[18F]FDG brain PET-MR versus PET/CT in a clinical cohort: intra-subject quantitative comparisons

Abstract There is limited evidence on the advantages of brain 2-deoxy-2-[ 18 F]fluoroglucose (FDG) positron emission tomography (PET) with simultaneous magnetic resonance imaging (PET-MR) over PET with computed tomography (PET/CT). We quantitatively compared brain FDG PET/CT and PET-MR images from a large clinical cohort. Participants with refractory epilepsy or cognitive problems underwent PET/CT at approximately 30 min post-injection, followed by PET-MR at approximately 100 min. Standardized uptake values (SUVs) were extracted in anatomical regions using a multi-atlas method, and voxel-wise comparisons were performed with statistical parametric mapping. Fifty-five image pairs passed quality control. Within-subject correlations between modalities were very high (0.96, range 0.84–0.99); within-region, across-participant correlations were 0.82 (0.72–0.90). PET-MR SUVs were on average 20% higher than PET/CT, more so for higher-uptake regions. Voxel-wise, PET-MR showed relatively higher uptake in visual and auditory cortices, lower uptake in medial frontal and anteromesial temporal cortices, and better grey-white matter contrast. We found a very close correspondence between the two modalities with their respective attenuation correction methods and sensory activation between scans; the different uptake times explain differences in absolute SUV levels and grey-white matter contrast. With awareness of these differences, brain PET-MR can be used clinically.