Biomarkers

Limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC) is a common pathologic finding in aged brain, however, definitive diagnosis of this disease is only possible at autopsy. This work aimed to develop a marker of LATE-NC based on in-vivo MRI features from a large group of community-based older adults.

This study included ex-vivo MRI, in-vivo MRI, and pathology data from four longitudinal clinicopathological cohort studies of aging conducted at the Rush Alzheimer's Disease Center (ROS, Rush MAP, MARS, LATC). LATE-NC was evaluated based on pathologic TDP-43 inclusions in 8 brain regions and categorized into 4 stages (Figure 1). MRI data was processed to obtain fractional anisotropy (FA), deformation-based morphometry (DBM) measurements, and lobar volumes. To develop a marker of LATE-NC, we first trained a classifier to distinguish between advanced (stages 2-3) and early LATE-NC stages (stages 0-1) based on ex-vivo MRI features (N = 863). Our classifier used a two-level stacking model for training and cross-validation, where level-1 estimators generated risk scores based on single-modality features, and level-2 estimator (logistic regression) provided the final LATE-NC prediction score based on the combined risk scores. We then translated the classifier to in-vivo and validated it on a separate set of participants (N = 60) with in-vivo MRI and pathology data. The entire pipeline was packaged into an automated software container named MARBLE (MARker of Brain LatE).

In the training group, the ex-vivo classifier demonstrated excellent performance, achieving an average AUC of 0.85±0.05 (sensitivity=78%, specificity=76%, balanced accuracy=77%) based on FA, DBM, and lobar volume features (Figure 2). In-vivo validation of MARBLE scores yielded an overall AUC of 0.76 in the test group. Additionally, ordinary least-squares regression revealed higher MARBLE scores with greater LATE-NC stages (p <0.001), controlling for antemortem interval (AMI) and scanners (Figure 3).

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 and showed decent performance in-vivo (AUC=0.76). 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.