Characterization of spatial distortion in a 0.35 T MRI-guided radiotherapy system

Abstract

Spatial distortion results in image deformation that can degrade accurate targeting and dose calculations in MRI-guided adaptive radiotherapy. The authors present a comprehensive assessment of a 0.35 T MRI-guided radiotherapy system’s spatial distortion using two commercially-available phantoms with regularly spaced markers. Images of the spatial integrity phantoms were acquired using five clinical protocols on the MRI-guided radiotherapy machine with the radiotherapy gantry positioned at various angles. Software was developed to identify and localize all phantom markers using a template matching approach. Rotational and translational corrections were implemented to account for imperfect phantom alignment. Measurements were made to assess uncertainties arising from susceptibility artifacts, image noise, and phantom construction accuracy. For a clinical 3D imaging protocol with a 1.5 mm reconstructed slice thickness, 100% of spheres within a 50 mm radius of isocenter had a 3D deviation of 1 mm or less. Of the spheres within 100 mm of isocenter, 99.9% had a 3D deviation less than 1 mm. 94.8% and 100% of the spheres within 175 mm were found to be within 1 mm and 2 mm of the expected positions in 3D respectively. Maximum 3D distortions within 50 mm, 100 mm and 175 mm of isocenter were 0.76 mm, 1.15 mm and 1.88 mm respectively. Distortions present in images acquired using the real-time imaging sequence were less than 1 mm for 98.1% and 95.0% of the cylinders within 50 mm and 100 mm of isocenter. The corresponding maximum distortion in these regions was 1.10 mm and 1.67 mm. These results may be used to inform appropriate planning target volume (PTV) margins for 0.35 T MRI-guided radiotherapy. Observed levels of spatial distortion should be explicitly considered when using PTV margins of 3 mm or less or in the case of targets displaced from isocenter by more than 50 mm.

Evaluation of a magnetic resonance guided linear accelerator for stereotactic radiosurgery treatment

“…The mean and range of the CI, HI, GI, and PCI were 1.26 [1.02–1.83], 1.22 [1.11–1.42], 5.38 [4.12–7.69], and 0.80 [0.750–.84], respectively. These results for CI, PCI, and HI are consistent with published results for other single iso-multi target studies [4,6]. The GI values are generally higher than published values, mainly due to the contribution for Patient 3 (GI = 7.69).…”

“…Since a low field strength MR scanner was used for this system, B 0 inhomogeneity and susceptibility artifacts were expected to be much less than for conventional 1.5 T or 3 T scanner. Ginn et al investigated both 2D and 3D spatial distortion of a 0.35 T scanner using the same image sequence with various field of views [6]. Their results showed that the average distortion was ~0.3 mm within 100 mm of the central axis using a 1.5 mm slice thickness scan.…”

Multislice motion modeling for MRI‐guided radiotherapy gating

“…All imaging studies were performed under an IRB‐approved research protocol using the ViewRay MRIdian . Briefly, MRIdian combines three 60 Co treatment heads with a 0.35 T split bore MRI.…”

Large field of view distortion assessment in a low‐field MR‐linac

“…Routine MRI spatial distortion arising from all system‐ and object‐induced distortions of a clinical sequence (termed “total distortion”) was assessed monthly and after major equipment repairs over a period of 15 months. A rectangular box phantom (Fluke Biomedical, model 76‐907, signal‐generating solution containing ~400 cylindrical objects spaced 1.45 cm apart in a ~30 × 30 cm 2 area) was imaged using a 3D gradient, steady‐state sequence (TR/TE/FA = 3.36/1.44 ms/60°, Voxel size = 1.5 × 1.5 × 3.0 mm 3 ). Three images of the phantom were acquired with the long axis of the phantom oriented along the axial, sagittal, and coronal planes.…”