Photo Credit: DC Studio

The following is a summary of “Automated Imaging Differentiation for Parkinsonism,” published in the March 2025 issue of JAMA Neurology by Vaillancourt et al. 

MRI with disease-specific machine learning may help differentiate PD, MSA, and PSP. A prospective study is needed to confirm its diagnostic value. 

Researchers conducted a retrospective study to evaluate the accuracy of Automated Imaging Differentiation for Parkinsonism (AIDP) using 3-T diffusion MRI and support vector machine (SVM) learning. 

They conducted a prospective multicenter cohort study across 21 sites from July 2021 to January 2024. Patients with PD, MSA, and PSP met established criteria, and their diagnoses were confirmed by 3 blinded neurologists. Participants were assigned to a training or independent testing set. 

The results showed that 316 patients were screened, and 249 (mean age 67.8 [7.7] years; 155 males [62.2%]) met inclusion criteria—99 had PD, 53 had MSA, and 97 had PSP. A retrospective cohort of 396 patients (mean age 65.8 [8.9] years; 234 males [59.1%]) included 211 with PD, 98 with MSA, and 87 with PSP. Patients were assigned to a training set (78%; 104 prospective, 396 retrospective) or an independent testing set (22%; 145 prospective: 60 PD, 27 MSA, 58 PSP; mean age 67.4 [7.7] years; 95 males [65.5%]). The model differentiated PD vs atypical parkinsonism (AUROC 0.96; positive predictive value [PPV] 0.91; negative predictive value [NPV], 0.83), MSA vs PSP (AUROC 0.98; PPV 0.98; NPV 0.81), PD vs MSA (AUROC 0.98; PPV 0.97; NPV 0.97), and PD vs PSP (AUROC 0.98; PPV 0.92; NPV 0.98). AIDP predictions were neuropathologically confirmed in 46 of 49 brains (93.9%). 

Investigators met the primary endpoints, confirming AIDP’s diagnostic value. Results supported its use in diagnosing common parkinsonian syndromes. 

Source: jamanetwork.com/journals/jamaneurology/fullarticle/2831631