Maier, Alexia R.

Objective Pre‐mortem diagnosis of parkinsonism is often challenging due to atypical presentations, overlapping syndromes, and co‐pathologies. This study aimed to develop a machine learning‐based algorithm predicting neuropathology in parkinsonism using chronological clinical presentations, which has previously been underexplored. Methods Clinical information was automatically abstracted from medical records of the Mayo Clinic Brain Bank using fine‐tuned Generative Pre‐trained Transformer 4 models. Patients who developed parkinsonism within 3 years of disease onset were included. Six machine learning models were trained with age, sex, family history, and 197 clinical presentations paired with onset information to predict neuropathologic diagnoses, including co‐pathologies. Results Among 7,825 donors, 949 met inclusion criteria, representing 9 neuropathologic categories: Lewy body disease (LBD; n = 128), LBD with Alzheimer's disease (AD; n = 136), progressive supranuclear palsy (PSP; n = 303), PSP with AD (n = 56), PSP with LBD (n = 27), multiple system atrophy (MSA; n = 120), corticobasal degeneration (CBD; n = 99), AD (n = 43), and frontotemporal lobar degeneration (FTLD; n = 37). The CatBoost algorithm achieved an area under the receiver operating characteristic curve of 0.83 across the 9 diagnostic categories at 3 years after onset. Important predictors included age at onset, restricted eye movement, and tremor. The model remained robust to incomplete data, requiring only 23 of 200 parameters for reliable predictions with an area under the curve of 0.80. The algorithm was implemented into a user‐friendly program providing diagnostic probabilities with visualizations of parameter contributions. Interpretation This neuropathology‐confirmed diagnostic algorithm provides a cost‐effective and interpretable screening tool for parkinsonism, bridging biomarker testing and molecular‐targeted therapies. ANN NEUROL 2026;99:1405–1414

Objectives Multiple system atrophy (MSA) is classified into parkinsonian (MSA‐P) and cerebellar (MSA‐C) phenotypes based on predominant motor features, with corresponding pathological classifications of striatonigral (SN) degeneration and olivopontocerebellar (OPC) atrophy; however, disease progression patterns remain poorly understood. We aimed to identify novel MSA subtypes based on neuronal loss patterns using unsupervised machine learning. Methods We applied the Subtype and Stage Inference (SuStaIn) algorithm to analyze neuronal loss patterns in 167 autopsy‐confirmed MSA. Neuronal loss was semi‐quantitatively assessed in 5 brain regions: putamen, substantia nigra, pontine nucleus, inferior olivary nucleus, and cerebellar Purkinje cells. Obtained subtypes were validated by clinicopathological information. Results Three distinct subtypes were identified: SN‐early subtype (54% [86/160]) with initial SN neuronal loss, OPC‐early subtype (28% [44/160]) with early OPC involvement, and SN‐OPC‐synchronous subtype (19% [30/160]) with concurrent early neuronal loss in both systems. These subtypes corresponded well with clinical phenotypes. The SN‐OPC‐synchronous subtype exhibited significantly shorter survival (median 6.2 years vs 6.9 and 7.4 years in SN‐early and OPC‐early subtypes; p = 0.0049), more frequent rapid progression (57% vs 24% and 41% in SN‐early and OPC‐early subtypes; p = 0.0048), and more frequent early falls (70% vs 36% and 50% in SN‐early and OPC‐early subtypes; p = 0.0039). Immunohistochemical validation confirmed extensive α‐synuclein pathology in both SN and OPC systems in the SN‐OPC‐synchronous subtype. Interpretation The SN‐OPC‐synchronous subtype provides insights into α‐synuclein propagation mechanisms, suggesting multiple initial seeding sites rather than unidirectional spread. This computational approach uncovered disease heterogeneity undetectable by conventional methods, potentially benefiting clinical trials through patient stratification. ANN NEUROL 2026;99:989–1000