Abstract Details

Title Predicting Clinical Course in Persons with Mild Cognitive Impairment using structural MRI.

Topic Aging, Dementia, and Behavioral Neurology

Presentation(s) P5 - Poster Session 5 (5:30 PM-6:30 PM)

Poster/Presentation
Number 1-003

Objective To test the hypothesis that machine learning (ML) models, applied to MRI data to distinguish normal controls (NC) from persons with AD dementia, can be applied to persons with MCI to identify prognostically distinct subgroups.

Background

Persons with MCI may progress to Alzheimer’s dementia (AD), remit to normal or remain MCI. Differentiating these subgroups is important for clinical trials design and in clinical practice.

Design/Methods

Of 1335 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) 436 were cognitively normal (CN), 665 had MCI, and 234 had AD. To distinguish between normal and AD groups we applied three classifiers (support vector machines (SVMs), linear discriminant analysis (LDA) and ensemble subspace discriminant (E-SD) algorithms on 44 automatically estimated brain region volumes. A 10-fold cross-validation procedure was applied to all models. We applied these models to prospectively predict the conversion from MCI to AD and the reversion from MCI to CN. Sensitivity, Specificity, positive predictive value (PPV) and negative predictive value (NPV) was calculated for each model.

Results Over an average follow up time of 3.0 years (SD = 2.2), 252 (37.9%) individuals progressed to AD and 37 (5.5%) persons with MCI reverted to CN. For the initial classification of CN vs AD, we obtained 92.2% accuracy (96% sensitivity and 83% specificity) for SVMs, 91.7% accuracy (95% sensitivity and 87% specificity) for both LDA and E-SD models. All three models, prospectively predicted future reversion from MCI to CN with 100% of reverters correctly classified. E-SD model had the highest predictive accuracy (64%) for future conversion from MCI to AD (sensitivity=64.3%, specificity=76.0%, PPV=62.1%, and NPV=77.7).

Conclusions

Classifiers trained to separate CN vs. AD based on volumetric MRI, can improve our prediction ability of MCI reversion to CN or MCI conversion to AD.

Authors/Disclosures
Ali Ezzati, MD (University of California, Irvine) PRESENTER	The institution of Dr. Ezzati has received research support from NIA. The institution of Dr. Ezzati has received research support from Alzheimer's Association. The institution of Dr. Ezzati has received research support from Cure Alzheimer's Fund.
Andrea R. Zammit, PhD (Albert Einstein College of Medicine)	Dr. Zammit has nothing to disclose.
Danielle Harvey, PA-C	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
Richard B. Lipton, MD, FAAN (Albert Einstein College of Medicine)	Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Allergan/Abbvie. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Amgen. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Biohaven. Dr. Lipton has received personal compensation in the range of $50,000-$99,999 for serving as a Consultant for Eli Lilly. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Lundbeck. Dr. Lipton has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for GlaxoSmithKline. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Teva. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Vedanta. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Merck. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Grifols. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Pfizer. Dr. Lipton has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Axon. Dr. Lipton has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Satsuma. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Genentech. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Cool Tech. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for BDSI. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Linpharma. Dr. Lipton has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Axsome. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Clexio. Dr. Lipton has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Shiratronics. Dr. Lipton has received personal compensation in the range of $5,000-$9,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Allergan/Abbvie. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Biohaven. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Eli Lilly. Dr. Lipton has received personal compensation in the range of $10,000-$49,999 for serving on a Scientific Advisory or Data Safety Monitoring board for Lundbeck. Dr. Lipton has or had stock in Biohaven.Dr. Lipton has or had stock in Manistee.Dr. Lipton has or had stock in Axon.Dr. Lipton has or had stock in CoolTech. The institution of Dr. Lipton has received research support from Teva. The institution of Dr. Lipton has received research support from Amgen. The institution of Dr. Lipton has received research support from Allergan/Abbvie. The institution of Dr. Lipton has received research support from Gammacore. The institution of Dr. Lipton has received research support from Axsome. The institution of Dr. Lipton has received research support from Charleston Labs. The institution of Dr. Lipton has received research support from Eli Lilly. The institution of Dr. Lipton has received research support from Satsuma. The institution of Dr. Lipton has received research support from NIH . The institution of Dr. Lipton has received research support from Veterans Administration. Dr. Lipton has received publishing royalties from a publication relating to health care.

��ɫ��

��ɫ��