Characteristics of geometric distortion correction with increasing field-of-view in open-configuration MRI

SC-GAN: Structure-completion generative adversarial network for synthetic CT generation from MR images with truncated anatomy

Creating synthetic CT (sCT) from magnetic resonance (MR) images enables MR-based treatment planning in radiation therapy. However, the MR images used for MR-guided adaptive planning are often truncated in the boundary regions due to the limited field of view and the need for sequence optimization. Consequently, the sCT generated from these truncated MR images lacks complete anatomic information, leading to dose calculation error for MR-based adaptive planning. We propose a novel structure-completion generative adversarial network (SC-GAN) to generate sCT with full anatomic details from the truncated MR images. To enable anatomy compensation, we expand input channels of the CT generator by including a body mask and introduce a truncation loss between sCT and real CT. The body mask for each patient was automatically created from the simulation CT scans and transformed to daily MR images by rigid registration as another input for our SC-GAN in addition to the MR images. The truncation loss was constructed by implementing either an auto-segmentor or an edge detector to penalize the difference in body outlines between sCT and real CT. The experimental results show that our SC-GAN achieved much improved accuracy of sCT generation in both truncated and untruncated regions compared to the original cycleGAN and conditional GAN methods.

Compensation cycle consistent generative adversarial networks (Comp-GAN) for synthetic CT generation from MR scans with truncated anatomy

BACKGROUND

MR scans used in radiotherapy can be partially truncated due to the limited field of view (FOV), affecting dose calculation accuracy in MR-based radiation treatment planning.

PURPOSE

We proposed a novel Compensation-cycleGAN (Comp-cycleGAN) by modifying the cycle-consistent generative adversarial network (cycleGAN), to simultaneously create synthetic CT (sCT) images and compensate the missing anatomy from the truncated MR images.

METHODS

Computed tomography (CT) and T1 MR images with complete anatomy of 79 head-and-neck patients were used for this study. The original MR images were manually cropped 10-25 mm off at the posterior head to simulate clinically truncated MR images. Fifteen patients were randomly chosen for testing and the rest of the patients were used for model training and validation. Both the truncated and original MR images were used in the Comp-cycleGAN training stage, which enables the model to compensate for the missing anatomy by learning the relationship between the truncation and known structures. After the model was trained, sCT images with complete anatomy can be generated by feeding only the truncated MR images into the model. In addition, the external body contours acquired from the CT images with full anatomy could be an optional input for the proposed method to leverage the additional information of the actual body shape for each test patient. The mean absolute error (MAE) of Hounsfield units (HU), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM) were calculated between sCT and real CT images to quantify the overall sCT performance. To further evaluate the shape accuracy, we generated the external body contours for sCT and original MR images with full anatomy. The Dice similarity coefficient (DSC) and mean surface distance (MSD) were calculated between the body contours of sCT and original MR images for the truncation region to assess the anatomy compensation accuracy.

RESULTS

The average MAE, PSNR, and SSIM calculated over test patients were 93.1 HU/91.3 HU, 26.5 dB/27.4 dB, and 0.94/0.94 for the proposed Comp-cycleGAN models trained without/with body-contour information, respectively. These results were comparable with those obtained from the cycleGAN model which is trained and tested on full-anatomy MR images, indicating the high quality of the sCT generated from truncated MR images by the proposed method. Within the truncated region, the mean DSC and MSD were 0.85/0.89 and 1.3/0.7 mm for the proposed Comp-cycleGAN models trained without/with body contour information, demonstrating good performance in compensating the truncated anatomy.

CONCLUSIONS

We developed a novel Comp-cycleGAN model that can effectively create sCT with complete anatomy compensation from truncated MR images, which could potentially benefit the MRI-based treatment planning.

Assessment of MRI image distortion based on 6 consecutive years of annual QAs and measurements on 14 MRI scanners used for radiation therapy

PURPOSE

To determine the magnitude of MRI image distortion based on 6 consecutive years of annual quality assurances/measurements on 14 MRI scanners used for radiation therapy and to provide evidence for the inclusion of additional margin for treatment planning.

METHODS AND MATERIALS

We used commercial MRI image phantoms to quantitatively study the MRI image distortion over period of 6 years for up to 14 1.5 and 3 T MRI scanners that could potentially be used to provide MRI images for treatment planning. With the phantom images collected from 2016 to 2022, we investigated the MRI image distortion, the dependence of distortion on the distance from the imaging isocenter, and the possible causes of large distortion discovered.

RESULTS

MRI image distortion increases with the distance from the imaging isocenter. For a region of interest (ROI) with a radius of 100 mm centered at the isocenter, the mean magnitude of distortion for all MRI scanners is 0.44 ± 0.18 mm $0.44 \pm 0.18\;{\rm{mm}}$ , and the maximum distortion varies from 0.52 to 1.31 mm $0.52\;{\rm{to}}\;1.31\;{\rm{mm}}$ depending on MRI scanners. For an ROI with a radius of 200 mm centered at the isocenter, the mean magnitude of distortion increases to 0.84 ± 0.45 mm $0.84 \pm 0.45\;{\rm{mm}}$ , and the range of the maximum distortion increases to 1.92 - 5.03 mm $1.92 - 5.03\;{\rm{mm}}$ depending on MRI scanners. The distortion could reach 2 mm at 150 mm from the isocenter.

CONCLUSION

An additional margin to accommodate image distortion should be considered for treatment planning. Imaging with proper patient alignment to the isocenter is vital to reducing image distortion. We recommend performing image distortion checks annually and after major upgrade on MRI scanners.

Technical Note: Extended field-of-view (FOV) MRI distortion determination through multi-positional phantom imaging

Comprehensive characterization of geometric distortions for MRI simulators and MRI‐guided treatment delivery systems is typically performed with large phantoms that are costly and unwieldy to handle. Here we propose an easily implementable methodology for MR distortion determination of the entire imaging space of the scanner through the use of a compact commercially available distortion phantom. The MagphanRT phantom was scanned at several locations within a MR scanner. From each scan, an approximate location of the phantom was determined from a subset of the fiducial spheres. The fiducial displacements were determined, and a displacement field was fitted to the displacement data using the entire multi‐scan data set. An orthogonal polynomial expansion fitting function was used that had been augmented to include independent rigid‐body transformations for each scan. The rigid‐body portions of the displacement field were thereafter discarded, and the resultant fit then represented the distortion field. Multi‐positional scans of the phantom were used successfully to determine the distortion field with extended coverage. A single scan of the phantom covered 20 cm in its smallest dimension. By stitching together overlapping scans we extended the distortion measurements to 30 cm. No information about the absolute location or orientation of each scan was required. The method, termed the Multi‐Scan Expansion (MSE) method, can be easily applied for larger field‐of‐views (FOVs) by using a combination of larger phantom displacements and more scans. The implementation of the MSE method allows for distortion determination beyond the physical limitations of the phantom. The method is scalable to the user’s needs and does not require any specialized equipment. This approach could open up for easier determination of the distortion magnitude at distances further from the scanner’s isocenter. This is especially important in the newly proposed methodologies of MR‐only simulation in RT and in adaptive replanning in MR linac systems.

Characterizing geometrical accuracy in clinically optimised 7T and 3T magnetic resonance images for high-precision radiation treatment of brain tumours

Background and purpose

In neuro-oncology, high spatial accuracy is needed for clinically acceptable high-precision radiation treatment planning (RTP). In this study, the clinical applicability of anatomically optimised 7-Tesla (7T) MR images for reliable RTP is assessed with respect to standard clinical imaging modalities.

Materials and methods

System- and phantom-related geometrical distortion (GD) were quantified on clinically-relevant MR sequences at 7T and 3T, and on CT images using a dedicated anthropomorphic head phantom incorporating a 3D grid-structure, creating 436 points-of-interest. Global GD was assessed by mean absolute deviation (MAD_Global). Local GD relative to the magnetic isocentre was assessed by MAD_Local. Using 3D displacement vectors of individual points-of-interest, GD maps were created. For clinically acceptable radiotherapy, 7T images need to meet the criteria for accurate dose delivery (GD < 1 mm) and present comparable GD as tolerated in clinically standard 3T MR/CT-based RTP.

Results

MAD_Global in 7T and 3T images ranged from 0.3 to 2.2 mm and 0.2-0.8 mm, respectively. MAD_Local increased with increasing distance from the isocentre, showed an anisotropic distribution, and was significantly larger in 7T MR sequences (MAD_Local = 0.2-1.2 mm) than in 3T (MAD_Local = 0.1-0.7 mm) (p < 0.05). Significant differences in GD were detected between 7T images (p < 0.001). However, maximum MAD_Local remained ≤1 mm within 68.7 mm diameter spherical volume. No significant differences in GD were found between 7T and 3T protocols near the isocentre.

Conclusions

System- and phantom-related GD remained ≤1 mm in central brain regions, suggesting that 7T MR images could be implemented in radiotherapy with clinically acceptable spatial accuracy and equally tolerated GD as in 3T MR/CT-based RTP. For peripheral regions, GD should be incorporated in safety margins for treatment uncertainties. Moreover, the effects of sequence-related factors on GD needs further investigation to obtain RTP-specific MR protocols.

Technical challenges of quantitative chest MRI data analysis in a large cohort pediatric study

Objectives

This study was conducted in order to evaluate the effect of geometric distortion (GD) on MRI lung volume quantification and evaluate available manual, semi-automated, and fully automated methods for lung segmentation.

Methods

A phantom was scanned with MRI and CT. GD was quantified as the difference in phantom’s volume between MRI and CT, with CT as gold standard. Dice scores were used to measure overlap in shapes. Furthermore, 11 subjects from a prospective population-based cohort study each underwent four chest MRI acquisitions. The resulting 44 MRI scans with 2D and 3D Gradwarp were used to test five segmentation methods. Intraclass correlation coefficient, Bland–Altman plots, Wilcoxon, Mann–Whitney U, and paired t tests were used for statistics.

Results

Using phantoms, volume differences between CT and MRI varied according to MRI positions and 2D and 3D Gradwarp correction. With the phantom located at the isocenter, MRI overestimated the volume relative to CT by 5.56 ± 1.16 to 6.99 ± 0.22% with body and torso coils, respectively. Higher Dice scores and smaller intraobject differences were found for 3D Gradwarp MR images. In subjects, semi-automated and fully automated segmentation tools showed high agreement with manual segmentations (ICC = 0.971–0.993 for end-inspiratory scans; ICC = 0.992–0.995 for end-expiratory scans). Manual segmentation time per scan was approximately 3–4 h and 2–3 min for fully automated methods.

Conclusions

Volume overestimation of MRI due to GD can be quantified. Semi-automated and fully automated segmentation methods allow accurate, reproducible, and fast lung volume quantification. Chest MRI can be a valid radiation-free imaging modality for lung segmentation and volume quantification in large cohort studies.

Key Points

• Geometric distortion varies according to MRI setting and patient positioning.

• Automated segmentation methods allow fast and accurate lung volume quantification.

• MRI is a valid radiation-free alternative to CT for quantitative data analysis.

Electronic supplementary material

The online version of this article (10.1007/s00330-018-5863-7) contains supplementary material, which is available to authorized users.

Estimating Accurate Target Coordinates with Magnetic Resonance Images by Using Multiple Phase-Encoding Directions during Acquisition

BACKGROUND

Stereotactic procedures are image guided, often using magnetic resonance (MR) images limited by image distortion, which may influence targets for stereotactic procedures.

OBJECTIVES

The aim of this work was to assess methods of identifying target coordinates for stereotactic procedures with MR in multiple phase-encoding directions.

METHODS

In 30 patients undergoing deep brain stimulation, we acquired 5 image sets: stereotactic brain computed tomography (CT), T2-weighted images (T2WI), and T1WI in both right-to-left (RL) and anterior-to-posterior (AP) phase-encoding directions. Using CT coordinates as a reference, we analyzed anterior commissure and posterior commissure coordinates to identify any distortion relating to phase-encoding direction.

RESULTS

Compared with CT coordinates, RL-directed images had more positive x-axis values (0.51 mm in T1WI, 0.58 mm in T2WI). AP-directed images had more negative y-axis values (0.44 mm in T1WI, 0.59 mm in T2WI). We adopted 2 methods to predict CT coordinates with MR image sets: parallel translation and selective choice of axes according to phase-encoding direction. Both were equally effective at predicting CT coordinates using only MR; however, the latter may be easier to use in clinical settings.

CONCLUSION

Acquiring MR in multiple phase-encoding directions and selecting axes according to the phase-encoding direction allows identification of more accurate coordinates for stereotactic procedures.

Prospective analysis of in vivo landmark point-based MRI geometric distortion in head and neck cancer patients scanned in immobilized radiation treatment position: Results of a prospective quality assurance protocol

Purpose

Uncertainties related to geometric distortion are a major obstacle for effectively utilizing MRI in radiation oncology. We aim to quantify the geometric distortion in patient images by comparing their in-treatment position MRIs with the corresponding planning CTs, using CT as the non-distorted gold standard.

Methods

Twenty-one head and neck cancer patients were imaged with MRI as part of a prospective Institutional Review Board approved study. MR images were acquired with a T2 SE sequence (0.5 × 0.5 × 2.5 mm voxel size) in the same immobilization position as in the CTs. MRI to CT rigid registration was then done and geometric distortion comparison was assessed by measuring the corresponding anatomical landmarks on both the MRI and the CT images. Several landmark measurements were obtained including; skin to skin (STS), bone to bone, and soft tissue to soft tissue at specific levels in horizontal and vertical planes of both scans. Inter-observer variability was assessed and interclass correlation (ICC) was calculated.

Results

A total of 430 landmark measurements were obtained. The median distortion for all landmarks in all scans was 1.06 mm (IQR 0.6–1.98). For each patient 48% of the measurements were done in the right-left direction and 52% were done in the anteroposterior direction. The measured geometric distortion was not statistically different in the right-left direction compared to the anteroposterior direction (1.5 ± 1.6 vs. 1.6 ± 1.7 mm, respectively, p = 0.4). The magnitude of distortion was higher in the STS peripheral landmarks compared to the more central landmarks (2.0 ± 1.9 vs. 1.2 ± 1.3 mm, p < 0.0001). The mean distortion measured by observer one was not significantly different compared to observer 2, 3, and 4 (1.05, 1.23, 1.06 and 1.05 mm, respectively, p = 0.4) with ICC = 0.84.

Conclusion

MRI geometric distortions were quantified in radiotherapy planning applications with a clinically insignificant error of less than 2 mm compared to the gold standard CT.

Technical Note: Characterization and correction of gradient nonlinearity induced distortion on a 1.0 T open bore MR-SIM

PURPOSE

Distortions in magnetic resonance imaging (MRI) compromise spatial fidelity, potentially impacting delineation and dose calculation. We characterized 2D and 3D large field of view (FOV), sequence-independent distortion at various positions in a 1.0 T high-field open MR simulator (MR-SIM) to implement correction maps for MRI treatment planning.

METHODS

A 36 × 43 × 2 cm(3) phantom with 255 known landmarks (∼1 mm(3)) was scanned using 1.0 T high-field open MR-SIM at isocenter in the transverse, sagittal, and coronal axes, and a 465 × 350 × 168 mm(3) 3D phantom was scanned by stepping in the superior-inferior direction in three overlapping positions to achieve a total 465 × 350 × 400 mm(3) sampled FOV yielding >13 800 landmarks (3D Gradient-Echo, TE/TR/α = 5.54 ms/30 ms/28°, voxel size = 1 × 1 × 2 mm(3)). A binary template (reference) was generated from a phantom schematic. An automated program converted MR images to binary via masking, thresholding, and testing for connectivity to identify landmarks. Distortion maps were generated by centroid mapping. Images were corrected via warping with inverse distortion maps, and temporal stability was assessed.

RESULTS

Over the sampled FOV, non-negligible residual gradient distortions existed as close as 9.5 cm from isocenter, with a maximum distortion of 7.4 mm as close as 23 cm from isocenter. Over six months, average gradient distortions were -0.07 ± 1.10 mm and 0.10 ± 1.10 mm in the x and y directions for the transverse plane, 0.03 ± 0.64 and -0.09 ± 0.70 mm in the sagittal plane, and 0.4 ± 1.16 and 0.04 ± 0.40 mm in the coronal plane. After implementing 3D correction maps, distortions were reduced to <1 pixel width (1 mm) for all voxels up to 25 cm from magnet isocenter.

CONCLUSIONS

Inherent distortion due to gradient nonlinearity was found to be non-negligible even with vendor corrections applied, and further corrections are required to obtain 1 mm accuracy for large FOVs. Statistical analysis of temporal stability shows that sequence independent distortion maps are consistent within six months of characterization.

Spatial Precision in Magnetic Resonance Imaging-Guided Radiation Therapy: The Role of Geometric Distortion

Because magnetic resonance imaging-guided radiation therapy (MRIgRT) offers exquisite soft tissue contrast and the ability to image tissues in arbitrary planes, the interest in this technology has increased dramatically in recent years. However, intrinsic geometric distortion stemming from both the system hardware and the magnetic properties of the patient affects MR images and compromises the spatial integrity of MRI-based radiation treatment planning, given that for real-time MRIgRT, precision within 2 mm is desired. In this article, we discuss the causes of geometric distortion, describe some well-known distortion correction algorithms, and review geometric distortion measurements from 12 studies, while taking into account relevant imaging parameters. Eleven of the studies reported phantom measurements quantifying system-dependent geometric distortion, while 2 studies reported simulation data quantifying magnetic susceptibility-induced geometric distortion. Of the 11 studies investigating system-dependent geometric distortion, 5 reported maximum measurements less than 2 mm. The simulation studies demonstrated that magnetic susceptibility-induced distortion is typically smaller than system-dependent distortion but still nonnegligible, with maximum distortion ranging from 2.1 to 2.6 mm at a field strength of 1.5 T. As expected, anatomic landmarks containing interfaces between air and soft tissue had the largest distortions. The evidence indicates that geometric distortion reduces the spatial integrity of MRI-based radiation treatment planning and likely diminishes the efficacy of MRIgRT. Better phantom measurement techniques and more effective distortion correction algorithms are needed to achieve the desired spatial precision.

Continuous table acquisition MRI for radiotherapy treatment planning: distortion assessment with a new extended 3D volumetric phantom

PURPOSE

Accurate geometry is required for radiotherapy treatment planning (RTP). When considering the use of magnetic resonance imaging (MRI) for RTP, geometric distortions observed in the acquired images should be considered. While scanner technology and vendor supplied correction algorithms provide some correction, large distortions are still present in images, even when considering considerably smaller scan lengths than those typically acquired with CT in conventional RTP. This study investigates MRI acquisition with a moving table compared with static scans for potential geometric benefits for RTP.

METHODS

A full field of view (FOV) phantom (diameter 500 mm; length 513 mm) was developed for measuring geometric distortions in MR images over volumes pertinent to RTP. The phantom consisted of layers of refined plastic within which vitamin E capsules were inserted. The phantom was scanned on CT to provide the geometric gold standard and on MRI, with differences in capsule location determining the distortion. MRI images were acquired with two techniques. For the first method, standard static table acquisitions were considered. Both 2D and 3D acquisition techniques were investigated. With the second technique, images were acquired with a moving table. The same sequence was acquired with a static table and then with table speeds of 1.1 mm/s and 2 mm/s. All of the MR images acquired were registered to the CT dataset using a deformable B-spline registration with the resulting deformation fields providing the distortion information for each acquisition.

RESULTS

MR images acquired with the moving table enabled imaging of the whole phantom length while images acquired with a static table were only able to image 50%-70% of the phantom length of 513 mm. Maximum distortion values were reduced across a larger volume when imaging with a moving table. Increased table speed resulted in a larger contribution of distortion from gradient nonlinearities in the through-plane direction and an increased blurring of capsule images, resulting in an apparent capsule volume increase by up to 170% in extreme axial FOV regions. Blurring increased with table speed and in the central regions of the phantom, geometric distortion was less for static table acquisitions compared to a table speed of 2 mm/s over the same volume. Overall, the best geometric accuracy was achieved with a table speed of 1.1 mm/s.

CONCLUSIONS

The phantom designed enables full FOV imaging for distortion assessment for the purposes of RTP. MRI acquisition with a moving table extends the imaging volume in the z direction with reduced distortions which could be useful particularly if considering MR-only planning. If utilizing MR images to provide additional soft tissue information to the planning CT, standard acquisition sequences over a smaller volume would avoid introducing additional blurring or distortions from the through-plane table movement.