Objective: To develop models that determine likeliness of positive detection of pathologic alpha-synuclein (asyn) in cerebrospinal fluid (CSF) as measured by seed amplification assay (SAA) using only baseline demographic and clinical features.

Background: The development of the SAA to detect asyn in CSF can reliably identify most individuals with Parkinson’s disease (PD) [1]. However, the proportion of people living with PD who are SAA-positive (SAA+) varies by demographic and genetic features (e.g., LRRK2 mutation status), suggesting that SAA status may be predictable based on these characteristics.

Method: Data used in this analysis come from the Parkinson’s Progression Marker Initiative (PPMI) observational study. The sample includes people with sporadic PD, genetic-associated PD, and people without PD or other neurologic disease (“healthy controls”). Logistic regression models were developed to estimate the likeliness of asyn SAA status (SAA+ vs. SAA-). Predictors assessed in models include age, sex, presence and type of LRRK2 genetic mutation regardless of manifest or non-manifest disease, APOE E4 carrier status, University of Pennsylvania Smell Identification Test (UPSIT) scores or percentiles, REM Sleep Behavior Disorder Questionnaire (RBDSQ) score, Montreal Cognitive Assessment score, and history of constipation problems. The process of model training and testing was conducted employing five-fold cross-validation. The performance of models were assessed by comparing model-derived SAA status likeliness versus the actual occurrence of CSF asyn, as determined by the asyn SAA.

Results: There were 1184 PPMI participants in the model training/testing dataset, consisting of 60% with sporadic PD, 20% with genetic PD, 19% healthy controls, and 1% deemed non-healthy controls. Using only raw UPSIT score as a predictor, upon testing, the model achieved 0.911 AUROC, 0.856 accuracy, and 0.904/0.737 sensitivity/specificity in determining SAA status.  A model incorporating age, sex, LRRK2 mutation, UPSIT score, RBDSQ score, and history of constipation problems achieved 0.923 AUROC, 0.876 accuracy, and 0.918/0.770 sensitivity/specificity.

Conclusion: Model-derived likeliness of asyn SAA status can be determined using UPSIT data with reasonably good reliability. Incorporating additional genetic and clinical data helps improve model performance.

References: 1. Siderowf A, Concha-Marambio L, Lafontant DE, Farris CM, Ma Y, Urenia PA, Nguyen H, Alcalay RN, Chahine LM, Foroud T, Galasko D, Kieburtz K, Merchant K, Mollenhauer B, Poston KL, Seibyl J, Simuni T, Tanner CM, Weintraub D, Videnovic A, Choi SH, Kurth R, Caspell-Garcia C, Coffey CS, Frasier M, Oliveira LMA, Hutten SJ, Sherer T, Marek K, Soto C; Parkinson’s Progression Markers Initiative. Assessment of heterogeneity among participants in the Parkinson’s Progression Markers Initiative cohort using α-synuclein seed amplification: a cross-sectional study. Lancet Neurol. 2023 May;22(5):407-417. doi: 10.1016/S1474-4422(23)00109-6. PMID: 37059509; PMCID: PMC10627170.

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/determining-cerebrospinal-fluid-alpha-synuclein-seed-amplification-assay-status-from-demographics-and-clinical-data/