Related Experiment Video
Updated: May 28, 2026
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
Predicting MCI outcome with clinically available MRI and CSF biomarkers
D Heister1, J B Brewer, S Magda
1Department of Radiology, University of California, San Diego, La Jolla, CA 92093-0841, USA.
Objective:
To determine the ability of clinically available volumetric MRI (vMRI) and CSF biomarkers, alone or in combination with a quantitative learning measure, to predict conversion to Alzheimer disease (AD) in patients with mild cognitive impairment (MCI).
Methods:
We stratified 192 MCI participants into positive and negative risk groups on the basis of 1) degree of learning impairment on the Rey Auditory Verbal Learning Test; 2) medial temporal atrophy, quantified from Food and Drug Administration-approved software for automated vMRI analysis; and 3) CSF biomarker levels(.) We also stratified participants based on combinations of risk factors. We computed Cox proportional hazards models, controlling for age, to assess 3-year risk of converting to AD as a function of risk group and used Kaplan-Meier analyses to determine median survival times.
Results:
When risk factors were examined separately, individuals testing positive showed significantly higher risk of converting to AD than individuals testing negative (hazard ratios [HR] 1.8-4.1). The joint presence of any 2 risk factors substantially increased risk, with the combination of greater learning impairment and increased atrophy associated with highest risk (HR 29.0): 85% of patients with both risk factors converted to AD within 3 years, vs 5% of those with neither. The presence of medial temporal atrophy was associated with shortest median dementia-free survival (15 months).
Conclusions:
Incorporating quantitative assessment of learning ability along with vMRI or CSF biomarkers in the clinical workup of MCI can provide critical information on risk of imminent conversion to AD.
