Evaluation of Brain Metabolic Change by Application of Network-based Analysis with Persistent Homology: [18F]FDG PET Study using a Chronic Cerebral Hypoperfusion Rat Model

Abstract

Purpose: Bilateral common carotid artery occlusion (BCCAO) in rats is an established animal model for vascular dementia. The conventional analytic method based on regional glucose metabolism and a novel network-based analysis with persistent homology were applied to evaluate brain metabolic change by using small animal PET with 2-deoxy-2-[18F]fluoro-D-glucose (FDG-PET) in BCCAO rat model.<br/> <br/> Methods: Wistar rats were divided into three groups: sham, BCCAO, and treated groups. Serial small animal brain FDG-PET scans were obtained at preoperative baseline and postoperative observation period up to 6 months. Two different FDG-PET images obtained under different anesthetic protocols were compared to<br/> investigate effect of prolonged anesthesia on brain metabolism. Visual assessment, conventional quantitative evaluation, and network-based analysis, volume entropy, and permutation test were introduced to evaluate significant change of PET parameters in chronic hypoperfused rat brain and this analytic approach was also<br/> applied to evaluate effect of a therapeutic agent.<br/> <br/> Results: Static PET acquisition after awakening showed less decreased regional glucose metabolism than dynamic PET acquisition throughout the whole brain except for cerebellum and brainstem and therefore it was selected as a default PET imaging protocol. All three groups did not show significant difference in inter-regional or temporal [18F]FDG uptake by visual assessment and conventional quantitative analysis. Network-based analysis with persistent homology including correlation matrices, single-linkage matrices, β0 curves, and volume entropy indicated more impaired metabolic connections throughout the whole brain regions in BCCAO group than in baseline and sham groups. Especially, single-linkage distances showed that BCCAO group had significant deterioration of inter-regional brain metabolic connectivity between the right entorhinal cortex and other brain regions compared with sham group. This approach also showed that treated group<br/> had improved metabolic connectivity than BCCAO group and there was significant difference in metabolic connectivity between the left retrosplenial cortex and other brain regions.<br/> <br/> Conclusion: Network-based analysis with persistent homology showed higher sensitivity to brain metabolic change than conventional regional glucose metabolism based analysis. This network-based analysis will be a promising approach to evaluate impairment of brain inter-regional connectivity related with a variety of<br/> neurodegenerative diseases and psychiatric disorders.<br/> <br/> Keywords: vascular dementia, Bilateral Common Carotid Artery Occlusion (BCCAO), [18F]FDG PET, brain connectivity <br/>