Effects of MR based attenuation correction on clinical interpretation and quantification of [18F]flutemetamol scans: A comparison between PET/CT and PET/MR

Purpose [ 18 F]Flutemetamol (FMM) binds to amyloid-β, which is enhanced in Alzheimer's disease (AD). Recently, PET/MR systems have become available enabling acquisition of MR and human amyloid images in a single scanning session. PET/MR systems use MR based attenuation correction (MR-AC), which does not include actual attenuation coefficients, but rather uses tissue segmentation with allocation of standard attenuation coefficients to different tissue classes. This might result in bias, because bone cannot be segmented using a standard MR image. The aim of this pilot study was to assess the effects of MR-AC on clinical interpretation and regional distribution of FMM uptake. Methods Ten memory-clinic patients underwent a PET/CT FMM scan, followed by a PET/MR scan. Regional standardized uptake value (SUV) and SUV ratios (SUVr), with cerebellar grey matter as reference, were calculated for all FMM scans. FMM SUVr scans were rated visually (amyloid-positive/amyloid-negative). PET/CT and PET/MR derived regional FMM SUVr were compared with each other. Results Scans of 2 subjects were excluded due to movement and poor quality. Visual rating showed good agreement (7/8). Regional SUVr showed good correlation between PET/CT and PET/MR (r=0.93), although PET/MR values were higher than corresponding PET/CT values (mean 6.1%). This could be due to differences in AC or differences in scan time. Conclusions PET/MR seems promising for routine clinical amyloid imaging, as it is patient friendly. Quantification of FMM SUVr on PET/MR should be explored further.