Geometric distortions of diffusion weighted imaging of the head/neck in combined PET/MR: optimization of image acquisition and post-processing correction for oncology applications

Abstract

The combination of PET with Diffusion Weighted Imaging (DWI) is a promising application of PET/MR. DWI geometric accuracy can be compromised in body regions with complex anatomy. Here, we assess DWI head/neck image quality following optimization of acquisition and post-processing.Preliminary data from 10 patients with cancer of the tonsil or base-of-tongue is presented. An integrated PET/MR system (Siemens Biograph mMR) with a 3 T magnet was used.PET was performed as a single-bed, 20 min acquisition, 120 min post injection of 4 MBq/kg [ 18 F]-FDG.MRI included axial T2 weighted STIR, B0 mapping and 2 DWI single-shot EPI measurements (DWI1 and DWI2) with b values 0, 500 and 1000 mm 2 /s. DWI2 was measured as 3 stacks, each at isocenter to maximize field homogeneity. DWI1/DWI2 had an effective echo spacing 0.375/0.145 ms. For DWI2, distortions were corrected using FSL with FUGUE and Topup algorithms. DWI geometric quality was evaluated in 44×44 mm 2 region centred on the tumor, with the STIR image as anatomical reference. Correlation and DICE coefficients between DWI b0 and STIR images were computed, for all DWI image sets. Geometric quality of DWI1 was poor, but improved both by optimization of acquisition (DWI2) and by post-processing. Visually, hyperintensities on the Topup corrected DWI2 images matched closely the STIR images. Also, regions with high FDG uptake and low diffusion had matching shapes. Correlation and DICE coefficients had a large variation for DWI1, and showed an overall increase to a high level >0.6 and 0.8, respectively, following optimization of image acquisition and post-processing correction.Diffusion weighted images of the head/neck with a quality suitable for dual-modality assessment of tumour characteristics on a voxel basis can be obtained following the optimization of acquisition and post-processing.Hansen et al. EJNMMI Physics 2014, 1(Suppl 1):A76

Preparing data for multiparametric PET/MR imaging: Influence of PET point spread function modelling and EPI distortion correction on the spatial correlation of [18F]FDG-PET and diffusion-weighted MRI in head and neck cancer

“…The current study also demonstrated that performing EPI distortion correction by FUGUE on the optimized DWI protocol further reduced geometric distortions significantly. Distortion correction with B 0 -field mapping are commonly applied for fMRI [20], but only recently for oncology studies [8,22,28,29]. However, with the increasing interest in preserving spatial accuracy, e.g.…”

“…EPI distortions were corrected using the algorithm FMRIB's Utility for Geometrically Unwarping EPIs (FUGUE) of the FSL software package (Analysis Group, FMRIB where FMRIB is Oxford Centre for Functional MRI of the Brain) [22]. FUGUE reduces artefacts by using a computed B 0 -field map based on two collected gradient echo images and a corresponding phase map.…”

“…Distortions may also be corrected with post-processing techniques, e.g., based on B 0 -field mapping [20]. Such techniques have frequently been utilized for brain fMRI [21] but only very little in context of oncology [8,22].…”

Feasibility of Multiparametric Imaging with PET/MR in Head and Neck Squamous Cell Carcinoma

“…Both image acquisition and postprocessing were optimized to minimize distortions. The visually judged correspondence of DWI and 18 F-FDG PET was overall good, as seen in the figures (25). After this FUGUE correction was implemented, no patient had to be excluded because of signal drop out (Supplemental Fig.…”

“…DWI was measured as 3 stacks, each at isocenter to maximize field homogeneity. DWI had an effective echo spacing of 0.145 ms, and DWI distortions were corrected using the algorithm (FMRIB's Utility for Geometrically Unwarping EPIs [where FMRIB is Oxford Centre for Functional MRI of the Brain] [FUGUE]) of the FSL software package (Analysis Group, FMRIB) (25).…”