Biomarkers.

Area of Science:

• Neuroimaging
• Neuropathology
• Geriatric Medicine

Background:

• Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC) is a common finding in aged brains, but definitive diagnosis requires autopsy.
• Current diagnostic limitations necessitate the development of in-vivo markers for LATE-NC.
• This study focused on identifying reliable MRI features for LATE-NC detection in community-based older adults.

Purpose of the Study:

• To develop and validate an automated, in-vivo magnetic resonance imaging (MRI) marker for Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC).
• To distinguish between early and advanced stages of LATE-NC using MRI-derived features.
• To package the developed marker into user-friendly software for potential clinical application.

Main Methods:

• Utilized ex-vivo MRI, in-vivo MRI, and neuropathology data from longitudinal aging studies.
• Developed a two-level stacking classifier using fractional anisotropy (FA), deformation-based morphometry (DBM), and lobar volumes to identify LATE-NC.
• Validated the classifier in-vivo and packaged the pipeline into an automated software container named MARBLE (MARker of Brain LatE).

Main Results:

• The ex-vivo classifier achieved an AUC of 0.85 (77% balanced accuracy) for distinguishing LATE-NC stages.
• In-vivo validation of the MARBLE marker demonstrated an AUC of 0.76 in the test group.
• Higher MARBLE scores correlated significantly with advanced LATE-NC stages (p < 0.001).

Conclusions:

• MARBLE is a novel, automated, in-vivo MRI marker for LATE-NC with promising diagnostic performance.
• The tool was trained on extensive ex-vivo data and validated in-vivo, showing potential for clinical utility.
• Further validation in independent cohorts is warranted, but MARBLE could aid in LATE-NC diagnosis, monitoring, and treatment development.