Discrimination and anatomical mapping of PET-positive lesions: comparison of CT attenuation-corrected PET images with coregistered MR and CT images in the abdomen

Abstract

PET/MR has the potential to become a powerful tool in clinical oncological imaging. The purpose of this prospective study was to evaluate the performance of a single T1-weighted (T1w) fat-suppressed unenhanced MR pulse sequence of the abdomen in comparison with unenhanced low-dose CT images to characterize PET-positive lesions. A total of 100 oncological patients underwent sequential whole-body (18)F-FDG PET with CT-based attenuation correction (AC), 40mAs low-dose CT and two-point Dixon-based T1w 3D MRI of the abdomen in a trimodality PET/CT-MR system. PET-positive lesions were assessed by CT and MRI with regard to their anatomical location, conspicuity and additional relevant information for characterization. From among 66 patients with at least one PET-positive lesion, 147 lesions were evaluated. No significant difference between MRI and CT was found regarding anatomical lesion localization. The MR pulse sequence used performed significantly better than CT regarding conspicuity of liver lesions (p < 0.001, Wilcoxon signed ranks test), whereas no difference was noted for extrahepatic lesions. For overall lesion characterization, MRI was considered superior to CT in 40% of lesions, equal to CT in 49%, and inferior to CT in 11%. Fast Dixon-based T1w MRI outperformed low-dose CT in terms of conspicuity and characterization of PET-positive liver lesions and performed similarly in extrahepatic tumour manifestations. Hence, under the assumption that the technical issue of MR AC for whole-body PET examinations is solved, in abdominal PET/MR imaging the replacement of low-dose CT by a single Dixon-based MR pulse sequence for anatomical lesion correlation appears to be valid and robust.