AD-FTD FDG Difference Map

Description: In April 2018, we queried our database looking for patients who fulfilled the following criteria: i) Had available both a [18C] Pittsburgh Compound B (PiB)-PET scan for beta-amyloid and a [18F] Fluorodeoxyglucose (FDG)-PET scan of brain glucose metabolism acquired on the Biograph scanner; ii) Had either a clinical diagnosis of AD (McKhann et al. 2011) and a “positive” PIB-PET read, or a clinical diagnosis of FTD (either of a three prototypic phenotypes described in (Neary et al. 1998) and a “negative” PIB-PET read. This query resulted in 36 AD and 39 FTD patients. Five patients were later excluded because of incomplete FDG-PET SUVR (missing at least one of the 6 frames between 30 and 60 min post-injection), resulting in a final count of 35 AD and 35 FTD patients. Demographic information can be found in the paper. Note there is no overlap between this sample and the sample described in La Joie et al. 2015, Neuron. All patients were seen at the at University of California, San Francisco Memory & Aging Center and imaged at the Lawrence Berkeley National Labs. PET acquisition details can be found elsewhere (Ossenkoppele et al. 2016, Brain). FDG-PET images were processed using SPM12 using a previously described pipeline (Ossenkoppele et al. 2016, Brain). Briefly, six five-minute frames were realigned and averaged, and the average image was coregistered onto patient specific anatomical T1-MRI scans. Standard uptake value ratios (SUVR) were calculated using the pons (Freesurfer segmentation of the brainstem with manual cleaning) as a reference region, and SUVR images were warped to the MNI template using MRI-derived parameters. All 70 patients were entered into a voxelwise t-test controlling for age and disease severity (Clinical Dementia Rating Sum of Boxes score) using SPM12, highlighting differences in glucose hypometabolism (a proxy for neurodegeneration) between AD and FTD patients.