Diagnostic Feature Associations in Atypical Parkinsonian Disorders: A Survey of Movement Disorders Specialists

Objective: For the atypical parkinsonian disorders (APDs), including dementia with Lewy bodies (DLB), multiple system atrophy (MSA), progressive supranuclear palsy (PSP), and corticobasal degeneration (CBD), accurate antemortem diagnosis remains clinically challenging because of overlapping features, phenotypic heterogeneity, and limited definitive biomarkers. Consensus diagnostic criteria are complex and may not reflect how clinicians operationalize diagnostic reasoning in practice. Methods: We surveyed movement disorders specialists regarding diagnostic implications of 32 clinical features, comparing the results with a 2024 consensus algorithmic framework called the "General Neurologist's Practical Diagnostic Algorithm for APDs". We evaluated inter-rater agreement, diagnosis-level framework coverage, feature-level attribution patterns, and exact-set agreement. Results: Respondents demonstrated strong inter-specialist consensus in feature attribution (raw agreement 81.9%; Gwet's AC1 0.67, 95% CI 0.64-0.70). Specialist endorsement increased markedly with framework association strength, from 12.9% to 36.2% to 82.6% for unmapped, modest, and moderate associations respectively. Diagnosis-level coverage was higher for MSA, PSP, and CBD than for DLB. Exact-set agreement was substantially more variable and depended on whether modest associations were incorporated into the framework definition. Despite high self-reported familiarity with and use of published diagnostic criteria, respondents overwhelmingly expressed interest in a unified diagnostic framework for APDs with automated clinical decision support (CDS). Conclusions: Movement disorders specialists showed internally consistent feature-attribution patterns for archetypal features, while differing in the weighting and breadth assigned to weaker or overlapping associations. These findings are evidence for a gap between consensus schemas and specialist diagnostic reasoning and may inform future diagnostic frameworks and CDS approaches for the APDs.