Value of a Dixon-based MR/PET attenuation correction sequence for the localization and evaluation of PET-positive lesions

Purpose In this study, the potential contribution of Dixon-based MR imaging with a rapid low-resolution breath-hold sequence, which is a technique used for MR-based attenuation correction (AC) for MR/positron emission tomography (PET), was evaluated for anatomical correlation of PET-positive lesions on a 3T clinical scanner compared to low-dose CT. This technique is also used in a recently installed fully integrated whole-body MR/PET system. Methods Thirty-five patients routinely scheduled for oncological staging underwent 18 F-fluorodeoxyglucose (FDG) PET/CT and a 2-point Dixon 3-D volumetric interpolated breath-hold examination (VIBE) T1-weighted MR sequence on the same day. Two PET data sets reconstructed using attenuation maps from low-dose CT (PET AC_CT ) or simulated MR-based segmentation (PET AC_MR ) were evaluated for focal PET-positive lesions. The certainty for the correlation with anatomical structures was judged in the low-dose CT and Dixon-based MRI on a 4-point scale (0–3). In addition, the standardized uptake values (SUVs) for PET AC_CT and PET AC_MR were compared. Results Statistically, no significant difference could be found concerning anatomical localization for all 81 PET-positive lesions in low-dose CT compared to Dixon-based MR (mean 2.51 ± 0.85 and 2.37 ± 0.87, respectively; p = 0.1909). CT tended to be superior for small lymph nodes, bone metastases and pulmonary nodules, while Dixon-based MR proved advantageous for soft tissue pathologies like head/neck tumours and liver metastases. For the PET AC_CT - and PET AC_MR -based SUVs (mean 6.36 ± 4.47 and 6.31 ± 4.52, respectively) a nearly complete concordance with a highly significant correlation was found ( r = 0.9975, p