Deep Learning to Differentiate Parkinsonian Syndromes Using Multimodal Magnetic Resonance Imaging: A Proof-of-Concept Study.
Authors
Affiliations (13)
Affiliations (13)
- Univ Toulouse, Inserm, ToNIC, Toulouse, France.
- Sorbonne Université, Institut du Cerveau-Paris Brain Institute-ICM, AP-HP, CNRS 7225, Inserm 1127, Hôpital de la Pitié Salpêtrière, Department of Neuroradiolog, Paris, France.
- The Neuro-Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada.
- Univ. de Bordeaux, CNRS, IMN, UMR 5293, Bordeaux, France.
- CHU Bordeaux, Service de Neurologie des Maladies Neurodégénératives, IMNc, CRMR AMS, NS-Park/FCRIN Network, Bordeaux, France.
- University of Bordeaux, INSERM, BPH, U1219, IPSED, Bordeaux, France.
- Department of Medicine, University of Otago, Christchurch, and New Zealand Brain Research Institute, Christchurch, New Zealand.
- MSA French Reference Center, Univ. Hospital Toulouse, Toulouse, France.
- University of Toulouse, CIC-1436, Departments of Clinical Pharmacology and Neurosciences, NeuroToul COEN Center, NS-Park/FCRIN Network, Toulouse, France.
- University Hospital, Inserm, U1048, Toulouse, France.
- Sorbonne Université, Institut du Cerveau-Paris Brain Institute-ICM, AP-HP, CNRS, Inserm, Hôpital de la Pitié Salpêtrière, Department of Neurology, Paris, France.
- Assistance Publique Hôpitaux de Paris, Service de Neurologie, Hôpital Avicenne, Hôpitaux Universitaires de Paris Seine-Saint-Denis, Sorbonne Paris Nord, NS-PARK/FCRIN Network, Bobigny, France.
- Dynamics and Pathophysiology of Neuronal Networks Team, Center for Interdisciplinary Research in Biology, Collège de France, CNRSUMR7241/INSERM U1050, Université PSL, Paris, France.
Abstract
The differentiation between multiple system atrophy (MSA) and Parkinson's disease (PD) based on clinical diagnostic criteria can be challenging, especially at an early stage. Leveraging deep learning methods and magnetic resonance imaging (MRI) data has shown great potential in aiding automatic diagnosis. The aim was to determine the feasibility of a three-dimensional convolutional neural network (3D CNN)-based approach using multimodal, multicentric MRI data for differentiating MSA and its variants from PD. MRI data were retrospectively collected from three MSA French reference centers. We computed quantitative maps of gray matter density (GD) from a T1-weighted sequence and mean diffusivity (MD) from diffusion tensor imaging. These maps were used as input to a 3D CNN, either individually ("monomodal," "GD" or "MD") or in combination ("bimodal," "GD-MD"). Classification tasks included the differentiation of PD and MSA patients. Model interpretability was investigated by analyzing misclassified patients and providing a visual interpretation of the most activated regions in CNN predictions. The study population included 92 patients with MSA (50 with MSA-P, parkinsonian variant; 33 with MSA-C, cerebellar variant; 9 with MSA-PC, mixed variant) and 64 with PD. The best accuracies were obtained for the PD/MSA (0.88 ± 0.03 with GD-MD), PD/MSA-C&PC (0.84 ± 0.08 with MD), and PD/MSA-P (0.78 ± 0.09 with GD) tasks. Patients misclassified by the CNN exhibited fewer and milder image alterations, as found using an image-based z score analysis. Activation maps highlighted regions involved in MSA pathophysiology, namely the putamen and cerebellum. Our findings hold promise for developing an efficient, MRI-based, and user-independent diagnostic tool suitable for differentiating parkinsonian syndromes in clinical practice. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.