Copathologies of Limbic-Predominant Age-Related TDP-43 Encephalopathy and Alzheimer Disease: [18F]FDG PET Statistical Mapping and Quantitative MRI Volumetry

Abstract

Limbic-predominant age-related transactive response DNA-binding protein 43 kDa encephalopathy (LATE) is emerging as a prevalent neurodegenerative disorder in aging populations, mimicking the clinical presentation of Alzheimer disease (AD). This study investigates in vivo [18F]FDG PET and MRI biomarkers in detecting probable LATE neuropathologic change. Methods: We retrospectively analyzed 944 [18F]FDG PET cases referred from cognitive disorder clinics in a tertiary care center. To characterize the LATE and AD findings objectively and quantitatively, we created 3-dimensional stereotactic surface projection PET templates for LATE neuropathologic change (n = 6) and AD neuropathologic change (n = 32) from autopsy-confirmed Alzheimer's Disease Neuroimaging Initiative and University of Utah datasets, respectively. All 3-dimensional stereotactic surface projection z score [18F]FDG maps were created in comparison to normal PET scans from 20 control cases whose amyloid PET scans were negative. Using the autopsy-derived z score maps, z score product indices (the individual z score map multiplied by the z scores of autopsy-confirmed cohorts) were generated for each subject, stratifying participants into probable LATE, probable LATE and AD (LATE+AD), and probable AD. Clinical and quantitative MRI volumetry data were compared across the groups using 1-way or Welch ANOVA and Fisher exact tests, depending on the assessed variables. Results: Of the 944 clinical cases, 13.0% were characterized as probable LATE (2.4% pure LATE and 10.6% LATE+AD) and 23.7% were characterized as probable AD without LATE. MRI volumetry revealed that the medial temporal lobe was most affected in pure LATE cases (P < 0.001), whereas the orbitofrontal gyrus and lateral temporal lobe were most vulnerable in mixed LATE+AD cases (P = 0.001; P < 0.001). Post hoc analysis identified the entorhinal cortex and amygdala as key regions for distinguishing mixed LATE+AD cases from pure LATE and pure AD cases, respectively (P = 0.05; P < 0.001). Subgroup analysis of the probable LATE+AD group demonstrated additive or synergistic effects of both pathologies, with three quarters of cases exhibiting concordant lateralized metabolic brain changes, predominantly left-sided, based on LATE and AD z score products (P < 0.001). A similar pattern of left-dominant brain atrophy was observed in MRI volumetry. Conclusion: Substantial numbers of our patients exhibited LATE features that were characterized objectively using scans from autopsy-proven cases. These LATE cases were associated with specific regional atrophy measured by quantitative MRI. Cases with LATE+AD copathologies demonstrated synergistic hemispheric involvement. Further investigations of such synergistic changes between LATE and AD are warranted.