Toward higher‐sensitivity, shorter‐interval MRI measures of atrophy in neurodegenerative dementias

Abstract

AbstractBackgroundConventional single acquisition T1w structural MRI can detect significant change in brain structure given ample latency between scans; most contemporary studies of neurodegenerative dementias collect MRI scans at one‐year intervals and calculate annualized rates of atrophy. MRI methods to detect atrophy over shorter (3‐ or 6‐month) intervals would be extremely valuable for basic human neuroscience research about neurodegenerative disease mechanisms as well as for translation into clinical trials. Measurement of change over short intervals would allow neurodegeneration to be monitored in smaller, early‐phase therapeutic trials—potentially even as a marker of target engagement in trials of disease‐modifying interventions.MethodOur pilot cohort consisted of 5 patients: 4 with primary progressive aphasia (2 semantic variant likely due to FTLD, 2 logopenic variant likely due to AD), 1 with posterior cortical atrophy likely due to AD. We collected 32 accelerated 1mm T1w scans in four 30‐minute blocks across two days for each of 3 timepoints (baseline, 3 months, 6 months). We used the FreeSurfer longitudinal processing stream to make robust estimates of the cortex and subcortical structures for each T1w volume. For each timepoint, cortical thickness and volume and subcortical volume estimates were derived from all 32 reconstructed scans. Rates of change were calculated from ROI measures. In addition, a cortical atrophy map was calculated for each subject at each timepoint in comparison with a group of age‐matched controls (N=32).ResultEach of the patients exhibits a typical pattern of atrophy for their clinical syndrome. Cortical atrophy maps are highly reliable between timepoints, and for each subject show visibly apparent but small changes in the magnitude and spread of atrophy. ROI measures demonstrate a slight but observable decrease in the size of many brain structures, even across a 3‐month interval. Cortical thickness changes in regions implicated in the patient’s individual clinical syndrome show the greatest magnitude of change over time—detectable in each individual at 3‐month intervals.ConclusionThis novel approach to MRI acquisition and analysis can sensitively measure neurodegeneration in individuals over shorter intervals of time, which could lead to substantial changes in methods for measuring disease mechanisms and in clinical trial design.