Predicting MCI outcome with clinically available MRI and CSF biomarkers

Abstract

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.

Figure 1. Survival curves according to risk category

Survival curves for the full mild cognitive impairment (MCI) cohort, and for negative and positive risk groups defined according to learning performance (Auditory Rey Verbal Learning Test [AVLT]), CSF T-tau, Aβ_1–42, and the tau/Aβ_1–42 ratio, as well as for medial temporal atrophy determined from the hippocampal occupancy score (HOC). Cox proportional hazard models controlled for age. The x-axis shows months to conversion to AD; the y-axis shows proportion of subjects who have not converted to Alzheimer disease. High risk is shown in red, low risk in blue.

Figure 2. Survival curves as a function of risk factor combinations

Survival curves are shown for patients with mild cognitive impairment stratified according to the combination of learning (Auditory Rey Verbal Learning Test [AVLT]) and atrophy (hippocampal occupancy score [HOC]) risk, learning and CSF risk, atrophy and CSF risk, and for individuals concordant on risk for all 3 measures. Cox proportional hazard model controlled for age. Green lines show those testing negative on all measures in the analysis, red lines show those testing positive on all measures. Blue and purple lines show survival for those with discordant risk factors.

Figure 3. Median survival times for those testing positive on each risk factor or combination of risk factors

Median survival time (in months) reflects the last time at which 50% of the subjects in the group retained the MCI diagnosis. AVLT = Auditory Rey Verbal Learning Test.