Computed tomography synthesis from magnetic resonance imaging using cycle Generative Adversarial Networks with multicenter learning.

Background and Purpose: Addressing the need for accurate dose calculation in MRI-only radiotherapy, the generation of synthetic Computed Tomography (sCT) from MRI has emerged. Deep learning (DL) techniques, have shown promising results in achieving high sCT accuracies. However, existing sCT synthesis methods are often center-specific, posing a challenge to their generalizability. To overcome this limitation, recent studies have proposed approaches, such as multicenter training . Material and methods: The purpose of this work was to propose a multicenter sCT synthesis by DL, using a 2D cycle-GAN on 128 prostate cancer patients, from four different centers. Four cases were compared: monocenter cases, monocenter training and test on another center, multicenter trainings and a test on a center not included in the training and multicenter trainings with an included center in the test. Trainings were performed using 20 patients. sCT accuracy evaluation was performed using Mean Absolute Error, Mean Error and Peak-Signal-to-Noise-Ratio. Dose accuracy was assessed with gamma index and Dose Volume Histogram comparison. Results: Qualitative, quantitative and dose results show that the accuracy of sCTs for monocenter trainings and multicenter trainings using a seen center in the test did not differ significantly. However, when the test involved an unseen center, the sCT quality was inferior. Conclusions: The aim of this work was to propose generalizable multicenter training for MR-to-CT synthesis. It was shown that only a few data from one center included in the training cohort allows sCT accuracy equivalent to a monocenter study.

A deep learning model to generate synthetic CT for prostate MR-only radiotherapy dose planning: a multicenter study.

Introduction: For radiotherapy based solely on magnetic resonance imaging (MRI), generating synthetic computed tomography scans (sCT) from MRI is essential for dose calculation. The use of deep learning (DL) methods to generate sCT from MRI has shown encouraging results if the MRI images used for training the deep learning network and the MRI images for sCT generation come from the same MRI device. The objective of this study was to create and evaluate a generic DL model capable of generating sCTs from various MRI devices for prostate radiotherapy. Materials and methods: In total, 90 patients from three centers (30 CT-MR prostate pairs/center) underwent treatment using volumetric modulated arc therapy for prostate cancer (PCa) (60 Gy in 20 fractions). T2 MRI images were acquired in addition to computed tomography (CT) images for treatment planning. The DL model was a 2D supervised conditional generative adversarial network (Pix2Pix). Patient images underwent preprocessing steps, including nonrigid registration. Seven different supervised models were trained, incorporating patients from one, two, or three centers. Each model was trained on 24 CT-MR prostate pairs. A generic model was trained using patients from all three centers. To compare sCT and CT, the mean absolute error in Hounsfield units was calculated for the entire pelvis, prostate, bladder, rectum, and bones. For dose analysis, mean dose differences of D 99% for CTV, V 95% for PTV, Dmax for rectum and bladder, and 3D gamma analysis (local, 1%/1 mm) were calculated from CT and sCT. Furthermore, Wilcoxon tests were performed to compare the image and dose results obtained with the generic model to those with the other trained models. Results: Considering the image results for the entire pelvis, when the data used for the test comes from the same center as the data used for training, the results were not significantly different from the generic model. Absolute dose differences were less than 1 Gy for the CTV D 99% for every trained model and center. The gamma analysis results showed nonsignificant differences between the generic and monocentric models. Conclusion: The accuracy of sCT, in terms of image and dose, is equivalent to whether MRI images are generated using the generic model or the monocentric model. The generic model, using only eight MRI-CT pairs per center, offers robust sCT generation, facilitating PCa MRI-only radiotherapy for routine clinical use.

Determination of acceptable Hounsfield units uncertainties via a sensitivity analysis for an accurate dose calculation in the context of prostate MRI-only radiotherapy.

Radiation therapy is moving from CT based to MRI guided planning, particularly for soft tissue anatomy. An important requirement of this new workflow is the generation of synthetic-CT (sCT) from MRI to enable treatment dose calculations. Automatic methods to determine the acceptable range of CT Hounsfield Unit (HU) uncertainties to avoid dose distribution errors is thus a key step toward safe MRI-only radiotherapy. This work has analysed the effects of controlled errors introduced in CT scans on the delivered radiation dose for prostate cancer patients. Spearman correlation coefficient has been computed, and a global sensitivity analysis performed following the Morris screening method. This allows the classification of different error factors according to their impact on the dose at the isocentre. sCT HU estimation errors in the bladder appeared to be the least influential factor, and sCT quality assessment should not only focus on organs surrounding the radiation target, as errors in other soft tissue may significantly impact the dose in the target volume. This methodology links dose and intensity-based metrics, and is the first step to define a threshold of acceptability of HU uncertainties for accurate dose planning.

Quality assurance for MRI-only radiation therapy: A voxel-wise population-based methodology for image and dose assessment of synthetic CT generation methods.

The quality assurance of synthetic CT (sCT) is crucial for safe clinical transfer to an MRI-only radiotherapy planning workflow. The aim of this work is to propose a population-based process assessing local errors in the generation of sCTs and their impact on dose distribution. For the analysis to be anatomically meaningful, a customized interpatient registration method brought the population data to the same coordinate system. Then, the voxel-based process was applied on two sCT generation methods: a bulk-density method and a generative adversarial network. The CT and MRI pairs of 39 patients treated by radiotherapy for prostate cancer were used for sCT generation, and 26 of them with delineated structures were selected for analysis. Voxel-wise errors in sCT compared to CT were assessed for image intensities and dose calculation, and a population-based statistical test was applied to identify the regions where discrepancies were significant. The cumulative histograms of the mean absolute dose error per volume of tissue were computed to give a quantitative indication of the error for each generation method. Accurate interpatient registration was achieved, with mean Dice scores higher than 0.91 for all organs. The proposed method produces three-dimensional maps that precisely show the location of the major discrepancies for both sCT generation methods, highlighting the heterogeneity of image and dose errors for sCT generation methods from MRI across the pelvic anatomy. Hence, this method provides additional information that will assist with both sCT development and quality control for MRI-based planning radiotherapy.

Deep learning methods to generate synthetic CT from MRI in radiotherapy: A literature review.

PURPOSE: In radiotherapy, MRI is used for target volume and organs-at-risk delineation for its superior soft-tissue contrast as compared to CT imaging. However, MRI does not provide the electron density of tissue necessary for dose calculation. Several methods of synthetic-CT (sCT) generation from MRI data have been developed for radiotherapy dose calculation. This work reviewed deep learning (DL) sCT generation methods and their associated image and dose evaluation, in the context of MRI-based dose calculation. METHODS: We searched the PubMed and ScienceDirect electronic databases from January 2010 to March 2021. For each paper, several items were screened and compiled in figures and tables. RESULTS: This review included 57 studies. The DL methods were either generator-only based (45% of the reviewed studies), or generative adversarial network (GAN) architecture and its variants (55% of the reviewed studies). The brain and pelvis were the most commonly investigated anatomical localizations (39% and 28% of the reviewed studies, respectively), and more rarely, the head-and-neck (H&N) (15%), abdomen (10%), liver (5%) or breast (3%). All the studies performed an image evaluation of sCTs with a diversity of metrics, with only 36 studies performing dosimetric evaluations of sCT. CONCLUSIONS: The median mean absolute errors were around 76 HU for the brain and H&N sCTs and 40 HU for the pelvis sCTs. For the brain, the mean dose difference between the sCT and the reference CT was <2%. For the H&N and pelvis, the mean dose difference was below 1% in most of the studies. Recent GAN architectures have advantages compared to generator-only, but no superiority was found in term of image or dose sCT uncertainties. Key challenges of DL-based sCT generation methods from MRI in radiotherapy is the management of movement for abdominal and thoracic localizations, the standardization of sCT evaluation, and the investigation of multicenter impacts.

Comparison of CBCT-based dose calculation methods in head and neck cancer radiotherapy: from Hounsfield unit to density calibration curve to deep learning.

PURPOSE: Anatomical variations occur during head and neck (H&N) radiotherapy treatment. kV cone-beam computed tomography (CBCT) images can be used for daily dose monitoring to assess dose variations owing to anatomic changes. Deep learning methods (DLMs) have recently been proposed to generate pseudo-CT (pCT) from CBCT to perform dose calculation. This study aims to evaluate the accuracy of a DLM and to compare this method with three existing methods of dose calculation from CBCT in H&N cancer radiotherapy. METHODS: Forty-four patients received VMAT for H&N cancer (70-63-56 Gy). For each patient, reference CT (Bigbore, Philips) and CBCT images (XVI, Elekta) were acquired. The DLM was based on a generative adversarial network. The three compared methods were: (a) a method using a density to Hounsfield Unit (HU) relation from phantom CBCT image (HU-D curve method), (b) a water-air-bone density assignment method (DAM), and iii) a method using deformable image registration (DIR). The imaging endpoints were the mean absolute error (MAE) and mean error (ME) of HU from pCT and reference CT (CTref ). The dosimetric endpoints were dose discrepancies and 3D gamma analyses (local, 2%/2 mm, 30% dose threshold). Dose discrepancies were defined as the mean absolute differences between DVHs calculated from the CTref and pCT of each method. RESULTS: In the entire body, the MAEs and MEs of the DLM, HU-D curve method, DAM, and DIR method were 82.4 and 17.1 HU, 266.6 and 208.9 HU, 113.2 and 14.2 HU, and 95.5 and -36.6 HU, respectively. The MAE obtained using the DLM differed significantly from those of other methods (Wilcoxon, P ≤ 0.05). The DLM dose discrepancies were 7 ± 8 cGy (maximum = 44 cGy) for the ipsilateral parotid gland Dmean and 5 ± 6 cGy (max = 26 cGy) for the contralateral parotid gland mean dose (Dmean ). For the parotid gland Dmean , no significant dose difference was observed between the DLM and other methods. The mean 3D gamma pass rate ± standard deviation was 98.1 ± 1.2%, 91.0 ± 5.3%, 97.9 ± 1.6%, and 98.8 ± 0.7% for the DLM, HU-D method, DAM, and DIR method, respectively. The gamma pass rates and mean gamma results of the HU-D curve method, DAM, and DIR method differed significantly from those of the DLM. CONCLUSIONS: For H&N radiotherapy, DIR method and DLM appears as the most appealing CBCT-based dose calculation methods among the four methods in terms of dose accuracy as well as calculation time. Using the DIR method or DLM with CBCT images enables dose monitoring in the parotid glands during the treatment course and may be used to trigger replanning.

Comparison of Deep Learning-Based and Patch-Based Methods for Pseudo-CT Generation in MRI-Based Prostate Dose Planning.

PURPOSE: Deep learning methods (DLMs) have recently been proposed to generate pseudo-CT (pCT) for magnetic resonance imaging (MRI) based dose planning. This study aims to evaluate and compare DLMs (U-Net and generative adversarial network [GAN]) using various loss functions (L2, single-scale perceptual loss [PL], multiscale PL, weighted multiscale PL) and a patch-based method (PBM). METHODS AND MATERIALS: Thirty-nine patients received a volumetric modulated arc therapy for prostate cancer (78 Gy). T2-weighted MRIs were acquired in addition to planning CTs. The pCTs were generated from the MRIs using 7 configurations: 4 GANs (L2, single-scale PL, multiscale PL, weighted multiscale PL), 2 U-Net (L2 and single-scale PL), and the PBM. The imaging endpoints were mean absolute error and mean error, in Hounsfield units, between the reference CT (CTref) and the pCT. Dose uncertainties were quantified as mean absolute differences between the dose volume histograms (DVHs) calculated from the CTref and pCT obtained by each method. Three-dimensional gamma indexes were analyzed. RESULTS: Considering the image uncertainties in the whole pelvis, GAN L2 and U-Net L2 showed the lowest mean absolute error (≤34.4 Hounsfield units). The mean errors were not different than 0 (P ≤ .05). The PBM provided the highest uncertainties. Very few DVH points differed when comparing GAN L2 or U-Net L2 DVHs and CTref DVHs (P ≤ .05). Their dose uncertainties were ≤0.6% for the prostate planning target Volume V95%, ≤0.5% for the rectum V70Gy, and ≤0.1% for the bladder V50Gy. The PBM, U-Net PL, and GAN PL presented the highest systematic dose uncertainties. The gamma pass rates were >99% for all DLMs. The mean calculation time to generate 1 pCT was 15 s for the DLMs and 62 min for the PBM. CONCLUSIONS: Generating pCT for MRI dose planning with DLMs and PBM provided low-dose uncertainties. In particular, the GAN L2 and U-Net L2 provided the lowest dose uncertainties together with a low computation time.

Pseudo-CT Generation for MRI-Only Radiation Therapy Treatment Planning: Comparison Among Patch-Based, Atlas-Based, and Bulk Density Methods.

PURPOSE: Methods have been recently developed to generate pseudo-computed tomography (pCT) for dose calculation in magnetic resonance imaging (MRI)-only radiation therapy. This study aimed to propose an original nonlocal mean patch-based method (PBM) and to compare this PBM to an atlas-based method (ABM) and to a bulk density method (BDM) for prostate MRI-only radiation therapy. MATERIALS AND METHODS: Thirty-nine patients received a volumetric modulated arc therapy for prostate cancer. In addition to the planning computed tomography (CT) scans, T2-weighted MRI scans were acquired. pCTs were generated from MRIs using 3 methods: an original nonlocal mean PBM, ABM, and BDM. The PBM was performed using feature extraction and approximate nearest neighbor search in a training cohort. The PBM accuracy was evaluated in a validation cohort by using imaging and dosimetric endpoints. Imaging endpoints included mean absolute error and mean error between Hounsfield units of the pCT and the reference CT (CTref). Dosimetric endpoints were based on dose-volume histograms calculated from the CTref and the pCTs for various volumes of interest and on 3-dimensional gamma analyses. The PBM uncertainties were compared with those of the ABM and BDM. RESULTS: The mean absolute error and mean error obtained from the PBM were 41.1 and -1.1 Hounsfield units. The PBM dose-volume histogram differences were 0.7% for prostate planning target volume V95%, 0.5% for rectum V70Gy, and 0.2% for bladder V50Gy. Compared with ABM and BDM, PBM provided significantly lower dose uncertainties for the prostate planning target volume (70-78 Gy), the rectum (8.5-29 Gy, 40-48 Gy, and 61-73 Gy), and the bladder (12-78 Gy). The PBM mean gamma pass rate (99.5%) was significantly higher than that of ABM (94.9%) or BDM (96.1%). CONCLUSIONS: The proposed PBM provides low uncertainties with dose planned on CTref. These uncertainties were smaller than those of ABM and BDM and are unlikely to be clinically significant.

Towards a patient-specific hepatic arterial modeling for microspheres distribution optimization in SIRT protocol.

Selective internal radiation therapy (SIRT) using Yttrium-90 loaded glass microspheres injected in the hepatic artery is an emerging, minimally invasive therapy of liver cancer. A personalized intervention can lead to high concentration dose in the tumor, while sparing the surrounding parenchyma. We propose a computational model for patient-specific simulation of entire hepatic arterial tree, based on liver, tumors, and arteries segmentation on patient's tomography. Segmentation of hepatic arteries down to a diameter of 0.5 mm is semi-automatically performed on 3D cone-beam CT angiography. The liver and tumors are extracted from CT-scan at portal phase by an active surface method. Once the images are registered through an automatic multimodal registration, extracted data are used to initialize a numerical model simulating liver vascular network. The model creates successive bifurcations from given principal vessels, observing Poiseuille's and matter conservation laws. Simulations provide a coherent reconstruction of global hepatic arterial tree until vessel diameter of 0.05 mm. Microspheres distribution under simple hypotheses is also quantified, depending on injection site. The patient-specific character of this model may allow a personalized numerical approximation of microspheres final distribution, opening the way to clinical optimization of catheter placement for tumor targeting.

Inexact coronary tree matching algorithm with artificial nodes.

Coronary tree matching is applied to plan percutaneous vascular procedures. This work, which allows following each segment of non-isomorphic coronary trees over time, precedes the determination of the best 2D angiography view from C-arm acquisition system for angioplasty procedure. To match two 3D coronary trees which represent two successive cardiac phases, we adapted a reference inexact tree matching algorithm based on association graph and maximum clique. To improve the pair-wise matching performance of our approach, artificial nodes are introduced to take into account the topology variation between 3D vascular trees. Different similarity measures using tree characteristics and geometric features of coronary branches are evaluated and compared to our previous work.

Turning Tangent Empirical Mode Decomposition: A Framework for Mono- and Multivariate Signals.

A novel Empirical Mode Decomposition (EMD) algorithm, called 2T-EMD, for both mono- and multivariate signals is proposed in this paper. It differs from the other approaches by its computational lightness and its algorithmic simplicity. The method is essentially based on a redefinition of the signal mean envelope, computed thanks to new characteristic points, which offers the possibility to decompose multivariate signals without any projection. The scope of application of the novel algorithm is specified, and a comparison of the 2T-EMD technique with classical methods is performed on various simulated mono- and multivariate signals. The monovariate behaviour of the proposed method on noisy signals is then validated by decomposing a fractional Gaussian noise and an application to real life EEG data is finally presented.

L0 constrained sparse reconstruction for multi-slice helical CT reconstruction.

In this paper, we present a Bayesian maximum a posteriori method for multi-slice helical CT reconstruction based on an L0-norm prior. It makes use of a very low number of projections. A set of surrogate potential functions is used to successively approximate the L0-norm function while generating the prior and to accelerate the convergence speed. Simulation results show that the proposed method provides high quality reconstructions with highly sparse sampled noise-free projections. In the presence of noise, the reconstruction quality is still significantly better than the reconstructions obtained with L1-norm or L2-norm priors.

A model-based reconstruction method for 3-D rotational coronary angiography.

This paper presents a model-based reconstruction method of the coronary tree from a few number of projections in rotational angiography imaging. The reconstruction relies on projections acquired at a same cardiac phase and an energy function minimization that aims to lead the deformation of the 3D model to fit projection data whereas preserving coherence both in time and space. Some preliminary results are provided on simulated rotational angiograms.

ECG gated tomographic reconstruction for 3-D rotational coronary angiography.

A method is proposed for 3-D reconstruction of coronary from a limited number of projections in rotational angiography. A Bayesian maximum a posteriori (MAP) estimation is applied with a Poisson distributed projection to reconstruct the 3D coronary tree at a given instant of the cardiac cycle. Several regularizers are investigated L0-norm, L1 and L2 -norm in order to take into account the sparsity of the data. Evaluations are reported on simulated data obtained from a 3D dynamic sequence acquired on a 64-slice GE LightSpeed CT scan. A performance study is conducted to evaluate the quality of the reconstruction of the structures.

Robust rigid registration of retinal angiograms through optimization.

Retinal fundus photographs are employed as standard diagnostic tools in ophthalmology. Serial photographs of the flow of fluorescein and indocyanine green (ICG) dye are used to determine the areas of the retinal lesions. For objective measurements of features, the registration of the images is a necessity. In this paper, we employ optimization techniques for registration with the help of 2-parameter translational motion model of retinal angiograms, based on non-linear pre-processing (Wiener filtering and morphological gradient) and computation of the similarity criteria for the alignment of the two gradient images for any given rigid transformation. The optimization methods are effectively employed to minimize the similarity criterion. The presence of noise, the variations in the background and the temporal variation of the fluorescence level pose serious problems in obtaining a robust registration of the retinal images. Moreover, local search strategies are not robust in the case of ICG angiograms, even if one uses a multiresolution approach. The present work makes a systematic comparison of different optimization techniques, namely the minimization method derived from the optical flow formulation, the Nelder-Mead local search and the HCIAC ant colony metaheuristic, each optimizing a similarity criterion for the gradient images. The impact of the resolution and median filtering of gradient image is studied and the robustness of the approaches is tested through experimental studies, performed on macular fluorescein and ICG angiographies. Our proposed optimization techniques have shown interesting results especially for high resolution difficult registration problems. Moreover, this approach seems promising for affine (6-parameter motion model) or elastical registrations.

ECG compression by modelling the instantaneous module/phase of its DCT.

Recent developments in compression methods on the non-linear and non-stationary data, such as electrocardiograms (ECG), have received large attention by the time-frequency analysts. The technique presented in this paper is based on parametrical modeling the instantaneous module as well as the instantaneous phase, estimated directly from the Discrete Cosine Transform (DCT) of each ECG beat. The estimated parameters are then used to reconstruct each recorded beat. In order to evaluate the performance of our technique, data recorded from the MIT-BIH arrhythmia database are used.