Non-invasive classification of limbic age-related TDP-43 encephalopathy (LATE) using MRI features
Poster No:
877
Submission Type:
Abstract Submission
Authors:
Mahir Tazwar1, Arnold Evia2, Abdur Raquib Ridwan2, David Bennett2, Julie Schneider2, Konstantinos Arfanakis2
Institutions:
1Illinois Institute of Technology, Chicago, IL, 2Rush University Medical Center, Chicago, IL
First Author:
Co-Author(s):
Introduction:
Limbic-predominant age-related TDP-43 encephalopathy (LATE), a common neurodegenerative disorder in older adults, is defined by neuropathological changes (LATE-NC) due to phosphorylated TDP-43 protein accumulation [6,7]. Clinically, it exhibits amnestic dementia syndrome similar to Alzheimer's disease (AD), and often co-occurs with AD [5,6]. However, definitive diagnosis of this disease is only possible at autopsy. Recent MRI studies have reported changes in morphometric and diffusion properties with LATE-NC that could serve as indirect markers to predict this disease in-vivo [2,9,10]. Therefore, this work aimed to develop a marker of LATE-NC based on in-vivo MRI features.
Methods:
Participants, MRI, neuropathology
This study utilized data from four longitudinal cohort studies of aging conducted at the Rush Alzheimer's Disease Center [1,3]. Participants with in-vivo MRI, ex-vivo MRI, and pathology data were included in this study (Fig.1A). LATE-NC was evaluated based on pathologic TDP-43 inclusions in 8 brain regions and categorized into 4 stages (Fig.1A). Dataset consisted of 863 participants for ex-vivo model training, 60 participants for in-vivo validation (test set), and a subset of training samples having both in-vivo and ex-vivo MRI were used for translation (Fig.1A,B). MRI data was processed to obtain fractional anisotropy (FA) values [8], logarithmic Jacobian determinant values of the deformation fields using deformation-based morphometry (DBM) [9], and lobar volumes (temporal, frontal, parietal, subcortical) using multi-atlas segmentation [4].
Ex-vivo classifier
MRI features included FA, DBM, and lobar volume measurements. As a preprocessing step, the lobar volumes were normalized by total gray matter and total hemisphere volume. A two-level stacking classifier was trained to distinguish individuals with advanced LATE stages (stages 2-3) from those with earlier stages (stages 0-1) based on ex-vivo MRI features (Fig.1C). Model performance was evaluated through 100-fold repeated cross-validation, employing 80/20% split for training/test data within each fold, and calculating the mean area under the curve (AUC).
In-vivo validation
Linear mixed-effects models were used to translate ex-vivo MRI features to in-vivo and then to translate the ex-vivo classifier to in-vivo. The resulting in-vivo marker was packaged into an automated software container named MARBLE (MARker of Brain LatE), which takes raw in-vivo MRI data as input and provides a score that is related to the likelihood a person suffers from LATE-NC (Fig.2B). The performance of MARBLE in predicting LATE-NC in-vivo was assessed on 60 participants with in-vivo MRI and pathology data.
This study utilized data from four longitudinal cohort studies of aging conducted at the Rush Alzheimer's Disease Center [1,3]. Participants with in-vivo MRI, ex-vivo MRI, and pathology data were included in this study (Fig.1A). LATE-NC was evaluated based on pathologic TDP-43 inclusions in 8 brain regions and categorized into 4 stages (Fig.1A). Dataset consisted of 863 participants for ex-vivo model training, 60 participants for in-vivo validation (test set), and a subset of training samples having both in-vivo and ex-vivo MRI were used for translation (Fig.1A,B). MRI data was processed to obtain fractional anisotropy (FA) values [8], logarithmic Jacobian determinant values of the deformation fields using deformation-based morphometry (DBM) [9], and lobar volumes (temporal, frontal, parietal, subcortical) using multi-atlas segmentation [4].
Ex-vivo classifier
MRI features included FA, DBM, and lobar volume measurements. As a preprocessing step, the lobar volumes were normalized by total gray matter and total hemisphere volume. A two-level stacking classifier was trained to distinguish individuals with advanced LATE stages (stages 2-3) from those with earlier stages (stages 0-1) based on ex-vivo MRI features (Fig.1C). Model performance was evaluated through 100-fold repeated cross-validation, employing 80/20% split for training/test data within each fold, and calculating the mean area under the curve (AUC).
In-vivo validation
Linear mixed-effects models were used to translate ex-vivo MRI features to in-vivo and then to translate the ex-vivo classifier to in-vivo. The resulting in-vivo marker was packaged into an automated software container named MARBLE (MARker of Brain LatE), which takes raw in-vivo MRI data as input and provides a score that is related to the likelihood a person suffers from LATE-NC (Fig.2B). The performance of MARBLE in predicting LATE-NC in-vivo was assessed on 60 participants with in-vivo MRI and pathology data.
·Figure 1. (A) Participant characteristics. (B) Venn-diagram. (C) Schematic of the ex-vivo classifier model.
Results:
In the training group, the ex-vivo classifier demonstrated excellent performance, achieving an average AUC of 0.85±0.05 (sensitivity=78%, specificity=76%) based on FA, DBM, and normalized lobar volume features (Fig.2A). The two-level classifier, incorporating all MRI features (FA, DBM, and lobar volume), demonstrated significantly higher AUCs than other classifier models trained on individual modality features or demographic data (Bonferroni-corrected p<0.0001, two-sided Wilcoxon signed rank test) (Fig.2A). In-vivo validation of MARBLE scores yielded an overall AUC of 0.81 in the test group. Additionally, linear regression revealed higher MARBLE scores with greater LATE-NC stages (p<0.001), controlling for antemortem interval (AMI) and scanners (Fig.2B).
·Figure 2. (A) Ex-vivo classifier performance. (B) In-vivo validation performance.
Conclusions:
This study developed MARBLE, a novel, automated, in-vivo marker of LATE-NC based on MRI features. MARBLE was trained on ex-vivo MRI and pathology data from a large number of community-based older adults, translated to in-vivo, automated, and validated in-vivo. MARBLE performance in-vivo was excellent (AUC=0.81). While further validation is needed in independent cohorts, MARBLE has the potential to significantly contribute towards in-vivo diagnosis, monitoring, prevention, and treatment of LATE-NC.
Disorders of the Nervous System:
Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)
Lifespan Development:
Aging 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
Keywords:
ADULTS
Aging
Degenerative Disease
Machine Learning
MRI
STRUCTURAL MRI
White Matter
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
Other - TDP-43, Limbic-predominant age-related TDP-43 encephalopathy (LATE), LATE-NC, Deformation-based morphometry (DBM), Fractional anisotropy (FA), volumetric segmentation, classifier
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
2. Bejanin, A. (2019). Antemortem volume loss mirrors TDP-43 staging in older adults with non-frontotemporal lobar degeneration. Brain, 142(11), 3621–3635.
3. Bennett, D. A. (2018). Religious Orders Study and Rush Memory and Aging Project. Journal of Alzheimer’s Disease, 64(s1), S161–S189.
4. Iglesias, J. E. (2015). Multi-atlas segmentation of biomedical images: A survey. Medical Image Analysis, 24(1), 205–219.
5. Kapasi, A. (2020). Limbic-predominant age-related TDP-43 encephalopathy, ADNC pathology, and cognitive decline in aging. Neurology, 95(14).
6. Nelson, P. T. (2019). Limbic-predominant age-related TDP-43 encephalopathy (LATE): Consensus working group report. Brain, 142(6), 1503–1527.
7. Nelson, P. T. (2023). LATE-NC staging in routine neuropathologic diagnosis: An update. Acta Neuropathologica, 145(2), 159–173.
8. Pierpaoli, C. (2010). TORTOISE: an integrated software package for processing of diffusion MRI data. ISMRM 18th Annual Meeting. International Society for Magnetic Resonance in Medicine.
9. Tazwar, M. (2024). Deformation-based morphometry reveals lower brain tissue volume in autopsy confirmed limbic age-related TDP-43 encephalopathy (LATE). 2024 ISMRM & ISMRT Annual Meeting & Exhibition. International Society for Magnetic Resonance in Medicine, Singapore, SG.
10. Tazwar, M. (2024). Limbic-predominant age-related TDP-43 encephalopathy neuropathological change (LATE-NC) is associated with abnormalities in white matter structural integrity and connectivity: An ex-vivo diffusion MRI and pathology investigation. Neurobiology of Aging, 140, 81–92.