Geometric distortion characterization and correction for the 1.0 T Australian MRI‐linac system using an inverse electromagnetic method

Shan, Shanshan; Liney, Gary; Tang, Fang‐Fang; Li, Mingyan; Wang, Yaohui; Ma, Huan; Weber, Ewald; Walker, Amy; Holloway, Lois; Wang, Qiuliang; Wang, Biao; F, Liu; Crozier, Stuart

doi:10.1002/mp.13979

Cited by 15 publications

(35 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…33 B 0 map results in each plane (2 × standard deviation of frequency shift) were as follows: axial plane (33 Hz), coronal plane (110 Hz), and sagittal plane (80 Hz), which showed that the largest distortion caused by the B 0 inhomogeneity was only half pixel in the ROI for the image acquisition protocol used here. 20 Two MR image sets scanned by the 1T MRI-Linac system were used to evaluate the performance of the proposed method. One set was acquired by scanning the phantom at various locations with turbo spin-echo (TSE) sequence, where no distortion corrections were implemented.…”

Section: A Data Preparationmentioning

confidence: 99%

“…For any positions within the ROI, the GNL deviations can be calculated by the GNLNet. Based on the obtained GNL deviation information, a "GNL-encoding" reconstruction method developed in our previous studies was used to restore the undistorted image by the equation 20,29 :…”

Section: G Model Trainingmentioning

confidence: 99%

“…[Color figure can be viewed at wileyonlinelibrary.com] MR image. 20 By contrast, although the SH method is much faster than the inverse EM model, the inferior correction accuracy in particular for through-plane distortions, see results in Table I, obstructs it to be used in high-precision radiotherapy. The presented GNLNet accounts for the lowest computational complexity without a compromise of correction accuracy.…”

Section: B Computational Complexitymentioning

confidence: 99%

“…A fifth-order SH expansion has been widely used in commercial whole-body MRI scanners, and it has shown effectiveness for correcting geometric distortions within the working DSV. 19 However, the correction accuracy of the fifth-order SH expansion method suffers from degradation on the boundaries of or outside the DSV, 20 due to a rapid field fall off. To enhance the capability of characterizing GNL outside the working DSV, higher (e.g., 7th-11th) order SH expansion methods have been investigated.…”

Section: Introductionmentioning

confidence: 99%

“…21 Recently, we developed an inverse electromagnetic (EM) method to simulate gradient field deviations both inside and outside the 30-cm DSV on the 1T MRI-Linac system. 20 Field samples measured by a grid phantom were used to inversely reconstitute the virtual current density distribution, and then the stream function was used to describe GNL fields in the entire region of interest (ROI). The experimental results showed that the inverse EM model could successfully improve the geometric accuracy in areas outside the DSV compared to the fifth-order SH method.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

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

Shan

Chandra

et al. 2020

Medical Physics

Self Cite

View full text Add to dashboard Cite

Purpose: Combining high-resolution magnetic resonance imaging (MRI) with a linear accelerator (Linac) as a single MRI-Linac system provides the capability to monitor intra-fractional motion and anatomical changes during radiotherapy, which facilitates more accurate delivery of radiation dose to the tumor and less exposure to healthy tissue. The gradient nonlinearity (GNL)-induced distortions in MRI, however, hinder the implementation of MRI-Linac system in image-guided radiotherapy where highly accurate geometry and anatomy of the target tumor is indispensable. Methods: To correct the geometric distortions in MR images, in particular, for the 1 Tesla (T) MRI-Linac system, a deep fully connected neural network was proposed to automatically learn the intricate relationship between the undistorted (theoretical) and distorted (real) space. A dataset, consisting of spatial samples acquired by phantom measurement that covers both inside and outside the working diameter of spherical volume (DSV), was utilized for training the neural network, which offers the ability to describe subtle deviations of the GNL field within the entire region of interest (ROI). Results: The performance of the proposed method was evaluated on MR images of a three-dimensional (3D) phantom and the pelvic region of an adult volunteer scanned in the 1T MRI-Linac system. The experimental results showed that the severe geometric distortions within the entire ROI had been successfully corrected with an error less than the pixel size. Also, the presented network is highly efficient, which achieved significant improvement in terms of computational efficiency compared to existing methods. Conclusions: The feasibility of the presented deep neural network for characterizing the GNL field deviations in the 1T MRI-Linac system was demonstrated in this study, which shows promise in facilitating the MRI-Linac system to be routinely implemented in real-time MRI-guided radiotherapy.

show abstract

Section: A Data Preparationmentioning

confidence: 99%

Section: G Model Trainingmentioning

confidence: 99%

Section: B Computational Complexitymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

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

Shan

Chandra

et al. 2020

Medical Physics

Self Cite

View full text Add to dashboard Cite

show abstract

Characterizing magnetically focused contamination electrons by off‐axis irradiation on an inline MRI‐Linac

Patterson¹,

Oborn

Cutajar³

et al. 2022

J Applied Clin Med Phys

View full text Add to dashboard Cite

The aim of this study is to investigate off -axis irradiation on the Australian MRI-Linac using experiments and Monte Carlo simulations. Simulations are used to verify experimental measurements and to determine the minimum offset distance required to separate electron contamination from the photon field. Methods: Dosimetric measurements were performed using a microDiamond detector, Gafchromic ® EBT3 film, and MOSkin TM . Three field sizes were investigated including 1.9 × 1.9, 5.8 × 5.8, and 9.7 × 9.6 cm 2 . Each field was offset a maximum distance, approximately 10 cm, from the central magnetic axis (isocenter). Percentage depth doses (PDDs) were collected at a source-tosurface distance (SSD) of 1.8 m for fields collimated centrally and off -axis. PDD measurements were also acquired at isocenter for each off -axis field to measure electron contamination. Monte Carlo simulations were used to verify experimental measurements, determine the minimum field offset distance, and demonstrate the use of a spoiler to absorb electron contamination. Results: Off -axis irradiation separates the majority of electron contamination from an x-ray beam and was found to significantly reduce in-field surface dose. For the 1.9 × 1.9, 5.8 × 5.8, and 9.7 × 9.6 cm 2 field, surface dose was reduced from 120.9% to 24.9%,229.7% to 39.2%,and 355.3% to 47.3%,respectively.Monte Carlo simulations generally were within experimental error to MOSkin TM and microDiamond, and used to determine the minimum offset distance, 2.1 cm, from the field edge to isocenter. A water spoiler 2 cm thick was shown to reduce electron contamination dose to near zero. Conclusions: Experimental and simulation data were acquired for a range of field sizes to investigate off -axis irradiation on an inline MRI-Linac. The skin sparing effect was observed with off -axis irradiation, a feature that cannot be achieved to the same extent with other methods, such as bolusing, for beams at isocenter.

show abstract

First experimental investigation of simultaneously tracking two independently moving targets on an MRI‐linac using real‐time MRI and MLC tracking

et al. 2020

Self Cite

View full text Add to dashboard Cite

Purpose High quality radiotherapy is challenging in cases where multiple targets with independent motion are simultaneously treated. A real‐time tumor tracking system that can simultaneously account for the motion of two targets was developed and characterized. Methods The multitarget tracking system was implemented on a magnetic resonance imaging (MRI)‐linac and utilized multi‐leaf collimator (MLC) tracking to adapt the radiation beam to phantom targets reproducing motion with prostate and lung motion traces. Multitarget tracking consisted of three stages: (a) pretreatment aperture segmentation where the treatment aperture was divided into segments corresponding to each target, (b) MR imaging where the positions of the two targets were localized, and (c) MLC tracking where an updated treatment aperture was calculated. Electronic portal images (EPID) acquired during irradiation were analyzed to characterize geometric uncertainty and tracking latency. Results Multitarget MLC tracking effectively accounted for the motion of both targets during treatment. The root‐mean‐square error between the centers of the targets and the centers of the corresponding MLC leaves were reduced from 5.5 mm without tracking to 2.7 mm with tracking for lung motion traces and reduced from 4.2 to 1.4 mm for prostate motion traces. The end‐to‐end latency of tracking was measured to be 328 ± 44 ms. Conclusions We have demonstrated the first experimental implementation of MLC tracking for multiple targets having independent motion. This technology takes advantage of the imaging capabilities of MRI‐linacs and would allow treatment margins to be reduced in cases where multiple targets are simultaneously treated.

show abstract

Geometric distortion characterization and correction for the 1.0 T Australian MRI‐linac system using an inverse electromagnetic method

Cited by 15 publications

References 38 publications

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

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

Characterizing magnetically focused contamination electrons by off‐axis irradiation on an inline MRI‐Linac

First experimental investigation of simultaneously tracking two independently moving targets on an MRI‐linac using real‐time MRI and MLC tracking

Contact Info

Product

Resources

About