Ten‐minute cognitive screening tool for mild cognitive impairment and prediction of pathological ß‐amyloid. (31st December 2021)

• Record Type: Journal Article
• Authors: Butler, P. Monroe; Osik, Jason; Schwartz, Joel; Irzhevsky, Victoria; Hansson, Oskar; Brown, Roland; Gao, Feng; Moss, Mark
• Is Part Of: Alzheimer's & dementia. Volume 17(2021)Supplement 6
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-12-31
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.054618
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Early detection of cognitive decline is critical to improve outcomes for Alzheimer's disease (AD) and other dementias. Screening for dementia risk during prodromal stages such as mild cognitive impairment (MCI) typically occurs through brief, bedside cognitive measures. While these tests readily detect dementia, the relative sensitivity/specificity for identifying MCI is variable and the ability to predict underlying AD is poor. Herein, we compare traditional cognitive screening tools with a novel ten‐minute battery to test predictive performance to discriminate MCI from normal cognition, and MCI with vs. without ß‐amyloid burden. The ten‐minute battery consisted of the Digit‐Symbol Substitution Test, Trail Making Tests A and B, and a delayed recall task. Method: Clinical data were assessed from healthy aging and MCI patients from: (i) the AD Neuroimaging Initiative Study (ADNI), (ii) the Swedish Biomarkers For Identifying Neurodegenerative Disorders Early and Reliably Study (BioFINDER), and (iii) the screening phase of aducanumab's Phase 2 EVOLVE study (NCT03639987) before the trial was prematurely terminated. Datasets were used to build and validate the performance of a statistical model derived from outcome measurements extracted from the ten‐minute battery to predict imaging or fluid biomarker‐based readouts of pathological ß‐amyloid load. Feature selection and model performance evaluations were achieved in two independent ways: stepwise model building and coefficient penalization. Neuropsychological variables were z‐scored against normative data. Additional predictors included age and APOE genotype. Model performance were compared with a standardized screening tool (MMSE) to assess relative ability to differentiate MCI from normal cognition, and MCI with vs. without ß‐amyloid burden based on imaging/fluid biomarkers. Result: The findings showed greater sensitivity and specificity achieved with the ten‐minute battery relative to the standard cognitive screening tools for MCI detection ( AUC‐ROC 0.98 vs. 0.78) and Aß prediction ( AUC‐ROC 0.81 vs. 0.58), respectively. Of striking interest, removing APOE as a predictor markedly reduced accuracy in the traditional, but not in the novel ten‐minute battery. Conclusion: These preliminary data suggest that the novel combination of neuropsychological tests performed in ten‐minutes may serve as a compact, more sensitive and specific screening tool to identify patients with MCI with Aß burden compared to traditional screening batteries.