Abstract:The authors showed that a patch-based method based on affine registrations and T1-weighted MRI could generate accurate pCTs of the pelvis. The main source of differences between pCT and CT was positional changes of air pockets and body outline.
“…For both the target and the OARs, both sCT-based dose distributions differed from the corresponding CT-based dose distribution, on average, no more than 1% of the original dose. These results are comparable to those already presented in the literature [8], [9], [11], [19], [23], [24]. Mean percentages of passing points within the external contour of 98–100% and 94–97% were achieved for both methods and cancer sites for the 3D local Ɣ-analyses with constraints of 3%_3 mm and 2%_2 mm, respectively.…”
Section: Discussionsupporting
confidence: 88%
“…Mean percentages of passing points within the external contour of 98–100% and 94–97% were achieved for both methods and cancer sites for the 3D local Ɣ-analyses with constraints of 3%_3 mm and 2%_2 mm, respectively. Ɣ-pass rates of the same order were reported by Korhonen et al [11] (93–97% for 1%_1 mm 2D Ɣ-test), Jonsson et al [15] (99% for 3%_3 mm Ɣ-test), Dowling et al [19] (93–96% for 2%_2 mm 3D global Ɣ-test), Uh et al [23] (98–99% for 2%_2 mm Ɣ-test), Andreasen et al [24] (97% for 1%_1 mm 2D global Ɣ-test) and Siversson et al [25] (99–100% for 2%_1 mm local Ɣ-test). Furthermore, dose distributions based on sCT a showed a better PTV agreement and a more homogeneous gamma map with lower gamma values than sCT bda .…”
Section: Discussionsupporting
confidence: 75%
“…The MAE obtained within the external contour for the prostate patients was on average 49.8 ± 4.6 HU, which is lower than the error obtained by Kim et al [20] (74.3 ± 10.9 HU) and is of the same order as the MAE obtained by Siversson et al [25], Dowling et al [19] and Andreasen et al [24] (36.5 ± 4.1 HU, 40.5 ± 8.2 HU and 54 ± 8 HU, respectively), when taking into account the fact that the images used in the present study had a lower resolution. The synthesis error was higher for the H&N patients as the neck is a more challenging area for registration algorithms because of the mixture of bone and air, and due to the presence of large-scale postural changes between patients, such as flexion or extension of the neck and the position of the jawbone.…”
Section: Discussionmentioning
confidence: 38%
“…The fusion can be obtained by computing the voxelwise median [21], using a probabilistic Bayesian framework [22], an arithmetic mean process or pattern recognition with Gaussian process [23] or a local image similarity measure [18], [19]. Instead of using a database of images, Andreasen et al [24] employed a dictionary of MR and CT patches. The sCT was predicted by extracting patches from the test subject MRI, running an intensity-based nearest neighbour search in the patch database and fusing the selected CT patches using a similarity-weighted average.…”
HighlightsEstablishing MRI-only RTP workflows requires synthetic CTs for dose calculation.This study evaluates the feasibility of using a multi-atlas CT synthesis approach.The proposed method was validated on head and neck and prostate cancer patients.Results showed an accurate bone estimation for future patient positioning.Results showed that synthetic CTs are suitable to perform clinical dose calculations.
“…For both the target and the OARs, both sCT-based dose distributions differed from the corresponding CT-based dose distribution, on average, no more than 1% of the original dose. These results are comparable to those already presented in the literature [8], [9], [11], [19], [23], [24]. Mean percentages of passing points within the external contour of 98–100% and 94–97% were achieved for both methods and cancer sites for the 3D local Ɣ-analyses with constraints of 3%_3 mm and 2%_2 mm, respectively.…”
Section: Discussionsupporting
confidence: 88%
“…Mean percentages of passing points within the external contour of 98–100% and 94–97% were achieved for both methods and cancer sites for the 3D local Ɣ-analyses with constraints of 3%_3 mm and 2%_2 mm, respectively. Ɣ-pass rates of the same order were reported by Korhonen et al [11] (93–97% for 1%_1 mm 2D Ɣ-test), Jonsson et al [15] (99% for 3%_3 mm Ɣ-test), Dowling et al [19] (93–96% for 2%_2 mm 3D global Ɣ-test), Uh et al [23] (98–99% for 2%_2 mm Ɣ-test), Andreasen et al [24] (97% for 1%_1 mm 2D global Ɣ-test) and Siversson et al [25] (99–100% for 2%_1 mm local Ɣ-test). Furthermore, dose distributions based on sCT a showed a better PTV agreement and a more homogeneous gamma map with lower gamma values than sCT bda .…”
Section: Discussionsupporting
confidence: 75%
“…The MAE obtained within the external contour for the prostate patients was on average 49.8 ± 4.6 HU, which is lower than the error obtained by Kim et al [20] (74.3 ± 10.9 HU) and is of the same order as the MAE obtained by Siversson et al [25], Dowling et al [19] and Andreasen et al [24] (36.5 ± 4.1 HU, 40.5 ± 8.2 HU and 54 ± 8 HU, respectively), when taking into account the fact that the images used in the present study had a lower resolution. The synthesis error was higher for the H&N patients as the neck is a more challenging area for registration algorithms because of the mixture of bone and air, and due to the presence of large-scale postural changes between patients, such as flexion or extension of the neck and the position of the jawbone.…”
Section: Discussionmentioning
confidence: 38%
“…The fusion can be obtained by computing the voxelwise median [21], using a probabilistic Bayesian framework [22], an arithmetic mean process or pattern recognition with Gaussian process [23] or a local image similarity measure [18], [19]. Instead of using a database of images, Andreasen et al [24] employed a dictionary of MR and CT patches. The sCT was predicted by extracting patches from the test subject MRI, running an intensity-based nearest neighbour search in the patch database and fusing the selected CT patches using a similarity-weighted average.…”
HighlightsEstablishing MRI-only RTP workflows requires synthetic CTs for dose calculation.This study evaluates the feasibility of using a multi-atlas CT synthesis approach.The proposed method was validated on head and neck and prostate cancer patients.Results showed an accurate bone estimation for future patient positioning.Results showed that synthetic CTs are suitable to perform clinical dose calculations.
“…This process is computationally intense with times ranging from ten minutes to several hours to generate sCT images. [8][9][10][11][12] Note also that computational time increases with the number of atlas pairs used. Ultimately, this computational burden makes these methods largely unsuitable for clinical practice at present, especially looking forward to where MR is used for online treatment adaptation.…”
Purpose
To evaluate pix2pix and CycleGAN and to assess the effects of multiple combination strategies on accuracy for patch‐based synthetic computed tomography (sCT) generation for magnetic resonance (MR)‐only treatment planning in head and neck (HN) cancer patients.
Materials and methods
Twenty‐three deformably registered pairs of CT and mDixon FFE MR datasets from HN cancer patients treated at our institution were retrospectively analyzed to evaluate patch‐based sCT accuracy via the pix2pix and CycleGAN models. To test effects of overlapping sCT patches on estimations, we (a) trained the models for three orthogonal views to observe the effects of spatial context, (b) we increased effective set size by using per‐epoch data augmentation, and (c) we evaluated the performance of three different approaches for combining overlapping Hounsfield unit (HU) estimations for varied patch overlap parameters. Twelve of twenty‐three cases corresponded to a curated dataset previously used for atlas‐based sCT generation and were used for training with leave‐two‐out cross‐validation. Eight cases were used for independent testing and included previously unseen image features such as fused vertebrae, a small protruding bone, and tumors large enough to deform normal body contours. We analyzed the impact of MR image preprocessing including histogram standardization and intensity clipping on sCT generation accuracy. Effects of mDixon contrast (in‐phase vs water) differences were tested with three additional cases. The sCT generation accuracy was evaluated using mean absolute error (MAE) and mean error (ME) in HU between the plan CT and sCT images. Dosimetric accuracy was evaluated for all clinically relevant structures in the independent testing set and digitally reconstructed radiographs (DRRs) were evaluated with respect to the plan CT images.
Results
The cross‐validated MAEs for the whole‐HN region using pix2pix and CycleGAN were 66.9 ± 7.3 vs 82.3 ± 6.4 HU, respectively. On the independent testing set with additional artifacts and previously unseen image features, whole‐HN region MAEs were 94.0 ± 10.6 and 102.9 ± 14.7 HU for pix2pix and CycleGAN, respectively. For patients with different tissue contrast (water mDixon MR images), the MAEs increased to 122.1 ± 6.3 and 132.8 ± 5.5 HU for pix2pix and CycleGAN, respectively. Our results suggest that combining overlapping sCT estimations at each voxel reduced both MAE and ME compared to single‐view non‐overlapping patch results. Absolute percent mean/max dose errors were 2% or less for the PTV and all clinically relevant structures in our independent testing set, including structures with image artifacts. Quantitative DRR comparison between planning CTs and sCTs showed agreement of bony region positions to <1 mm.
Conclusions
The dosimetric and MAE based accuracy, along with the similarity between DRRs from sCTs, indicate that pix2pix and CycleGAN are promising methods for MR‐only treatment planning for HN cancer. Our methods investigated for overlapping patch‐based HU estimations also ...
Purpose
The purpose of this study was to address the dosimetric accuracy of synthetic computed tomography (sCT) images of patients with brain tumor generated using a modified generative adversarial network (GAN) method, for their use in magnetic resonance imaging (MRI)‐only treatment planning for proton therapy.
Methods
Dose volume histogram (DVH) analysis was performed on CT and sCT images of patients with brain tumor for plans generated for intensity‐modulated proton therapy (IMPT). All plans were robustly optimized using a commercially available treatment planning system (RayStation, from RaySearch Laboratories) and standard robust parameters reported in the literature. The IMPT plan was then used to compute the dose on CT and sCT images for dosimetric comparison, using RayStation analytical (pencil beam) dose algorithm. We used a second, independent Monte Carlo dose calculation engine to recompute the dose on both CT and sCT images to ensure a proper analysis of the dosimetric accuracy of the sCT images.
Results
The results extracted from RayStation showed excellent agreement for most DVH metrics computed on the CT and sCT for the nominal case, with a mean absolute difference below 0.5% (0.3 Gy) of the prescription dose for the clinical target volume (CTV) and below 2% (1.2 Gy) for the organs at risk (OARs) considered. This demonstrates a high dosimetric accuracy for the generated sCT images, especially in the target volume. The metrics obtained from the Monte Carlo doses mostly agreed with the values extracted from RayStation for the nominal and worst‐case scenarios (mean difference below 3%).
Conclusions
This work demonstrated the feasibility of using sCT generated with a GAN‐based deep learning method for MRI‐only treatment planning of patients with brain tumor in intensity‐modulated proton therapy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.