JAMA Neurol
AI-driven MRI analysis enhances parkinsonism diagnosis
Automated Imaging Differentiation for Parkinsonism (AIDP) offers a non-invasive, scalable approach to improve diagnostic precision for Parkinsonian disorders, potentially enhancing patient care and treatment outcomes.
Study details: The prospective, multicenter AIDP cohort study, conducted from July 2021 to January 2024 across 21 Parkinson Study Group sites in the U.S. and Canada, evaluated the discriminative performance of AIDP using 3-T diffusion MRI and support vector machine learning. The study included 249 patients (mean age, 67.8 years; 62.2% male) with Parkinson disease (PD), multiple system atrophy (MSA), and progressive supranuclear palsy (PSP), confirmed by three independent, blinded neurologists specializing in movement disorders.
Results: AIDP demonstrated high accuracy, calculated by area under the receiver operating characteristic curve (AUROC), in distinguishing the following:
- PD vs. atypical parkinsonism: 0.96
- MSA vs. PSP: 0.98
- PD vs. MSA: 0.98
- PD vs. PSP: 0.98
Positive predictive values (PPV) and negative predictive values (NPV) were also robust, with PPV ranging from 0.91 to 0.98 and NPV from 0.81 to 0.98. Neuropathological confirmation in a subset of autopsy cases showed 93.9% accuracy.
Source:
Vaillancourt DE, et al. (2025, March 17). JAMA Neurol. Automated Imaging Differentiation for Parkinsonism. https://pubmed.ncbi.nlm.nih.gov/40094699/
