Abstract:The purpose of this work is to develop a learning-based method to generate synthetic CT from routine anatomical MRI for MRI-only radiation treatment of prostate cancer. We propose to integrate auto-context model and patch-based anatomical signature into machine learning framework to effectively capture the relationship between the CT and MR images for CT synthesis. The proposed prediction of CT images consists of two major stages: the training stage and the prediction stage. During the training stage, patch-based anatomical features are extracted from the aligned MRI-CT training images, and the most informative features are identified to train a random forest based on auto-context model. During the prediction stage, we extract the selected features from the new MRI and feed them into the well-trained forests for the iterative CT prediction. Our predicted CT images generated from MRI were registered to original (planning) CT images for generating a synthetic CT-based treatment plan. Mean absolute error (MAE) and normalized cross-correlation (NCC) were used to quantify the differences between the synthetic and original CT images. Clinically relevant dose volume histogram (DVH) metrics were extracted from original CT-based and synthetic CT-based treatment plans for quantitative dosimetric evaluation. Gamma analysis was performed for the comparison of absorbed dose distributions between the both plans of each patient. This CT prediction method was tested using 22 prostate patients with 34 treatment plans (mean prescription dose: 48.79±15.20 Gy), each of whom has planning CT and MRI images. The mean MAE and NCC between MRI-based synthetic CT and planning CT were 29.86±10.04HU and 0.98±0.03 for all patients' data. Mean dose differences between PTVs on synthetic and planning CT images across all the plans were -0.23±0.36%, -0.12±0.14%, -0.13±0.13%, -0.14±0.14%, -0.13±0.14%, and -0.15±0.24% for Dmin, D10, D50, D95, Dmean and Dmax, respectively. Mean dose differences between bladder, rectum, femur head and penile bulb on both CT images ranged from -0.25% to 0.20% for all DVH metrics. No significant differences were observed in the DVH metrics of the PTV and OARs between synthetic and planning CT-based plans (p > 0.05). The average pass rate of gamma analysis was over 99% with 3%/3 mm (dose difference/distance to agreement) acceptance criteria in every plan. We have investigated a novel learning-based approach to generate CT images from routine MR images for MRI-based prostate cancer radiotherapy. We have demonstrated that the reported method is capable of reliably generating synthetic CT images from the MR images and providing dose calculation of comparable accuracy to the standard CT for prostate cancer treatment planning. The image similarity and dosimetric agreement between synthetic and planning CT warrants further development of a MRI-only workflow for radiotherapy of prostate cancer.

MRI to CT Synthesis Using Contrastive Learning

A Novel Method Of Synthetic Ct Generation From Mr Images Based On Convolutional Neural Networks

Mutual information-based contrastive learning for the generation of pseudo-CT images of the head from magnetic resonance imaging

Synthetic CT generation from cone-beam CT using deep-learning for breast adaptive radiotherapy

MRI-based Synthetic CT Generation Using Deep Convolutional Neural Network

Magnetic Resonance-Based Synthetic Computed Tomography Using Generative Adversarial Networks for Intracranial Tumor Radiotherapy Treatment Planning

Development of an Unsupervised Cycle Contrastive Unpaired Translation Network for MRI‐to‐CT Synthesis

MRI-only brain radiotherapy: Assessing the dosimetric accuracy of synthetic CT images generated using a deep learning approach

Evaluation of Synthetic CT Image Generated Using a Neural Network for MR-Only Radiotherapy

CT synthesis from MRI with an improved multi-scale learning network

MRI-based Pseudo CT Synthesis Using Anatomical Signature and Alternating Random Forest with Iterative Refinement Model

MRI-only Based Synthetic CT Generation Using Dense Cycle Consistent Generative Adversarial Networks.

MRI-based Treatment Planning for Liver Stereotactic Body Radiotherapy: Validation of a Deep Learning-Based Synthetic CT Generation Method

MRI-Based Synthetic CT for Radiation Treatment of Prostate Cancer

Generation of Synthetic CT Images From MRI for Treatment Planning and Patient Positioning Using a 3-Channel U-Net Trained on Sagittal Images

Multi-planar dual adversarial network based on dynamic 3D features for MRI-CT head and neck image synthesis

Magnetic Resonance Imaging-Based Pseudo Computed Tomography Using Anatomic Signature and Joint Dictionary Learning

CBCT-Based Synthetic MRI Generation for CBCT-Guided Adaptive Radiotherapy.

Abdominal synthetic CT generation from MR Dixon images using a U-net trained with ‘semi-synthetic’ CT data

MRI-based Synthetic CT Generation Using Semantic Random Forest with Iterative Refinement.

Deep learning-based synthetic CT generation from MR images: comparison of generative adversarial and residual neural networks