Deep Learning for Ultra High Resolution T2‐weighted 7 Tesla Ex vivo Magnetic Resonance Imaging Reveals Differential Subcortical Atrophy across Neurodegenerative Diseases. (20th December 2022)

• Record Type: Journal Article
• Authors: Khandelwal, Pulkit; Duong, Michael Tran; Chung, Eunice; Sadaghiani, Shokufeh; Lim, Sydney A; Ravikumar, Sadhana; Arezoumandan, Sanaz; Peterson, Claire; Bedard, Madigan L; Capp, Noah; Ittyerah, Ranjit; Migdal, Elyse; Choi, Grace; Kopp, Emily; Patino, Bridget Loja; Hasan, Eusha; Li, Jiacheng; Prabhakaran, Karthik; Mizsei, Gabor; Gabrielyan, Marianna; Schuck, Theresa; Robinson, John L.; Ohm, Daniel T; Nasrallah, Ilya M.; Lee, Eddie B; Trojanowski, John Q; McMillan, Corey T; Grossman, Murray; Irwin, David J.; Tisdall, Dylan M; Das, Sandhitsu R.; Wisse, Laura EM; Wolk, David A.; Yushkevich, Paul A.;
• Is Part Of: Alzheimer's & dementia. Volume 18(2022)Supplement 5
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-12-20
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.062628
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Ex vivo magnetic resonance imaging (MRI) of the brain provides remarkable advantages over in vivo MRI for linking neuroanatomy and morphometry to underlying pathology (Yushkevich et al. 2021, Ravikumar et al. 2021). Subcortical structures show atrophy in certain neurodegenerative diseases, especially Frontotemporal Lobar Degeneration with TDP‐43 (FTLD‐TDP) and four‐repeat (4R) tauopathies (i.e., Corticobasal Degeneration, Progressive Supranuclear Palsy) (Miletić et al. 2022), yet few methods exist to measure subcortical atrophy in ex vivo MRI. We present a framework to quantify subcortical morphometry using 7 Tesla ex vivo MRI and distinguish atrophy patterns across neurodegenerative spectrums. Method: A deep learning method, nnU‐Net (Isensee et al. 2021), was trained on manual segmentations from only 3 brain hemispheres to obtain automated segmentations of 4 subcortical structures (caudate, putamen, globus pallidus, thalamus) across 38 subjects spanning Alzheimer's Disease (AD), Lewy Body Disease (LBD), FTLD‐TDP, 4R tauopathies and miscellaneous tauopathies (Figure 1, Table 1). Subcortical volumes were extracted from automated segmentations. Cerebral cortical volume was computed via cortical segmentation method in Khandelwal et al. 2021. Regional volumes were evaluated via likelihood ratio tests (Figure 2), adjusted for covariates (age, sex and intracranial volume from in vivo MRI) and multiple tests. Separately, correlations were computed between subcortical volumes, cortical thicknesses at 18 landmark locations and neuropathological ratings (Khandelwal et al. 2021, Wisse et al. 2021, Figure 3). Result: The pipeline validated regional volumetric relationships in neurodegeneration. Global cortex volume did not significantly differ among disease groups (Figure 2). Compared to AD, FTLD‐TDP had significantly lower putamen and thalamus volumes while 4R tauopathies had reduced putamen and caudate volumes ( P 's<0.05, adjusted for covariates/multiple comparisons). Before multiple tests correction, there were decreased covariate‐adjusted volumes in globus pallidus and caudate in FTLD‐TDP and thalamus in 4R tauopathy relative to AD. Subcortical volumes correlated with each other ( P 's<0.05) but not with cortical thickness, with trends in motor cortex (Figure 3). Subcortical volumes also trended with local tau pathology (Figure 4). Conclusion: Our ex vivo neuroimaging framework differentiates subcortical atrophy patterns in FTLD‐TDP and 4R tauopathies compared to AD, highlighting utility in ex vivo imaging for diagnosing and investigating neurodegeneration.