Correcting geometric image distortions in slice‐based 4D‐MRI on the MR‐linac

Abstract

Purpose: The importance of four-dimensional-magnetic resonance imaging (4D-MRI) is increasing in guiding online plan adaptation in thoracic and abdominal radiotherapy. Many 4D-MRI sequences are based on multislice two-dimensional (2D) acquisitions which provide contrast flexibility. Intrinsic to MRI, however, are machine-and subject-related geometric image distortions. Full correction of slice-based 4D-MRIs acquired on the Unity MR-linac (Elekta AB, Stockholm, Sweden) is challenging, since through-plane corrections are currently not available for 2D sequences. In this study, we implement a full three-dimensional 3D correction and quantify the geometric and dosimetric effects of machine-related (residual) geometric image distortions. Methods: A commercial three-dimensional (3D) geometric QA phantom (Philips, Best, the Netherlands) was used to quantify the effect of gradient nonlinearity (GNL) and static-field inhomogeneity (B0I) on geometric accuracy. Additionally, the effectiveness of 2D (in-plane, machine-generic), 3D (machine-generic), and in-house developed 3D þ (machine-specific) corrections was investigated. Corrections were based on deformable vector fields derived from spherical harmonics coefficients. Three patients with oligometastases in the liver were scanned with axial 4D-MRIs on our MR-linac (total: 10 imaging sessions). For each patient, a step-and-shoot IMRT plan (3 9 20 Gy) was created based on the simulation mid-position (midP)-CT. The 4D-MRIs were then warped into a daily midP-MRI and geometrically corrected. Next, the treatment plan was adapted according to the position offset of the tumor between midP-CT and the 3D-corrected midP-MRIs. The midP-CT was also deformably registered to the daily midP-MRIs (different corrections applied) to quantify the dosimetric effects of (residual) geometric image distortions. Results: Using phantom data, median GNL distortions were 0.58 mm (no correction), 0.42-0.48 mm (2D), 0.34 mm (3D), and 0.34 mm (3D þ ), measured over a diameter of spherical volume (DSV) of 200 mm. Median B0I distortions were 0.09 mm for the same DSV. For DSVs up to 500 mm, through-plane corrections are necessary to keep the median residual GNL distortion below 1 mm. 3D and 3D þ corrections agreed within 0.15 mm. 2D-corrected images featured uncorrected through-plane distortions of up to 21.11 mm at a distance of 20-25 cm from the machine's isocenter. Based on the 4D-MRI patient scans, the average external body contour distortions were 3.1 mm (uncorrected) and 1.2 mm (2D-corrected), with maximum local distortions of 9.5 mm in the uncorrected images. No (residual) distortions were visible for the metastases, which were all located within 10 cm of the machine's isocenter. The interquartile range (IQR) of dose differences between planned and daily dose caused by variable patient setup, patient anatomy, and online plan adaptation was 1.37 Gy/Fx for the PTV D95%. When comparing dose on 3D-corrected with uncorrected (2D-corrected) images, the IQR was 0.61 (0.31) Gy/Fx. Conclusions: GNL is the ma...

Fast geometric distortion correction using a deep neural network: Implementation for the 1 Tesla MRI‐Linac system

“…Further, a newly designed phantom incorporating with the registration method was used to correct geometric distortions for the 0.35 T MR‐Linac system, and the mean distortion after the correction was 0.60 ± 0.28 mm within a range of 14 cm 50 . Similarly, a recent study showed that the registration method has successfully reduced the median distortion below 1.0 mm in the DSV up to 50 cm 51 . The accuracy of the registration‐based method, however, depends on the convergence of the co‐registration with CT images, potentially making this method challenging to implement routinely 34…”

“…50 Similarly, a recent study showed that the registration method has successfully reduced the median distortion below 1.0 mm in the DSV up to 50 cm. 51 The accuracy of the registration-based method, however, depends on the convergence of the co-registration with CT images, potentially making this method challenging to implement routinely. 34 Standard Cartesian TSE sequences were used for scanning the 3D phantom and human pelvis in this examination.…”

Applicator visualization using ultrashort echo time MRI for high-dose-rate endorectal brachytherapy

“…To quantify geometric distortions, we acquired MRI images of a commercial 3D geometric QA phantom (Philips Healthcare, Best, The Netherlands) (13,14). The phantom was placed on the treatment table centered around the isocenter.…”

“…Subsequently, we acquired a UTE image with the same isotropic voxel size as the two axial 3D T 1weighted images ( Table 1). The marker detection algorithm described in the study by Keesman et al (13) was used to detect the markers, which produced lists of marker positions, one for each image.…”

Tumor-site specific geometric distortions in high field integrated magnetic resonance linear accelerator radiotherapy

“…air cavities and tissue boundaries) distorting the local B 0 field [9]. Geometric distortion in MR-linacs has been studied by others [9][10][11][12][13][14][15], i.e. in the study by Tijssen et al where a method for cumulative total geometric distortion is presented [10], but larger sample sizes and anatomical tumor-site specific investigation is still warranted.…”