Current and future developments of synthetic computed tomography generation for radiotherapy

A deep learning algorithm to generate synthetic computed tomography images for brain treatments from 0.35 T magnetic resonance imaging

Background and Purpose

The development of Magnetic Resonance Imaging (MRI)-only Radiotherapy (RT) represents a significant advancement in the field. This study introduces a Deep Learning (DL) algorithm designed to quickly generate synthetic CT (sCT) images from low-field MR images in the brain, an area not yet explored.

Methods

Fifty-six patients were divided into training (32), validation (8), and test (16) groups. A conditional Generative Adversarial Network (cGAN) was trained on pre-processed axial paired images. sCTs were validated using mean absolute error (MAE) and mean error (ME) calculated within the patient body. Intensity Modulated Radiation Therapy (IMRT) plans were optimised on simulation MRI and calculated considering sCT and original CT as electron density (ED) map. Dose distributions using sCT and CT were compared using global gamma analysis at different tolerance criteria (2 %/2mm and 3 %/3mm) and evaluating the difference in estimating different Dose Volume Histogram (DVH) parameters for target and organs at risk (OARs).

Results

The network generated sCTs of each single patient in less than two minutes (mean time = 103 ± 41 s). For test patients, the MAE was 62.1 ± 17.7 HU, and the ME was -7.3 ± 13.4 HU. Dose parameters on sCTs were within 0.5 Gy of those on original CTs. Gamma passing rates 2 %/2mm, and 3 %/3mm criteria were 99.5 %±0.5 %, and 99.7 %±0.3 %, respectively.

Conclusion

The proposed DL algorithm generates in less than 2 min accurate sCT images in the brain for online adaptive radiotherapy, potentially eliminating the need for CT simulation in MR-only workflows for brain treatments.

Toward Closing the Loop in Image-to-Image Conversion in Radiotherapy: A Quality Control Tool to Predict Synthetic Computed Tomography Hounsfield Unit Accuracy

In recent years, synthetic Computed Tomography (CT) images generated from Magnetic Resonance (MR) or Cone Beam Computed Tomography (CBCT) acquisitions have been shown to be comparable to real CT images in terms of dose computation for radiotherapy simulation. However, until now, there has been no independent strategy to assess the quality of each synthetic image in the absence of ground truth. In this work, we propose a Deep Learning (DL)-based framework to predict the accuracy of synthetic CT in terms of Mean Absolute Error (MAE) without the need for a ground truth (GT). The proposed algorithm generates a volumetric map as an output, informing clinicians of the predicted MAE slice-by-slice. A cascading multi-model architecture was used to deal with the complexity of the MAE prediction task. The workflow was trained and tested on two cohorts of head and neck cancer patients with different imaging modalities: 27 MR scans and 33 CBCT. The algorithm evaluation revealed an accurate HU prediction (a median absolute prediction deviation equal to 4 HU for CBCT-based synthetic CTs and 6 HU for MR-based synthetic CTs), with discrepancies that do not affect the clinical decisions made on the basis of the proposed estimation. The workflow exhibited no systematic error in MAE prediction. This work represents a proof of concept about the feasibility of synthetic CT evaluation in daily clinical practice, and it paves the way for future patient-specific quality assessment strategies.

RobMedNAS: searching robust neural network architectures for medical image synthesis

Investigating U-Net model robustness in medical image synthesis against adversarial perturbations, this study introduces RobMedNAS, a neural architecture search strategy for identifying resilient U-Net configurations. Through retrospective analysis of synthesized CT from MRI data, employing Dice coefficient and mean absolute error metrics across critical anatomical areas, the study evaluates traditional U-Net models and RobMedNAS-optimized models under adversarial attacks. Findings demonstrate RobMedNAS's efficacy in enhancing U-Net resilience without compromising on accuracy, proposing a novel pathway for robust medical image processing.

A treatment-site-specific evaluation of commercial synthetic computed tomography solutions for proton therapy

Background and purpose

Despite the superior dose conformity of proton therapy, the dose distribution is sensitive to daily anatomical changes, which can affect treatment accuracy. This study evaluated the dose recalculation accuracy of two synthetic computed tomography (sCT) generation algorithms in a commercial treatment planning system.

Materials and methods

The evaluation was conducted for head-and-neck, thorax-and-abdomen, and pelvis sites treated with proton therapy. Thirty patients with two cone-beam computed tomography (CBCT) scans each were selected. The sCT images were generated from CBCT scans using two algorithms, Corrected CBCT (corrCBCT) and Virtual CT (vCT). Dose recalculations were performed based on these images for comparison with "ground truth" deformed CTs.

Results

The choice of algorithm influenced dose recalculation accuracy, particularly in high dose regions. For head-and-neck cases, the corrCBCT method showed closer agreement with the "ground truth", while for thorax-and-abdomen and pelvis cases, the vCT algorithm yielded better results (mean percentage dose discrepancy of 0.6 %, 1.3 % and 0.5 % for the three sites, respectively, in the high dose region). Head-and-neck and pelvis cases exhibited excellent agreement in high dose regions (2 %/2 mm gamma passing rate >98 %), while thorax-and-abdomen cases exhibited the largest differences, suggesting caution in sCT algorithm usage for this site. Significant systematic differences were observed in the clinical target volume and organ-at-risk doses in head-and-neck and pelvis cases, highlighting the importance of using the correct algorithm.

Conclusions

This study provided treatment site-specific recommendations for sCT algorithm selection in proton therapy. The findings offered insights for proton beam centers implementing adaptive radiotherapy workflows.

A Deep Learning Approach for the Fast Generation of Synthetic Computed Tomography from Low-Dose Cone Beam Computed Tomography Images on a Linear Accelerator Equipped with Artificial Intelligence

Artificial Intelligence (AI) is revolutionising many aspects of radiotherapy (RT), opening scenarios that were unimaginable just a few years ago. The aim of this study is to propose a Deep Leaning (DL) approach able to quickly generate synthetic Computed Tomography (CT) images from low-dose Cone Beam CT (CBCT) acquired on a modern linear accelerator integrating AI. Methods: A total of 53 patients treated in the pelvic region were enrolled and split into training (30), validation (9), and testing (14). A Generative Adversarial Network (GAN) was trained for 200 epochs. The image accuracy was evaluated by calculating the mean and mean absolute error (ME and ME) between sCT and CT. RT treatment plans were calculated on CT and sCT images, and dose accuracy was evaluated considering Dose Volume Histogram (DVH) and gamma analysis. Results: A total of 4507 images were selected for training. The MAE and ME values in the test set were 36 ± 6 HU and 7 ± 6 HU, respectively. Mean gamma passing rates for 1%/1 mm, 2%/2 mm, and 3%/3 mm tolerance criteria were respectively 93.5 ± 3.4%, 98.0 ± 1.3%, and 99.2 ± 0.7%, with no difference between curative and palliative cases. All the DVH parameters analysed were within 1 Gy of the difference between sCT and CT. Conclusion: This study demonstrated that sCT generation using the DL approach is feasible on low-dose CBCT images. The proposed approach can represent a valid tool to speed up the online adaptive procedure and remove CT simulation from the RT workflow.