Hippocampal Volumetry: A Promising and Accessible Tool for Early AD Diagnosis in Regions with Limited 18F‐FDG PET Imaging Access?

Background Hippocampal atrophy and hypometabolism are key neuroimaging biomarkers for early diagnosis of Alzheimer's Disease (AD). Artificial intelligence methods for quantitative automated volumetric analysis of brain magnetic resonance imaging (MRI) are enabling objective assessment of brain atrophy in a fast and accessible way. Thus, the objective of this work is to compare hippocampal volume derived from MRI with the hypometabolism observed in 18F‐FDG positron emission tomography (PET) in the context of early diagnosis of AD. Methods Twenty‐two patients with AD (11 women, [51–84yo], mean age±SD: 69.91±9.51) and twenty‐six normal individuals (16 women, age range [51–80yo], mean age±SD: 66.23±6.87) were enrolled in this study. Structural T1‐weighted and 18F‐FDG scans were acquired on a 3T MR‐PET system (Biograph mMR, Siemens Healthineers, Forchheim, Germany). A deep‐learning‐based research application (MorphoBox) was used to derive brain volume normalized by the total intracranial volume from MRI (Figure 1), and p ‐mod software was applied to calculate standardized uptake value ratio (SUVR) from PET imaging. Pearson correlations between normalized brain volumes and SUVR, as well as visual interpretations from diagnostic radiologists, were assessed. Receiver operating characteristics (ROC) analyses were conducted to investigate the discrimination ability of hippocampus relative volume on one hand and SUVR on the other hand. Results A significant difference in SUVR normalized by cerebellar SUV was observed between the AD and control groups in the following brain regions: hippocampus ( p