T2 Relaxometry and Diffusion Tensor Indices of the Hippocampus and Entorhinal Cortex Improve Sensitivity and Specificity of MRI to Detect Amnestic Mild Cognitive Impairment and Alzheimer's Disease Dementia

BACKGROUND: Quantitative T2 and diffusion MRI indices inform about tissue state and microstructure, both of which may be affected by pathology before tissue atrophy.

PURPOSE: To evaluate the capability of both volumetric and quantitative MRI (qMRI) of the hippocampus and entorhinal cortex (EC) for classification of amnestic mild cognitive impairment (aMCI) and Alzheimer's disease dementia (ADD).

STUDY TYPE: Retrospective cross-sectional study.

POPULATION: Consecutive cohorts of healthy age-matched controls (n = 62), aMCI patients (n = 25), and ADD patients (n = 14).

FIELD STRENGTH/SEQUENCE: 3T using T1-weighted imaging, T2-weighted imaging, T2 relaxometry and diffusion tensor imaging (DTI).

ASSESSMENT: Montreal Cognitive Assessment and paired associate learning tests for cognitive state. Hippocampal subfield volumes by the automated segmentation of hippocampal subfields system from structural brain images. T2 relaxation time and DTI indices quantified for hippocampal subfields. The fraction of voxels with high T2 values (>20 ms above subfield median) was calculated and regionalized for hippocampus and EC.

STATISTICAL TESTS: Support vector machine and receiver operating characteristic analyses from cognitive and MRI data.

RESULTS: qMRI classified aMCI and ADD with excellent sensitivity (79.0% and 94.5%, respectively) and specificity (85.6% and 86.1%, respectively), superior to volumes alone (70.0% and 84.5% for respective sensitivities; 82.2 and 91.1 for respective specificities) and similar to cognitive tests (61.7% and 87.5% for respective sensitivities; 88.2% and 90.7% for respective specificities). Regions of high T2 are dispersed throughout each hippocampal subfield in aMCI and ADD with higher concentration than controls, and was most pronounced in the EC. No other individual qMRI marker than EC volume can separate aMCI from ADD, however. DATA CONCLUSION: qMRI markers of hippocampal and entorhinal tissue states are sensitive and specific classifiers of aMCI and ADD. They may serve as markers of a neurodegenerative state preceding volume loss.

LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2019;49:445-455.