B – 156 Predicting Alzheimer’s Neuropathology in MCI and Mild Dementia: A Comparison of Three Memory Measures

Abstract Objective Early identification of symptoms related to Alzheimer's disease (AD), such as episodic memory deficits, enables timely intervention and monitoring. We compared the ability of three memory measures administered during early neurocognitive impairment in predicting AD at autopsy. Method Data from 831 participants in the National Alzheimer’s Coordinating Center database with a diagnosis of mild cognitive impairment (MCI; n=220) or mild dementia (Clinical Dementia Rating score of 0.5 or 1; n=611) who underwent evaluation for AD neuropathological change (ADNC) were included. The Montreal Cognitive Assessment Memory Index Score (MoCA-MIS), Benson Figure Delayed Recall (BFDR), and Craft Story Delayed Recall (CSDR) were administered at baseline. ADNC was dichotomized (i.e., none/low or intermediate/high). Paired-sample receiver operating characteristic (ROC) analyses compared the discriminant ability of the memory measures. Results The average age at baseline was 75.1 (SD=10.4), with an average of 4.1 years between assessment and death. Most participants (74%) displayed intermediate/high ADNC. All three baseline memory scores significantly differentiated the groups (AUC ps.001), but AUC values were in the low range. The MoCA-MIS (.64) and BFDR (.68) displayed similar discriminative ability (p=.10), and both were superior to the CSDR (.58, ps.01). Conclusion In those with MCI/mild dementia, all memory measures demonstrated low identification accuracy for ADNC. However, the MoCA-MIS and BFDR outperformed the CSDR in predicting intermediate/high ADNC at autopsy. The MoCA-MIS shows promise as a brief alternative for memory assessment, potentially supporting broader clinical applicability. Further investigation is warranted to identify the most effective brief approaches to predicting AD pathology.