Assessment of geometric distortion of eleven clinical MRI scanners across the UK

Abstract

Purpose or Objective MRI is increasingly being used in radiotherapy (RT) in the UK for delineation, MRI-only planning and image guidance on an MR-linac. Geometric distortions are a known challenge of using MRI in RT. The aim of this work was to assess large FOV MRI distortions in a number of scanners used for RT throughout the UK. Material and Methods MRI scans of a GRADE phantom (Spectronic Medical AB, Helsingborg, Sweden) were acquired using a protocol recommended by the manufacturer (3D gradient echo, FoV 50x50x50 cm, voxel size 1x1x2 mm and Bandwidth 490 Hz/pixel) on the following 11 MRI scanners that are used for RT in the UK: Siemens Aera (Leeds, Guy’s and St Thomas’, Royal Marsden and Christie); Siemens Prisma (Leeds); Siemens Sola (Leeds); Siemens Skyra (Christie); Siemens Espree (Newcastle); Elekta Unity (Christie); Philips Ingenia 1.5T (Christie); GE Signa PET/MR (Newcastle). Scanners are referred to as 1-11 in a randomised order (for anonymity). Analysis was performed using MriPlanner (Spectronic Medical AB, Helsingborg, Sweden), the mean and max distortions and variance (δ2) between scanners were reported at <100, 100-150, 150- 200, 200-250 and ≥250 mm from the isocentre. Results The range of the mean and max distortions for all scanners at 100-150 mm (representing the typical size of a brain) were 0.28-0.62 mm and 0.69-2.81 mm respectively. The range of the mean and max distortions for all scanners at 200-250 mm (representing the typical size of a pelvis) were 0.67-2.52 mm and 2.0-10.75 mm respectively. The mean and maximum distortions are shown in figure 1 and at <100, 100-150, 150-200, 200-250 and ≥250 mm from the isocentre, δ2 was 0.004, 0.009, 0.045, 0.232 and 0.779 mm2 respectively. The greatest mean geometric distortion (5.04 mm) observed was scanner 9 at ≥250 mm from the isocentre. For uses requiring a small FOV (less than 150 mm from the isocentre) such as brain imaging, both mean and maximum distortions would be < 2 mm except for the maximum distortion of one scanner. For uses requiring a large FOV (up to 250 mm from the isocentre) such as the head and neck and the entire pelvis, the mean distortions would be < 2 mm except for those from one scanner. Conclusion Geometric distortion was assessed on 11 scanners used for RT across the UK and for all scanners distortion increased with distance to isocentre. Variance of geometric distortions between scanners was found to be low (<0.8) and increased with distance to isocentre. The vast majority of scanners showed mean distortions of > 2 mm at a distance ≥ 250 mm from the isocentre, but these values have limited clinical relevance. This assessment demonstrates feasibility of the technique to be repeated in a UK wide audit of all MRI scanners used clinically for RT.