However, measurements provided by different software tools or acquisition techniques are not always consistent, which raises concerns about their reliability. This project will investigate new methods to reconcile disparate measures obtained from different types of image analyses, resulting in more reliable techniques to characterize and predict Parkinson's disease.
Hypothesis: Our hypothesis is that dissimilar MRI-based measurements of brain atrophy provided by different software tools or measured in different datasets can be aggregated using machine learning to provide more reliable insights on Parkinson's disease.
Study Design: We will replicate the most promising image-based measures related to Parkinson's disease found in the existing literature and we will re-analyse them using multiple different software tools and multiple different data sets. We will then aggregate the outcome of these different analytical conditions using machine learning or advanced statistics, and determine whether the resulting aggregation improves the reliability of the initial measurements.
Impact on Diagnosis/Treatment of Parkinson’s disease: This project will map the influence of image analysis software and data acquisition conditions on the reliability of MRI-based measurements for Parkinson's disease. It will inform clinicians on the reliability of existing imaging studies, and help researchers design more reliable imaging studies.
Next Steps for Development: If successful, this project will provide a review of the reliability of existing MRI-based measures for Parkinson's disease investigations. It will also provide new MRI measures based on the aggregation of existing ones. The most reliable ones, if any, could be further examined in a clinical context. This project will also result in a set of practical guidelines to evaluate the reliability of future MRI-based measures, and therefore provide insights on their clinical applicability.