[OA141] Preparing data for multiparametric PET/MRI: Influence of PET point spread function modelling and EPI distortion correction on the spatial correlation of 18F-FDG-PET uptake and diffusion weighted MRI in head/neck cancer. (August 2018)

• Record Type: Journal Article
• Authors: Krogager, Louise; Rasmussen, Jacob Høygaard; Andersen, Flemming Littrup; Specht, Lena; Beyer, Thomas; Kjaer, Andreas; Fischer, Barbara Malene; Hansen, Adam Espe
• Is Part Of: Physica medica. Volume 52(2018)Supplement 1
• Page Start: 54
• Publication Date: 2018-08
• Subjects: Medical physics -- Periodicals; Biophysics -- Periodicals; Biophysics -- Periodicals; Imagerie médicale -- Périodiques; Radiothérapie -- Périodiques; Rayons X -- Sécurité -- Mesures -- Périodiques; Physique -- Périodiques; Médecine -- Périodiques; 610.153
• DOI: 10.1016/j.ejmp.2018.06.213
• Languages: English
• ISSNs: 1120-1797

Abstract

Abstract : Purpose: Multiparametric imaging may improve disease characterization. The combination of 18 F-flourodeoxyglucose (FDG) Positron Emission Tomography (PET) and Diffusion-Weighted Magnetic Resonance Imaging (DW-MRI) has been suggested to improve the assessment of tumour heterogeneity[1] . However, PET image resolution is limited compared to MRI while DW-MRI is prone to image distortions, potentially compromising the spatial correspondence of the modalities. The aim of this study was to assess relative differences and improvements of the quality of PET and DW-MRI by employing Point Spread Function (PSF) modelling and distortion correction, respectively. Methods: Data was collected from 11 patients with head/neck cancers, each scanned on a fully-integrated PET/MRI (Siemens Biograph mMR) twice on separate days[2] . PET [60 min post injection of 18 F-FDG (4MBq/kg)] was reconstructed [3D OP-OSEM, 3 iterations, 21 subsets] with and without PSF modelling. DW-MRI [repetition time/echo time 3000/84 ms, parallel imaging factor 4, effective echo spacing 0.145 ms] was corrected for distortions using B0-mapping and compared to an anatomical reference [T2-weighted turbo spin echo]. Voxelwise correlations of FDG uptake and the apparent diffusion coefficient (ADC) derived from DW-MRI were measured in tumours. Results: The inclusion of PSF modelling in PET image reconstruction increased overall FDG uptake in tumours [max: 5 %, mean: 12%] and enhanced image quality. Distortion correction improved DW-MRI fidelity compared to the reference. The impact of PSF modelling on voxelwise correlation coefficients of FDG uptake and the ADC was not significant [p = 0.17]. DW-MRI distortion correction had a large and significant influence on correlation coefficients [p = 0.001]. Conclusions: This study investigated the consequences of improving quality of PET and DW-MRI for the spatial correlation of the modalities in patients with head/neck cancers. Voxelwise correlation of FDG uptake and ADC was affected by DW-MRI distortion correction, but not by using PSF modelling during PET image reconstruction. Accordingly, optimized data acquisition and processing can help increase the value of PET/MRI data for voxelwise multiparametric analysis.