Progression of metabolic brain patterns in common dementias

Background Neurodegenerative dementias are chronic progressive syndromes. The neurodegenerative process can be well observed with the metabolic brain imaging, which reveals typical patterns of regional hypometabolism. The specific metabolic brain patterns have been previously identified for the common dementia syndromes using 18 F‐FDG PET scans and SSM/PCA network analysis and they represent a metabolic biomarker of neurodegenerative process. Aims: The aim of this retrospective study was to analyse changes in the expression of the three characteristic metabolic patterns: Alzheimer’s disease related pattern (ADRP), frontotemporal dementia related pattern (FTDRP) and dementia with Lewy bodies related pattern (DLBRP), from the 18 F‐FDG PET scans in a cohort of patients with cognitive decline, who were scanned with 18 F‐FDG PET twice. Method We analysed clinical data and 18 F‐FDG PET brain scans from 25 patients with cognitive decline due to AD ( n =5), FTD ( n =6), DLB ( n =7) and non‐AD cognitively stable mild cognitive impairment (MCI, n =7) who underwent two 18 F‐FDG PET imaging, within 2.9 years. We performed TPR network analysis to measure the expression of ADRP, FTDRP and DLBRP in all 18 F‐FDG PET scans. Then we calculated absolute difference in pattern expression between the two scans and fitted a series of linear models to examine slopes of progression. Results In AD group we have observed a larger increase of ADRP‐Z scores ( β =1.21, p <0.001) in comparison to FTDRP‐z ( β =0.98) or DLBRP‐Z ( β =0.07). In FTD group we have observed a larger increase of FTDRP‐Z scores ( β =0.85, p <0.001) in comparison to ADRP‐Z ( β =0.43) or DLBRP‐Z ( β =‐0.36). In DLB group we have observed the highest coefficient for ADRP‐Z ( β= 0.84), followed by DLBRP‐Z ( β= 0.45) and FTDRP‐Z ( β= 0.14). On the other hand, DLBRP expression progressed in time in DLB group, but not in AD or FTD group. In MCI group FTDRP‐Z expression scores remained stable, while ADRP‐Z and DLBRP‐Z scores decreased in time. Conclusions Characteristic and dementia specific metabolic brain patterns have proved to be a reliable imaging biomarker of the disease progression in common dementia syndromes. As the diseases progresses, the expression of corresponding patterns increases.