Predictive value of brain 18F-FDG PET neuroimaging in a mouse model of sudden death

Abstract

Sudden unexpected death (SUD) is a major cause of mortality in neurological disorders, with poorly understood mechanisms and no reliable biomarkers. Brain-heart axis dysfunction has been implicated in SUD pathogenesis. Using the FUSdelta14 mouse model, we tested whether cerebral glucometabolism, assessed by 18F-FDG PET, reflects autonomic dysregulation and predicts SUD vulnerability. 18F-FDG PET/CT and ECG were acquired at 10 weeks of age (FUSdelta14/delta14 n=12, FUS+/delta14 n=14, FUS+/+ n=12). Heart rate variability was analyzed in time and frequency domains, brain uptake via volume of interest and voxel-wise approaches, and correlations assessed by Pearson's analysis. A classification-tree model tested SUD predictability within the FUSdelta14/delta14 group. Female FUSdelta14/delta14 mice showed reduced 18-FFDG uptake in basal forebrain septum, left amygdala, and brainstem, with increased uptake in hypothalamus and right midbrain, alongside cardiac size, metabolic and heart rate variability alterations. 50% experienced SUD; in this subgroup, PET showed cortical hypometabolism (insular, entorhinal) and hypermetabolism in hypothalamus, amygdala, septum, and brainstem. Although brainstem uptake was able to singlehandedly predict SUD in this particular FUSdelta14/delta14 group (AUC=0.787), cross-validation also highlighted cortical metabolism as a relevant feature (UsageRate=0.38 and 0.44, respectively). 18F-FDG PET identifies region-specific metabolic signatures of brain-heart axis dysfunction linked to SUD, supporting its potential as a non-invasive biomarker for risk stratification.