DeepDose: Towards a fast dose calculation engine for radiation therapy using deep learning

Deep learning applied to dose prediction in external radiation therapy: A narrative review

Over the last decades, the use of artificial intelligence, machine learning and deep learning in medical fields has skyrocketed. Well known for their results in segmentation, motion management and posttreatment outcome tasks, investigations of machine learning and deep learning models as fast dose calculation or quality assurance tools have been present since 2000. The main motivation for this increasing research and interest in artificial intelligence, machine learning and deep learning is the enhancement of treatment workflows, specifically dosimetry and quality assurance accuracy and time points, which remain important time-consuming aspects of clinical patient management. Since 2014, the evolution of models and architectures for dose calculation has been related to innovations and interest in the theory of information research with pronounced improvements in architecture design. The use of knowledge-based approaches to patient-specific methods has also considerably improved the accuracy of dose predictions. This paper covers the state of all known deep learning architectures and models applied to external radiotherapy with a description of each architecture, followed by a discussion on the performance and future of deep learning predictive models in external radiotherapy.

Application and progress of artificial intelligence in radiation therapy dose prediction

Radiation therapy (RT) nowadays is a main treatment modality of cancer. To ensure the therapeutic efficacy of patients, accurate dose distribution is often required, which is a time-consuming and labor-intensive process. In addition, due to the differences in knowledge and experience among participants and diverse institutions, the predicted dose are often inconsistent. In last several decades, artificial intelligence (AI) has been applied in various aspects of RT, several products have been implemented in clinical practice and confirmed superiority. In this paper, we will review the research of AI in dose prediction, focusing on the progress in deep learning (DL).

AS-NeSt: A Novel 3D Deep Learning Model for Radiation Therapy Dose Distribution Prediction in Esophageal Cancer Treatment With Multiple Prescriptions

PURPOSE

Implementing artificial intelligence technologies allows for the accurate prediction of radiation therapy dose distributions, enhancing treatment planning efficiency. However, esophageal cancers present unique challenges because of tumor complexity and diverse prescription types. Additionally, limited data availability hampers the effectiveness of existing artificial intelligence models. This study developed a deep learning model, trained on a diverse data set of esophageal cancer prescriptions, to improve dose prediction accuracy.

METHODS AND MATERIALS

We retrospectively collected data from 530 patients with esophageal cancer, including single-target and simultaneous integrated boost prescriptions, for model building. The proposed Asymmetric ResNeSt (AS-NeSt) model features novel 3-dimensional (3D) ResNeSt blocks and an asymmetrical architecture. We constructed a loss function targeting global and local doses and validated the model's performance against existing alternatives. Model-assisted experiments were used to validate its clinical benefits.

RESULTS

The AS-NeSt model maintained an absolute prediction error below 5% for each dosimetric metric. The average Dice similarity coefficient for isodose volumes was 0.93. The model achieved an average relative prediction error of 2.02%, statistically lower than Hierarchically Densely Connected U-net (4.17%), DoseNet (2.35%), and Densely Connected Network (3.65%). It also demonstrated significantly fewer parameters and shorter prediction times. Clinically, the AS-NeSt model raised physicians' ability to accurately preassess appropriate treatment methods before planning from 95.24% to 100%, reduced planning time by over 61% for junior dosimetrists and 52% for senior dosimetrists, and decreased both inter- and intra-dosimetrist discrepancies by more than 50%.

CONCLUSIONS

The AS-NeSt model, developed with innovative 3D ResNeSt blocks and an asymmetrical encoder-decoder structure, has been validated using clinical esophageal cancer patient data. It accurately predicts 3D dose distributions for various prescriptions, including simultaneous integrated boost, showing potential to improve the management of esophageal cancer treatment in a clinical setting.

A generalization performance study on the boosting radiotherapy dose calculation engine based on super-resolution

PURPOSE

During the radiation treatment planning process, one of the time-consuming procedures is the final high-resolution dose calculation, which obstacles the wide application of the emerging online adaptive radiotherapy techniques (OLART). There is an urgent desire for highly accurate and efficient dose calculation methods. This study aims to develop a dose super resolution-based deep learning model for fast and accurate dose prediction in clinical practice.

METHOD

A Multi-stage Dose Super-Resolution Network (MDSR Net) architecture with sparse masks module and multi-stage progressive dose distribution restoration method were developed to predict high-resolution dose distribution using low-resolution data. A total of 340 VMAT plans from different disease sites were used, among which 240 randomly selected nasopharyngeal, lung, and cervix cases were used for model training, and the remaining 60 cases from the same sites for model benchmark testing, and additional 40 cases from the unseen site (breast and rectum) was used for model generalizability evaluation. The clinical calculated dose with a grid size of 2 mm was used as baseline dose distribution. The input included the dose distribution with 4 mm grid size and CT images. The model performance was compared with HD U-Net and cubic interpolation methods using Dose-volume histograms (DVH) metrics and global gamma analysis with 1%/1 mm and 10% low dose threshold. The correlation between the prediction error and the dose, dose gradient, and CT values was also evaluated.

RESULTS

The prediction errors of MDSR were 0.06-0.84% of Dmean indices, and the gamma passing rate was 83.1-91.0% on the benchmark testing dataset, and 0.02-1.03% and 71.3-90.3% for the generalization dataset respectively. The model performance was significantly higher than the HD U-Net and interpolation methods (p < 0.05). The mean errors of the MDSR model decreased (monotonously by 0.03-0.004%) with dose and increased (by 0.01-0.73%) with the dose gradient. There was no correlation between prediction errors and the CT values.

CONCLUSION

The proposed MDSR model achieved good agreement with the baseline high-resolution dose distribution, with small prediction errors for DVH indices and high gamma passing rate for both seen and unseen sites, indicating a robust and generalizable dose prediction model. The model can provide fast and accurate high-resolution dose distribution for clinical dose calculation, particularly for the routine practice of OLART.

A deep learning based dynamic arc radiotherapy photon dose engine trained on Monte Carlo dose distributions

Background and purpose

Despite hardware acceleration, state-of-the-art Monte Carlo (MC) dose engines require considerable computation time to reduce stochastic noise. We developed a deep learning (DL) based dose engine reaching high accuracy at strongly reduced computation times.

Materials and methods

Radiotherapy treatment plans and computed tomography scans were collected for 350 treatments in a variety of tumor sites. Dose distributions were computed using a MC dose engine for ∼ 30,000 separate segments at 6 MV and 10 MV beam energies, both flattened and flattening filter free. For dynamic arcs these explicitly incorporated the leaf, jaw and gantry motions during dose delivery. A neural network was developed, combining two-dimensional convolution and recurrence using 64 hidden channels. Parameters were trained to minimize the mean squared log error loss between the MC computed dose and the model output. Full dose distributions were reconstructed for 100 additional treatment plans. Gamma analyses were performed to assess accuracy.

Results

DL dose evaluation was on average 82 times faster than MC computation at a 1 % accuracy setting. In voxels receiving at least 10 % of the maximum dose the overall global gamma pass rate using a 2 % and 2 mm criterion was 99.6 %, while mean local gamma values were accurate within 2 %. In the high dose region over 50 % of maximum the mean local gamma approached a 1 % accuracy.

Conclusions

A DL based dose engine was implemented, able to accurately reproduce MC computed dynamic arc radiotherapy dose distributions at high speed.

Monte Carlo Dosimetry Validation for X-Ray Guided Endovascular Procedures

BACKGROUND

Endovascular treatment is continuously gaining ground in vascular surgery procedures. However, current patient radiation dose estimation does not take into account the exact patient morphology and organs' composition. Monte Carlo (MC) simulation can accurately estimate the dose by recreating the irradiation process generated during X-ray-guided interventions. This study aimed to validate the MC simulation models by comparing simulated and measured dose distributions in endovascular aortic aneurysm repair (EVAR) procedures.

METHODS

We conducted a clinical study in patients treated for EVAR. Patient dose measurements were taken with passive dosimeters using Optically Stimulated Luminescence technology in 4 specific anatomical points on the skin: xiphoid process, pubic symphysis, right and left iliac crest. Dose measurements were compared to the corresponding simulated doses with the Geant4 Application for Emission Tomography (GATE) and GPU Geant4-based Monte Carlo Simulations (GGEMS) MC simulations softwares. The MC simulation took as input the computed tomography scan of the patient and the parameters of the imaging system (orientation angles, tube voltage, and aluminum filtration) and gives as output the three-dimensional (3D) dose map for each patient and angulation.

RESULTS

A good agreement with real doses was found for doses simulated by the MC GATE method (P < 0.0001; r = 0.97; 95% confidence interval [CI] [0.96-0.98]), as well as for doses simulated by the GGEMS method (P < 0.0001; r = 0.96; 95% CI [0.94-0.97]). The mean relative error for all measurements was 5 ± 5% in the MC GATE group and 6 ± 5% in the GGEMS group. Process execution on GGEMS (6 sec) was faster than the GATE MC simulation (5 hr).

CONCLUSION

Considering the current imaging settings, this study shows the potential of using the GATE and GGEMS MC simulations platforms to model the 3D dose distributions during EVAR procedures.

Invertible and Variable Augmented Network for Pretreatment Patient-Specific Quality Assurance Dose Prediction

Pretreatment patient-specific quality assurance (prePSQA) is conducted to confirm the accuracy of the radiotherapy dose delivered. However, the process of prePSQA measurement is time consuming and exacerbates the workload for medical physicists. The purpose of this work is to propose a novel deep learning (DL) network to improve the accuracy and efficiency of prePSQA. A modified invertible and variable augmented network was developed to predict the three-dimensional (3D) measurement-guided dose (MDose) distribution of 300 cancer patients who underwent volumetric modulated arc therapy (VMAT) between 2018 and 2021, in which 240 cases were randomly selected for training, and 60 for testing. For simplicity, the present approach was termed as "IVPSQA." The input data include CT images, radiotherapy dose exported from the treatment planning system, and MDose distribution extracted from the verification system. Adam algorithm was used for first-order gradient-based optimization of stochastic objective functions. The IVPSQA model obtained high-quality 3D prePSQA dose distribution maps in head and neck, chest, and abdomen cases, and outperformed the existing U-Net-based prediction approaches in terms of dose difference maps and horizontal profiles comparison. Moreover, quantitative evaluation metrics including SSIM, MSE, and MAE demonstrated that the proposed approach achieved a good agreement with ground truth and yield promising gains over other advanced methods. This study presented the first work on predicting 3D prePSQA dose distribution by using the IVPSQA model. The proposed method could be taken as a clinical guidance tool and help medical physicists to reduce the measurement work of prePSQA.

Advancing equitable and personalized cancer care: Novel applications and priorities of artificial intelligence for fairness and inclusivity in the patient care workflow

Patient care workflows are highly multimodal and intertwined: the intersection of data outputs provided from different disciplines and in different formats remains one of the main challenges of modern oncology. Artificial Intelligence (AI) has the potential to revolutionize the current clinical practice of oncology owing to advancements in digitalization, database expansion, computational technologies, and algorithmic innovations that facilitate discernment of complex relationships in multimodal data. Within oncology, radiation therapy (RT) represents an increasingly complex working procedure, involving many labor-intensive and operator-dependent tasks. In this context, AI has gained momentum as a powerful tool to standardize treatment performance and reduce inter-observer variability in a time-efficient manner. This review explores the hurdles associated with the development, implementation, and maintenance of AI platforms and highlights current measures in place to address them. In examining AI's role in oncology workflows, we underscore that a thorough and critical consideration of these challenges is the only way to ensure equitable and unbiased care delivery, ultimately serving patients' survival and quality of life.

MRI-Guided Adaptive Radiation Therapy

Magnetic resonance imaging-guided radiation therapy (MRIgRT) has improved soft tissue contrast over computed tomography (CT) based image-guided RT. Superior visualization of the target and surrounding radiosensitive structures has the potential to improve oncological outcomes partly due to safer dose-escalation and adaptive planning. In this review, we highlight the workflow of adaptive MRIgRT planning, which includes simulation imaging, daily MRI, identifying isocenter shifts, contouring, plan optimization, quality control, and delivery. Increased utilization of MRIgRT will depend on addressing technical limitations of this technology, while addressing treatment efficacy, cost-effectiveness, and workflow training.

A modified U-net with graph representation for dose prediction in esophageal cancer radiotherapy plans

The manual design of esophageal cancer radiotherapy plan is time-consuming and labor-intensive. Automatic planning (AP) is prevalent nowadays to increase physicists' work efficiency. Because of the intuitiveness of dose distribution in AP evaluation, obtaining reasonable dose prediction provides effective guarantees to generate a satisfactory AP. Existing fully convolutional network-based methods for predicting dose distribution in esophageal cancer radiotherapy plans often capture features in a limited receptive field. Additionally, the correlations between voxel pairs are often ignored. This work modifies the U-net architecture and exploits graph convolution to capture long-range information for dose prediction in esophageal cancer plans. Meanwhile, attention mechanism gets correlations between planning target volume (PTV) and organs at risk, and adaptively learns their feature weights. Finally, a novel loss function that considers features between voxel pairs is used to highlight the predictions. 152 subjects with prescription doses of 50 Gy or 60 Gy are collected in this study. The mean absolute error and standard deviation of conformity index, homogeneity index, and max dose for PTV achieved by the proposed method are 0.036 ± 0.030, 0.036 ± 0.027, and 0.930 ± 1.162, respectively, which outperform other state-of-the-art models. The superior performance demonstrates that our proposed method has great potential for AP generation.

Real-time motion management in MRI-guided radiotherapy: Current status and AI-enabled prospects

MRI-guided radiotherapy (MRIgRT) is a highly complex treatment modality, allowing adaptation to anatomical changes occurring from one treatment day to the other (inter-fractional), but also to motion occurring during a treatment fraction (intra-fractional). In this vision paper, we describe the different steps of intra-fractional motion management during MRIgRT, from imaging to beam adaptation, and the solutions currently available both clinically and at a research level. Furthermore, considering the latest developments in the literature, a workflow is foreseen in which motion-induced over- and/or under-dosage is compensated in 3D, with minimal impact to the radiotherapy treatment time. Considering the time constraints of real-time adaptation, a particular focus is put on artificial intelligence (AI) solutions as a fast and accurate alternative to conventional algorithms.

Generalisation of radiotherapy dose calculation for Monte Carlo algorithm combined with 3D Swin-Unet: a multi-institutional IMRT evaluation

Objective.Accurate dose calculations are essential prerequisites for precise radiotherapy. The integration of deep learning into dosimetry could consider computational accuracy and efficiency and has potential applicability to clinical dose calculation. The generalisation of a deep learning dose calculation method (hereinafter referred to as TERMA-Monte Carlo network, T-MC net) was evaluated in clinical practice using intensity-modulated radiotherapy (IMRT) plans for various human body regions and multiple institutions, with the Monte Carlo (MC) algorithm serving as a benchmark.Approach. Sixty IMRT plans were selected from four institutions for testing the head and neck, chest and abdomen, and pelvis regions. Using the MC results as the benchmark, the T-MC net calculation results were used to perform three-dimensional dose distribution and dose-volume histogram (DVH) comparisons of the entire body, planning target volume (PTV) and organs at risk (OARs), respectively, and calculate the mean ±95% confidence interval of gamma pass rate (GPR), percentage of agreement (PA) and dose difference ratio (DDR) of dose indices D95, D50, and D5.Main results. For the entire body, the GPRs of 3%/3 mm, 2%/2 mm, 2%/1 mm, and the PA were 99.62 ± 0.32%, 98.50 ± 1.09%, 95.60 ± 2.90% and 97.80 ± 1.12%, respectively. For the PTV, the GPRs of 3%/3 mm, 2%/2 mm, 2%/1 mm and the PA were 98.90 ± 1.00%, 95.78 ± 2.83%, 92.23 ± 4.74% and 98.93 ± 0.62%, respectively. The absolute value of average DDR was less than 1.4%.Significance. We proposed a general dose calculation framework based on deep learning, using the MC algorithm as a benchmark, performing a generalisation test for IMRT treatment plans across multiple institutions. The framework provides high computational speed while maintaining the accuracy of MC and may become an effective dose algorithm engine in treatment planning, adaptive radiotherapy, and dose verification.

Deep-Learning-Based Dose Predictor for Glioblastoma-Assessing the Sensitivity and Robustness for Dose Awareness in Contouring

External beam radiation therapy requires a sophisticated and laborious planning procedure. To improve the efficiency and quality of this procedure, machine-learning models that predict these dose distributions were introduced. The most recent dose prediction models are based on deep-learning architectures called 3D U-Nets that give good approximations of the dose in 3D almost instantly. Our purpose was to train such a 3D dose prediction model for glioblastoma VMAT treatment and test its robustness and sensitivity for the purpose of quality assurance of automatic contouring. From a cohort of 125 glioblastoma (GBM) patients, VMAT plans were created according to a clinical protocol. The initial model was trained on a cascaded 3D U-Net. A total of 60 cases were used for training, 15 for validation and 20 for testing. The prediction model was tested for sensitivity to dose changes when subject to realistic contour variations. Additionally, the model was tested for robustness by exposing it to a worst-case test set containing out-of-distribution cases. The initially trained prediction model had a dose score of 0.94 Gy and a mean DVH (dose volume histograms) score for all structures of 1.95 Gy. In terms of sensitivity, the model was able to predict the dose changes that occurred due to the contour variations with a mean error of 1.38 Gy. We obtained a 3D VMAT dose prediction model for GBM with limited data, providing good sensitivity to realistic contour variations. We tested and improved the model's robustness by targeted updates to the training set, making it a useful technique for introducing dose awareness in the contouring evaluation and quality assurance process.

Performance assessment of variant UNet-based deep-learning dose engines for MR-Linac-based prostate IMRT plans

Objective. UNet-based deep-learning (DL) architectures are promising dose engines for traditional linear accelerator (Linac) models. Current UNet-based engines, however, were designed differently with various strategies, making it challenging to fairly compare the results from different studies. The objective of this study is to thoroughly evaluate the performance of UNet-based models on magnetic-resonance (MR)-Linac-based intensity-modulated radiation therapy (IMRT) dose calculations.Approach. The UNet-based models, including the standard-UNet, cascaded-UNet, dense-dilated-UNet, residual-UNet, HD-UNet, and attention-aware-UNet, were implemented. The model input is patient CT and IMRT field dose in water, and the output is patient dose calculated by DL model. The reference dose was calculated by the Monaco Monte Carlo module. Twenty training and ten test cases of prostate patients were included. The accuracy of the DL-calculated doses was measured using gamma analysis, and the calculation efficiency was evaluated by inference time.Results. All the studied models effectively corrected low-accuracy doses in water to high-accuracy patient doses in a magnetic field. The gamma passing rates between reference and DL-calculated doses were over 86% (1%/1 mm), 98% (2%/2 mm), and 99% (3%/3 mm) for all the models. The inference times ranged from 0.03 (graphics processing unit) to 7.5 (central processing unit) seconds. Each model demonstrated different strengths in calculation accuracy and efficiency; Res-UNet achieved the highest accuracy, HD-UNet offered high accuracy with the fewest parameters but the longest inference, dense-dilated-UNet was consistently accurate regardless of model levels, standard-UNet had the shortest inference but relatively lower accuracy, and the others showed average performance. Therefore, the best-performing model would depend on the specific clinical needs and available computational resources.Significance. The feasibility of using common UNet-based models for MR-Linac-based dose calculations has been explored in this study. By using the same model input type, patient training data, and computing environment, a fair assessment of the models' performance was present.

Fast dose calculation in x-ray guided interventions by using deep learning

Objective.Patient dose estimation in x-ray-guided interventions is essential to prevent radiation-induced biological side effects. Current dose monitoring systems estimate the skin dose based in dose metrics such as the reference air kerma. However, these approximations do not take into account the exact patient morphology and organs composition. Furthermore, accurate organ dose estimation has not been proposed for these procedures. Monte Carlo simulation can accurately estimate the dose by recreating the irradiation process generated during the x-ray imaging, but at a high computation time, limiting an intra-operative application. This work presents a fast deep convolutional neural network trained with MC simulations for patient dose estimation during x-ray-guided interventions.Approach.We introduced a modified 3D U-Net that utilizes a patient's CT scan and the numerical values of imaging settings as input to produce a Monte Carlo dose map. To create a dataset of dose maps, we simulated the x-ray irradiation process for the abdominal region using a publicly available dataset of 82 patient CT scans. The simulation involved varying the angulation, position, and tube voltage of the x-ray source for each scan. We additionally conducted a clinical study during endovascular abdominal aortic repairs to validate the reliability of our Monte Carlo simulation dose maps. Dose measurements were taken at four specific anatomical points on the skin and compared to the corresponding simulated doses. The proposed network was trained using a 4-fold cross-validation approach with 65 patients, and evaluating the performance on the remaining 17 patients during testing.Main results.The clinical validation demonstrated a average error within the anatomical points of 5.1%. The network yielded test errors of 11.5 ± 4.6% and 6.2 ± 1.5% for peak and average skin doses, respectively. Furthermore, the mean errors for the abdominal region and pancreas doses were 5.0 ± 1.4% and 13.1 ± 2.7%, respectively.Significance.Our network can accurately predict a personalized 3D dose map considering the current imaging settings. A short computation time was achieved, making our approach a potential solution for dose monitoring and reporting commercial systems.

TrDosePred: A deep learning dose prediction algorithm based on transformers for head and neck cancer radiotherapy

BACKGROUND

Intensity-Modulated Radiation Therapy (IMRT) has been the standard of care for many types of tumors. However, treatment planning for IMRT is a time-consuming and labor-intensive process.

PURPOSE

To alleviate this tedious planning process, a novel deep learning based dose prediction algorithm (TrDosePred) was developed for head and neck cancers.

METHODS

The proposed TrDosePred, which generated the dose distribution from a contoured CT image, was a U-shape network constructed with a convolutional patch embedding and several local self-attention based transformers. Data augmentation and ensemble approach were used for further improvement. It was trained based on the dataset from Open Knowledge-Based Planning Challenge (OpenKBP). The performance of TrDosePred was evaluated with two mean absolute error (MAE) based scores utilized by OpenKBP challenge (i.e., Dose score and DVH score) and compared to the top three approaches of the challenge. In addition, several state-of-the-art methods were implemented and compared to TrDosePred.

RESULTS

The TrDosePred ensemble achieved the dose score of 2.426 Gy and the DVH score of 1.592 Gy on the test dataset, ranking at 3rd and 9th respectively in the leaderboard on CodaLab as of writing. In terms of DVH metrics, on average, the relative MAE against the clinical plans was 2.25% for targets and 2.17% for organs at risk.

CONCLUSIONS

A transformer-based framework TrDosePred was developed for dose prediction. The results showed a comparable or superior performance as compared to the previous state-of-the-art approaches, demonstrating the potential of transformer to boost the treatment planning procedures.

Analytical models for external photon beam radiotherapy out-of-field dose calculation: a scoping review

A growing body of scientific evidence indicates that exposure to low dose ionizing radiation (< 2 Gy) is associated with a higher risk of developing radio-induced cancer. Additionally, it has been shown to have significant impacts on both innate and adaptive immune responses. As a result, the evaluation of the low doses inevitably delivered outside the treatment fields (out-of-field dose) in photon radiotherapy is a topic that is regaining interest at a pivotal moment in radiotherapy. In this work, we proposed a scoping review in order to identify evidence of strengths and limitations of available analytical models for out-of-field dose calculation in external photon beam radiotherapy for the purpose of implementation in clinical routine. Papers published between 1988 and 2022 proposing a novel analytical model that estimated at least one component of the out-of-field dose for photon external radiotherapy were included. Models focusing on electrons, protons and Monte-Carlo methods were excluded. The methodological quality and potential limitations of each model were analyzed to assess their generalizability. Twenty-one published papers were selected for analysis, of which 14 proposed multi-compartment models, demonstrating that research efforts are directed towards an increasingly detailed description of the underlying physical phenomena. Our synthesis revealed great inhomogeneities in practices, in particular in the acquisition of experimental data and the standardization of measurements, in the choice of metrics used for the evaluation of model performance and even in the definition of regions considered out-of-the-field, which makes quantitative comparisons impossible. We therefore propose to clarify some key concepts. The analytical methods do not seem to be easily suitable for massive use in clinical routine, due to the inevitable cumbersome nature of their implementation. Currently, there is no consensus on a mathematical formalism that comprehensively describes the out-of-field dose in external photon radiotherapy, partly due to the complex interactions between a large number of influencing factors. Out-of-field dose calculation models based on neural networks could be promising tools to overcome these limitations and thus favor a transfer to the clinic, but the lack of sufficiently large and heterogeneous data sets is the main obstacle.

Accurate and Fast Deep Learning Dose Prediction for a Preclinical Microbeam Radiation Therapy Study Using Low-Statistics Monte Carlo Simulations

Microbeam radiation therapy (MRT) utilizes coplanar synchrotron radiation beamlets and is a proposed treatment approach for several tumor diagnoses that currently have poor clinical treatment outcomes, such as gliosarcomas. Monte Carlo (MC) simulations are one of the most used methods at the Imaging and Medical Beamline, Australian Synchrotron to calculate the dose in MRT preclinical studies. The steep dose gradients associated with the 50μm-wide coplanar beamlets present a significant challenge for precise MC simulation of the dose deposition of an MRT irradiation treatment field in a short time frame. The long computation times inhibit the ability to perform dose optimization in treatment planning or apply online image-adaptive radiotherapy techniques to MRT. Much research has been conducted on fast dose estimation methods for clinically available treatments. However, such methods, including GPU Monte Carlo implementations and machine learning (ML) models, are unavailable for novel and emerging cancer radiotherapy options such as MRT. In this work, the successful application of a fast and accurate ML dose prediction model for a preclinical MRT rodent study is presented for the first time. The ML model predicts the peak doses in the path of the microbeams and the valley doses between them, delivered to the tumor target in rat patients. A CT imaging dataset is used to generate digital phantoms for each patient. Augmented variations of the digital phantoms are used to simulate with Geant4 the energy depositions of an MRT beam inside the phantoms with 15% (high-noise) and 2% (low-noise) statistical uncertainty. The high-noise MC simulation data are used to train the ML model to predict the energy depositions in the digital phantoms. The low-noise MC simulations data are used to test the predictive power of the ML model. The predictions of the ML model show an agreement within 3% with low-noise MC simulations for at least 77.6% of all predicted voxels (at least 95.9% of voxels containing tumor) in the case of the valley dose prediction and for at least 93.9% of all predicted voxels (100.0% of voxels containing tumor) in the case of the peak dose prediction. The successful use of high-noise MC simulations for the training, which are much faster to produce, accelerates the production of the training data of the ML model and encourages transfer of the ML model to different treatment modalities for other future applications in novel radiation cancer therapies.

A deep-learning-based dose verification tool utilizing fluence maps for a cobalt-60 compensator-based intensity-modulated radiation therapy system

Background and purpose

A novel cobalt-60 compensator-based intensity-modulated radiation therapy (IMRT) system was developed for a resource-limited environment but lacked an efficient dose verification algorithm. The aim of this study was to develop a deep-learning-based dose verification algorithm for accurate and rapid dose predictions.

Materials and methods

A deep-learning network was employed to predict the doses from static fields related to beam commissioning. Inputs were a cube-shaped phantom, a beam binary mask, and an intersecting volume of the phantom and beam binary mask, while output was a 3-dimensional (3D) dose. The same network was extended to predict patient-specific doses for head and neck cancers using two different approaches. A field-based method predicted doses for each field and combined all calculated doses into a plan, while the plan-based method combined all nine fluences into a plan to predict doses. Inputs included patient computed tomography (CT) scans, binary beam masks, and fluence maps truncated to the patient's CT in 3D.

Results

For static fields, predictions agreed well with ground truths with average deviations of less than 0.5% for percent depth doses and profiles. Even though the field-based method showed excellent prediction performance for each field, the plan-based method showed better agreement between clinical and predicted dose distributions. The distributed dose deviations for all planned target volumes and organs at risk were within 1.3 Gy. The calculation speed for each case was within two seconds.

Conclusions

A deep-learning-based dose verification tool can accurately and rapidly predict doses for a novel cobalt-60 compensator-based IMRT system.

Prediction of dose deposition matrix using voxel features driven machine learning approach

OBJECTIVES

A dose deposition matrix (DDM) prediction method using several voxel features and a machine learning (ML) approach is proposed for plan optimization in radiation therapy.

METHODS

Head and lung cases with the inhomogeneous medium are used as training and testing data. The prediction model is a cascade forward backprop neural network where the input is the features of the voxel, including 1) voxel to body surface distance along the beamlet axis, 2) voxel to beamlet axis distance, 3) voxel density, 4) heterogeneity corrected voxel to body surface distance, 5) heterogeneity corrected voxel to beamlet axis, and (6) the dose of voxel obtained from the pencil beam (PB) algorithm. The output is the predicted voxel dose corresponding to a beamlet. The predicted DDM was used for plan optimization (ML method) and compared with the dose of MC-based plan optimization (MC method) and the dose of pencil beam-based plan optimization (PB method). The mean absolute error (MAE) value was calculated for full volume relative to the dose of the MC method to evaluate the overall dose performance of the final plan.

RESULTS

For patient with head tumor, the ML method achieves MAE value 0.49 × 10-4 and PB has MAE 1.86 × 10-4. For patient with lung tumor, the ML method has MAE 1.42 × 10-4 and PB has MAE 3.72 × 10-4. The maximum percentage difference in PTV dose coverage (D98) between ML and MC methods is no more than 1.2% for patient with head tumor, while the difference is larger than 10% using the PB method. For patient with lung tumor, the maximum percentage difference in PTV dose coverage (D98) between ML and MC methods is no more than 2.1%, while the difference is larger than 16% using the PB method.

CONCLUSIONS

In this work, a reliable DDM prediction method is established for plan optimization by applying several voxel features and the ML approach. The results show that the ML method based on voxel features can obtain plans comparable to the MC method and is better than the PB method in achieving accurate dose to the patient, which is helpful for rapid plan optimization and accurate dose calculation.

ADVANCES IN KNOWLEDGE

Establishment of a new machine learning method based on the relationship between the voxel and beamlet features for dose deposition matrix prediction in radiation therapy.

An ultra-fast deep-learning-based dose engine for prostate VMAT via knowledge distillation framework with limited patient data

Objective. Deep-learning (DL)-based dose engines have been developed to alleviate the intrinsic compromise between the calculation accuracy and efficiency of the traditional dose calculation algorithms. However, current DL-based engines typically possess high computational complexity and require powerful computing devices. Therefore, to mitigate their computational burdens and broaden their applicability to a clinical setting where resource-limited devices are available, we proposed a compact dose engine via knowledge distillation (KD) framework that offers an ultra-fast calculation speed with high accuracy for prostate Volumetric Modulated Arc Therapy (VMAT).Approach. The KD framework contains two sub-models: a large pre-trained teacher and a small to-be-trained student. The student receives knowledge transferred from the teacher for better generalization. The trained student serves as the final engine for dose calculation. The model input is patient computed tomography and VMAT dose in water, and the output is DL-calculated patient dose. The ground-truth \dose was computed by the Monte Carlo module of the Monaco treatment planning system. Twenty and ten prostate cases were included for model training and assessment, respectively. The model's performance (teacher/student/student-only) was evaluated by Gamma analysis and inference efficiency.Main results. The dosimetric comparisons (input/DL-calculated/ground-truth doses) suggest that the proposed engine can effectively convert low-accuracy doses in water to high-accuracy patient doses. The Gamma passing rate (2%/2 mm, 10% threshold) between the DL-calculated and ground-truth doses was 98.64 ± 0.62% (teacher), 98.13 ± 0.76% (student), and 96.95 ± 1.02% (student-only). The inference time was 16 milliseconds (teacher) and 11 milliseconds (student/student-only) using a graphics processing unit device, while it was 936 milliseconds (teacher) and 374 milliseconds (student/student-only) using a central processing unit device.Significance. With the KD framework, a compact dose engine can achieve comparable accuracy to that of a larger one. Its compact size reduces the computational burdens and computing device requirements, and thus such an engine can be more clinically applicable.

Small beams, fast predictions: a comparison of machine learning dose prediction models for proton minibeam therapy

BACKGROUND

Dose calculations for novel radiotherapy cancer treatments such as proton minibeam radiation therapy is often done using full Monte Carlo (MC) simulations. As MC simulations can be very time consuming for this kind of application, deep learning models have been considered to accelerate dose estimation in cancer patients.

PURPOSE

This work systematically evaluates the dose prediction accuracy, speed and generalization performance of three selected state-of-the-art deep learning models for dose prediction applied to the proton minibeam therapy. The strengths and weaknesses of those models are thoroughly investigated, helping other researchers to decide on a viable algorithm for their own application.

METHODS

The following recently published models are compared: first, a 3D U-Net model trained as a regression network, second, a 3D U-Net trained as a generator of a generative adversarial network (GAN) and third, a dose transformer model which interprets the dose prediction as a sequence translation task. These models are trained to emulate the result of MC simulations. The dose depositions of a proton minibeam with a diameter of 800μm and an energy of 20-100 MeV inside a simple head phantom calculated by full Geant4 MC simulations are used as a case study for this comparison. The spatial resolution is 0.5 mm. Special attention is put on the evaluation of the generalization performance of the investigated models.

RESULTS

Dose predictions with all models are produced in the order of a second on a GPU, the 3D U-Net models being fastest with an average of 130 ms. An investigated 3D U-Net regression model is found to show the strongest performance with overall 61.0 % ± $\%\pm$ 0.5% of all voxels exhibiting a deviation in energy deposition prediction of less than 3% compared to full MC simulations with no spatial deviation allowed. The 3D U-Net models are observed to show better generalization performance for target geometry variations, while the transformer-based model shows better generalization with regard to the proton energy.

CONCLUSIONS

This paper reveals that (1) all studied deep learning models are significantly faster than non-machine learning approaches predicting the dose in the order of seconds compared to hours for MC, (2) all models provide reasonable accuracy, and (3) the regression-trained 3D U-Net provides the most accurate predictions.

Advances in Automated Treatment Planning

Treatment planning in radiation therapy has progressed enormously over the past several decades. Such advancements came in the form of innovative hardware and algorithms, giving rise to modalities such as intensity-modulated radiation therapy and volume modulated arc therapy, greatly improving patient outcome and quality of life. While these developments have improved the overall plan quality, they have also given rise to higher treatment planning complexity. This has resulted in increased treatment planning time and higher variability in the final approved plan quality. Radiation oncology, as an already technologically advanced field, has much research and implementation involving the use of AI. The field has begun to show the efficacy of using such technologies in many of its sub-areas, such as in diagnosis, imaging, segmentation, treatment planning, quality assurance, treatment delivery, and follow-up. Some AI technologies have already been clinically implemented by commercial systems. In this article, we will provide an overview to methods involved with treatment planning in radiation therapy. In particular, we will review the recent research and literature related to automation of the treatment planning process, leading to potentially higher efficiency and higher quality plans. We will then present the current and future challenges, as well as some future perspectives.

Enabling non-isocentric dynamic trajectory radiotherapy by integration of dynamic table translations

Objective. The purpose of this study is to develop a treatment planning process (TPP) for non-isocentric dynamic trajectory radiotherapy (DTRT) using dynamic gantry rotation, collimator rotation, table rotation, longitudinal, vertical and lateral table translations and intensity modulation and to validate the dosimetric accuracy. Approach. The TPP consists of two steps. First, a path describing the dynamic gantry rotation, collimator rotation and dynamic table rotation and translations is determined. Second, an optimization of the intensity modulation along the path is performed. We demonstrate the TPP for three use cases. First, a non-isocentric DTRT plan for a brain case is compared to an isocentric DTRT plan in terms of dosimetric plan quality and delivery time. Second, a non-isocentric DTRT plan for a craniospinal irradiation (CSI) case is compared to a multi-isocentric intensity modulated radiotherapy (IMRT) plan. Third, a non-isocentric DTRT plan for a bilateral breast case is compared to a multi-isocentric volumetric modulated arc therapy (VMAT) plan. The non-isocentric DTRT plans are delivered on a TrueBeam in developer mode and their dosimetric accuracy is validated using radiochromic films. Main results. The non-isocentric DTRT plan for the brain case is similar in dosimetric plan quality and delivery time to the isocentric DTRT plan but is expected to reduce the risk of collisions. The DTRT plan for the CSI case shows similar dosimetric plan quality while reducing the delivery time by 45% in comparison with the IMRT plan. The DTRT plan for the breast case showed better treatment plan quality in comparison with the VMAT plan. The gamma passing rates between the measured and calculated dose distributions are higher than 95% for all three plans. Significance. The versatile benefits of non-isocentric DTRT are demonstrated with three use cases, namely reduction of collision risk, reduced setup and delivery time and improved dosimetric plan quality.

A novel multichannel deep learning model for fast denoising of Monte Carlo dose calculations: preclinical applications

Objective. In preclinical radiotherapy with kilovolt (kV) x-ray beams, accurate treatment planning is needed to improve the translation potential to clinical trials. Monte Carlo based radiation transport simulations are the gold standard to calculate the absorbed dose distribution in external beam radiotherapy. However, these simulations are notorious for their long computation time, causing a bottleneck in the workflow. Previous studies have used deep learning models to speed up these simulations for clinical megavolt (MV) beams. For kV beams, dose distributions are more affected by tissue type than for MV beams, leading to steep dose gradients. This study aims to speed up preclinical kV dose simulations by proposing a novel deep learning pipeline. Approach. A deep learning model is proposed that denoises low precision (∼106 simulated particles) dose distributions to produce high precision (109 simulated particles) dose distributions. To effectively denoise the steep dose gradients in preclinical kV dose distributions, the model uses the novel approach to use the low precision Monte Carlo dose calculation as well as the Monte Carlo uncertainty (MCU) map and the mass density map as additional input channels. The model was trained on a large synthetic dataset and tested on a real dataset with a different data distribution. To keep model inference time to a minimum, a novel method for inference optimization was developed as well. Main results. The proposed model provides dose distributions which achieve a median gamma pass rate (3%/0.3 mm) of 98% with a lower bound of 95% when compared to the high precision Monte Carlo dose distributions from the test set, which represents a different dataset distribution than the training set. Using the proposed model together with the novel inference optimization method, the total computation time was reduced from approximately 45 min to less than six seconds on average. Significance. This study presents the first model that can denoise preclinical kV instead of clinical MV Monte Carlo dose distributions. This was achieved by using the MCU and mass density maps as additional model inputs. Additionally, this study shows that training such a model on a synthetic dataset is not only a viable option, but even increases the generalization of the model compared to training on real data due to the sheer size and variety of the synthetic dataset. The application of this model will enable speeding up treatment plan optimization in the preclinical workflow.

Artificial intelligence in radiotherapy

Radiotherapy is a discipline closely integrated with computer science. Artificial intelligence (AI) has developed rapidly over the past few years. With the explosive growth of medical big data, AI promises to revolutionize the field of radiotherapy through highly automated workflow, enhanced quality assurance, improved regional balances of expert experiences, and individualized treatment guided by multi-omics. In addition to independent researchers, the increasing number of large databases, biobanks, and open challenges significantly facilitated AI studies on radiation oncology. This article reviews the latest research, clinical applications, and challenges of AI in each part of radiotherapy including image processing, contouring, planning, quality assurance, motion management, and outcome prediction. By summarizing cutting-edge findings and challenges, we aim to inspire researchers to explore more future possibilities and accelerate the arrival of AI radiotherapy.

Integrated MRI-guided radiotherapy - opportunities and challenges

MRI can help to categorize tissues as malignant or non-malignant both anatomically and functionally, with a high level of spatial and temporal resolution. This non-invasive imaging modality has been integrated with radiotherapy in devices that can differentially target the most aggressive and resistant regions of tumours. The past decade has seen the clinical deployment of treatment devices that combine imaging with targeted irradiation, making the aspiration of integrated MRI-guided radiotherapy (MRIgRT) a reality. The two main clinical drivers for the adoption of MRIgRT are the ability to image anatomical changes that occur before and during treatment in order to adapt the treatment approach, and to image and target the biological features of each tumour. Using motion management and biological targeting, the radiation dose delivered to the tumour can be adjusted during treatment to improve the probability of tumour control, while simultaneously reducing the radiation delivered to non-malignant tissues, thereby reducing the risk of treatment-related toxicities. The benefits of this approach are expected to increase survival and quality of life. In this Review, we describe the current state of MRIgRT, and the opportunities and challenges of this new radiotherapy approach.

TransDose: a transformer-based UNet model for fast and accurate dose calculation for MR-LINACs

Objective. To present a transformer-based UNet model (TransDose) for fast and accurate dose calculation for magnetic resonance-linear accelerators (MR-LINACs). Approach. A 2D fluence map from each beam was first projected into a 3D fluence volume and then fed into the TransDose model together with patient density volume and output predicted beam dose. The proposed TransDose model combined a 3D residual UNet with a transformer encoder, where convolutional layers extracted the volumetric spatial features, and the transformer encoder processed the long-range dependencies in a global space. Ninety-eight cases with four tumor sites (brain, nasopharynx, lung, and rectum) treated with fixed-beam intensity-modulated radiotherapy were included in the dataset; 78 cases were used for model training and validation; and 20 cases were used for testing. The ground-truth beam doses were calculated with Monte Carlo (MC) simulations within 1% statistical uncertainty and magnetic field strength B = 1.5 T in the superior and inferior direction. Beam angles from the training and validation datasets were rotated 2–5 times, and doses were recalculated to augment the datasets. Results. The dose-volume histograms and indices between the predicted and MC doses showed good consistency. The average 3D γ-passing rates (3%/2 mm, for dose regions above 10% of maximum dose) were 99.13 ± 0.89% (brain), 98.31 ± 1.92% (nasopharynx), 98.74 ± 0.70% (lung), and 99.28 ± 0.25% (rectum). The average dose calculation time, which included the fluence projection and model prediction, was less than 310 ms for each beam. Significance. We successfully developed a transformer-based UNet dose calculation model—TransDose in magnetic fields. Its accuracy and efficiency indicated its potential for use in online adaptive plan optimization for MR-LINACs.

A Survey on Deep Learning for Precision Oncology

Precision oncology, which ensures optimized cancer treatment tailored to the unique biology of a patient’s disease, has rapidly developed and is of great clinical importance. Deep learning has become the main method for precision oncology. This paper summarizes the recent deep-learning approaches relevant to precision oncology and reviews over 150 articles within the last six years. First, we survey the deep-learning approaches categorized by various precision oncology tasks, including the estimation of dose distribution for treatment planning, survival analysis and risk estimation after treatment, prediction of treatment response, and patient selection for treatment planning. Secondly, we provide an overview of the studies per anatomical area, including the brain, bladder, breast, bone, cervix, esophagus, gastric, head and neck, kidneys, liver, lung, pancreas, pelvis, prostate, and rectum. Finally, we highlight the challenges and discuss potential solutions for future research directions.

Towards a safe and efficient clinical implementation of machine learning in radiation oncology by exploring model interpretability, explainability and data-model dependency

The interest in machine learning (ML) has grown tremendously in recent years, partly due to the performance leap that occurred with new techniques of deep learning, convolutional neural networks for images, increased computational power, and wider availability of large datasets. Most fields of medicine follow that popular trend and, notably, radiation oncology is one of those that are at the forefront, with already a long tradition in using digital images and fully computerized workflows. ML models are driven by data, and in contrast with many statistical or physical models, they can be very large and complex, with countless generic parameters. This inevitably raises two questions, namely, the tight dependence between the models and the datasets that feed them, and the interpretability of the models, which scales with its complexity. Any problems in the data used to train the model will be later reflected in their performance. This, together with the low interpretability of ML models, makes their implementation into the clinical workflow particularly difficult. Building tools for risk assessment and quality assurance of ML models must involve then two main points: interpretability and data-model dependency. After a joint introduction of both radiation oncology and ML, this paper reviews the main risks and current solutions when applying the latter to workflows in the former. Risks associated with data and models, as well as their interaction, are detailed. Next, the core concepts of interpretability, explainability, and data-model dependency are formally defined and illustrated with examples. Afterwards, a broad discussion goes through key applications of ML in workflows of radiation oncology as well as vendors' perspectives for the clinical implementation of ML.

Millisecond speed deep learning based proton dose calculation with Monte Carlo accuracy

Objective.Next generation online and real-time adaptive radiotherapy workflows require precise particle transport simulations in sub-second times, which is unfeasible with current analytical pencil beam algorithms (PBA) or Monte Carlo (MC) methods. We present a deep learning based millisecond speed dose calculation algorithm (DoTA) accurately predicting the dose deposited by mono-energetic proton pencil beams for arbitrary energies and patient geometries.Approach.Given the forward-scattering nature of protons, we frame 3D particle transport as modeling a sequence of 2D geometries in the beam's eye view. DoTA combines convolutional neural networks extracting spatial features (e.g. tissue and density contrasts) with a transformer self-attention backbone that routes information between the sequence of geometry slices and a vector representing the beam's energy, and is trained to predict low noise MC simulations of proton beamlets using 80 000 different head and neck, lung, and prostate geometries.Main results.Predicting beamlet doses in 5 ± 4.9 ms with a very high gamma pass rate of 99.37 ± 1.17% (1%, 3 mm) compared to the ground truth MC calculations, DoTA significantly improves upon analytical pencil beam algorithms both in precision and speed. Offering MC accuracy 100 times faster than PBAs for pencil beams, our model calculates full treatment plan doses in 10-15 s depending on the number of beamlets (800-2200 in our plans), achieving a 99.70 ± 0.14% (2%, 2 mm) gamma pass rate across 9 test patients.Significance.Outperforming all previous analytical pencil beam and deep learning based approaches, DoTA represents a new state of the art in data-driven dose calculation and can directly compete with the speed of even commercial GPU MC approaches. Providing the sub-second speed required for adaptive treatments, straightforward implementations could offer similar benefits to other steps of the radiotherapy workflow or other modalities such as helium or carbon treatments.

A feasibility study for in vivo treatment verification of IMRT using Monte Carlo dose calculation and deep learning-based modelling of EPID detector response

Background

This paper describes the development of a predicted electronic portal imaging device (EPID) transmission image (TI) using Monte Carlo (MC) and deep learning (DL). The measured and predicted TI were compared for two-dimensional in vivo radiotherapy treatment verification.

Methods

The plan CT was pre-processed and combined with solid water and then imported into PRIMO. The MC method was used to calculate the dose distribution of the combined CT. The U-net neural network-based deep learning model was trained to predict EPID TI based on the dose distribution of solid water calculated by PRIMO. The predicted TI was compared with the measured TI for two-dimensional in vivo treatment verification.

Results

The EPID TI of 1500 IMRT fields were acquired, among which 1200, 150, and 150 fields were used as the training set, the validation set, and the test set, respectively. A comparison of the predicted and measured TI was carried out using global gamma analyses of 3%/3 mm and 2%/2 mm (5% threshold) to validate the model's accuracy. The gamma pass rates were greater than 96.7% and 92.3%, and the mean gamma values were 0.21 and 0.32, respectively.

Conclusions

Our method facilitates the modelling process more easily and increases the calculation accuracy when using the MC algorithm to simulate the EPID response, and has potential to be used for in vivo treatment verification in the clinic.

A portable magnet for radiation biology and dosimetry studies in magnetic fields

BACKGROUND AND PURPOSE

In the current and rapidly evolving era of real-time MRI-guided radiotherapy, our radiation biology and dosimetry knowledge is being tested in a novel way. This paper presents the successful design and implementation of a portable device used to generate strong localized magnetic fields. These are ideally suited for small scale experiments that mimic the magnetic field environment inside an MRI-linac system, or more broadly MRI-guided particle therapy as well.

MATERIALS AND METHODS

A portable permanent magnet based device employing an adjustable steel yoke and magnetic field focusing cones has been designed, constructed and tested. The apparatus utilises two banks of Nd₂ Fe₁₄ B permanent magnets totalling around 50 kg in mass to generate a strong magnetic field throughout a small volume between two pole tips. The yoke design allows adjustment of the pole tip gap and exchanging of the focusing cones. Further to this, beam portal holes are present in the yoke and focusing cones, allowing for radiation beams of up to 5 x 5 cm² to pass through the region of high magnetic field between the focusing cone tips. Finite element magnetic modelling was performed to design and characterise the performance of the device. Automated physical measurements of the magnetic field components at various locations were measured to confirm the performance. The adjustable pole gap and interchangeable cones allows rapid changing of the experimental set-up to allow different styles of measurements to be performed.

RESULTS

A mostly uniform magnetic field of 1.2 T can be achieved over a volume of at least 3 x 3 x 3 cm³ . This can be reduced in strength to 0.3 T but increased in volume to 10 x 10 x 10 cm³ via removal of the cone tips and/or adjustment of the steel yoke. Although small, these volumes are sufficient to house radiation detectors, cell culture dishes and various phantom arrangements targeted at examining small radiation field dosimetry inside magnetic field strengths that can be changed with ease. Most important is the ability to align the magnetic field both perpendicular to, or inline with the radiation beam. To date, the system has been successfully used to conduct published research in the areas of radiation detector performance, lung phantom dosimetry, and how small clinical electron beams behave in these strong magnetic fields.

CONCLUSIONS

A portable, relatively inexpensive, and simple to operate device has successfully been constructed and used for performing radiation oncology studies around the theme of MRI-guided radiotherapy. This can be in either inline and perpendicular magnetic fields of up to 1.2 T with x-ray and particle beams.

Convolution neural network toward Monte Carlo photon dose calculation in radiation therapy

PURPOSE

The Monte Carlo (MC) algorithm has been widely accepted as the most accurate algorithm for dosimetric calculations under various conditions in radiotherapy. However, the calculation time remains an important obstacle hindering the routine use of MC in clinical settings. In this study, full MC three-dimensional dose distributions were obtained with the inputs of the total energy release per unit mass (TERMA) distributions and the electron density (ED) distributions using a convolutional neural network (CNN). A new Dose-mixup data augmentation routine and training strategy are proposed and applied in the training process. Attempts were made to reduce the calculation time while ensuring that the calculation accuracy is comparable to that of the MC.

METHODS

Datasets were generated via the MC with random rectangular field sizes, random iso-centers, and random gantry angles for head and neck computed tomography (CT) images with Mohan 6-MV spectrum photon beams. 1500 samples were generated for the training set, and 150 samples were generated for the validation set. The T-MC Net model was obtained with the Dose-mixup data augmentation routine. The new CTs were used to test the performance of the model in the rectangular fields and the intensity-modulated radiation therapy (IMRT) fields. The mean ± 95% confidence interval of gamma pass rates were calculated.

RESULTS

For 150 rectangular field test samples, the 1%/2 mm, 2%/2 mm, and 3%/2 mm criteria gamma pass rates were 90.11% ± 0.65%, 97.65% ± 0.31%, and 99.16% ± 0.19%, respectively. For the 100 IMRT field test samples, the 1%/2 mm, 2%/2 mm, and 3%/2 mm criteria gamma pass rates were 96.48% ± 0.28%, 99.14% ± 0.10%, and 99.63% ± 0.06%, respectively. For the 7-fields IMRT plan, the 1%/2 mm, 2%/2 mm, and 3%/2 mm criteria gamma pass rates were 97.06%, 99.10%, and 99.52%, respectively. For the 9-fields IMRT plan, the 1%/2 mm, 2%/2 mm, and 3%/2 mm criteria gamma pass rates were 98.16%, 99.61%, and 99.89%, respectively.

CONCLUSIONS

The feasibility of calculating dose distribution using a CNN with the TERMA three-dimensional distribution and ED distribution was established. The dosimetric results were comparable to those of the full MC. The accuracy and speed of the proposed approach make it a potential solution for full MC in radiotherapy. This method may be used as an acceleration engine for the dose algorithm and shows great potential for cases where fast dose calculations are needed.

Improving robustness of a deep learning-based lung-nodule classification model of CT images with respect to image noise

Objective. Robustness is an important aspect to consider, when developing methods for medical image analysis. This study investigated robustness properties of deep neural networks (DNNs) for a lung nodule classification problem based on CT images and proposed a solution to improve robustness.Approach. We firstly constructed a class of four DNNs with different widths, each predicting an output label (benign or malignant) for an input CT image cube containing a lung nodule. These networks were trained to achieve Area Under the Curve of 0.891-0.914 on a testing dataset. We then added to the input CT image cubes noise signals generated randomly using a realistic CT image noise model based on a noise power spectrum at 100 mAs, and monitored the DNNs output change. We definedSAR₅(%) to quantify the robustness of the trained DNN model, indicating that for 5% of CT image cubes, the noise can change the prediction results with a chance of at leastSAR₅(%). To understand robustness, we viewed the information processing pipeline by the DNN as a two-step process, with the first step using all but the last layers to extract representations of the input CT image cubes in a latent space, and the second step employing the last fully-connected layer as a linear classifier to determine the position of the sample representations relative to a decision plane. To improve robustness, we proposed to retrain the last layer of the DNN with a Supporting Vector Machine (SVM) hinge loss function to enforce the desired position of the decision plane.Main results.SAR₅ranged in 47.0%-62.0% in different DNNs. The unrobustness behavior may be ascribed to the unfavorable placement of the decision plane in the latent representation space, which made some samples be perturbed to across the decision plane and hence susceptible to noise. The DNN-SVM model improved robustness over the DNN model and reducedSAR₅by 8.8%-21.0%.Significance. This study provided insights about the potential reason for the unrobustness behavior of DNNs and the proposed DNN-SVM model improved model robustness.

Deep learning-based 3Din vivodose reconstruction with an electronic portal imaging device for magnetic resonance-linear accelerators: a proof of concept study

Objective.To develop a novel deep learning-based 3Din vivodose reconstruction framework with an electronic portal imaging device (EPID) for magnetic resonance-linear accelerators (MR-LINACs).Approach.The proposed method directly back-projected 2D portal dose into 3D patient coarse dose, which bypassed the complicated patient-to-EPID scatter estimation step used in conventional methods. A pre-trained convolutional neural network (CNN) was then employed to map the coarse dose to the final accurate dose. The electron return effect caused by the magnetic field was captured with the CNN model. Patient dose and portal dose datasets were synchronously generated with Monte Carlo simulation for 96 patients (78 cases for training and validation and 18 cases for testing) treated with fixed-beam intensity-modulated radiotherapy in four different tumor sites, including the brain, nasopharynx, lung, and rectum. Beam angles from the training dataset were further rotated 2-3 times, and doses were recalculated to augment the datasets.Results.The comparison between reconstructed doses and MC ground truth doses showed mean absolute errors <0.88% for all tumor sites. The averaged 3Dγ-passing rates (3%, 2 mm) were 97.42%±2.66% (brain), 98.53%±0.95% (nasopharynx), 99.41%±0.46% (lung), and 98.63%±1.01% (rectum). The dose volume histograms and indices also showed good consistency. The average dose reconstruction time, including back projection and CNN dose mapping, was less than 3 s for each individual beam.Significance.The proposed method can be potentially used for accurate and fast 3D dosimetric verification for online adaptive radiotherapy using MR-LINACs.

Comparison of Library of Plans with two daily adaptive strategies for whole bladder radiotherapy

Background and purpose

Whole bladder radiotherapy is challenging due to inter- and intrafraction size and shape changes. To account for these changes, currently a Library of Plans (LoP) technique is often applied, but daily adaptive radiotherapy is also increasingly becoming available. The aim of this study was to compare LoP with two magnetic resonance imaging guided radiotherapy (MRgRT) strategies by comparing target coverage and volume of healthy tissue inside the planning target volume (PTV) for whole bladder treatments.

Methods and materials

Data from 25 MRgRT lymph node oligometastases treatments (125 fractions) were used, with three MRI scans acquired at each fraction at 0, 15 and 30 min. Bladders were delineated and used to evaluate three strategies: 1) LoP with two plans for a 15 min fraction, 2) MRgRT_15min for a 15 min fraction and 3) MRgRT_30min for a 30 min fraction. The volumes of healthy tissue inside and bladder outside the PTV were analyzed on the simulated post-treatment images.

Results

MRgRT_30min had 120% and 121% more healthy tissue inside the PTV than LoP and MRgRT_15min. For LoP slightly more target outside the PTV was found than for MRgRT_30min and MRgRT_15min, with median 0% (range 0-23%) compared to 0% (0-20%) and 0% (0-10%), respectively.

Conclusions

Taking into account both target coverage and volume of healthy tissue inside the PTV, MRgRT_15min performed better than LoP and MRgRT_30min for whole bladder treatments. A 15 min daily adaptive radiotherapy workflow is needed to potentially benefit from replanning compared to LoP.

Automatic 3D Monte-Carlo-based secondary dose calculation for online verification of 1.5 T magnetic resonance imaging guided radiotherapy

•

First implementation of an independent 3D-secondary dose calculation (3D-SDC).

•

Validation of the 3D-SDC solution using patient plans and experimental plan QA.

•

Online SDC of central targets is feasible with a median calculation time of 1:23 min.

•

Peripheral targets with small beam numbers need alternative validation strategies.

Background and purpose

Hybrid magnetic resonance linear accelerator (MR-Linac) systems represent a novel technology for online adaptive radiotherapy. 3D secondary dose calculation (SDC) of online adapted plans is required to assure patient safety. Currently, no 3D-SDC solution is available for 1.5T MR-Linac systems. Therefore, the aim of this project was to develop and validate a method for online automatic 3D-SDC for adaptive MR-Linac treatments.

Materials and methods

An accelerator head model was designed for an 1.5T MR-Linac system, neglecting the magnetic field. The use of this model for online 3D-SDC of MR-Linac plans was validated in a three-step process: (1) comparison to measured beam data, (2) investigation of performance and limitations in a planning phantom and (3) clinical validation using n = 100 patient plans from different tumor entities, comparing the developed 3D-SDC with experimental plan QA.

Results

The developed model showed median gamma passing rates compared to MR-Linac base data of 84.7%, 100% and 99.1% for crossplane, inplane and depth-dose-profiles, respectively. Comparison of 3D-SDC and full dose calculation in a planning phantom revealed that with ⩾5 beams gamma passing rates >95% can be achieved for central target locations. With a median calculation time of 1:23 min, 3D-SDC of online adapted clinical MR-Linac plans demonstrated a median gamma passing rate of 98.9% compared to full dose calculation, whereas experimental plan QA reached 99.5%.

Conclusion

Here, we describe the first technical 3D-SDC solution for online adaptive MR-guided radiotherapy. For clinical situations with peripheral targets and a small number of beams additional verification appears necessary. Further improvement may include 3D-SDC with consideration of the magnetic field.

Prediction of Three-Dimensional Radiotherapy Optimal Dose Distributions for Lung Cancer Patients With Asymmetric Network

The iterative design of radiotherapy treatment plans is time-consuming and labor-intensive. In order to provide a guidance to treatment planning, Asymmetric network (A-Net) is proposed to predict the optimal 3D dose distribution for lung cancer patients. A-Net was trained and tested in 392 lung cancer cases with the prescription doses of 50Gy and 60Gy. In A-Net, the encoder and decoder are asymmetric, able to preserve input information and to adapt the limitation of GPU memory. Squeeze and excitation (SE) units are used to improve the data-fitting ability. A loss function involving both the dose distribution and prescription dose as ground truth are designed. In the experiment, A-Net is separately trained and tested in the 50Gy and 60Gy dataset and most of the metrics A-Net achieve similar performance as HD-Unet and 3D-Unet, and some metrics slightly better. In the 50Gy-and-60Gy-combined dataset, most of the A-Net's metrics perform better than the other two. In conclusion, A-Net can accurately predict the IMRT dose distribution in the three datasets of 50Gy and 50Gy-and-60Gy-combined dataset.

Long short-term memory networks for proton dose calculation in highly heterogeneous tissues

PURPOSE

To investigate the feasibility and accuracy of proton dose calculations with artificial neural networks (ANNs) in challenging three-dimensional (3D) anatomies.

METHODS

A novel proton dose calculation approach was designed based on the application of a long short-term memory (LSTM) network. It processes the 3D geometry as a sequence of two-dimensional (2D) computed tomography slices and outputs a corresponding sequence of 2D slices that forms the 3D dose distribution. The general accuracy of the approach is investigated in comparison to Monte Carlo reference simulations and pencil beam dose calculations. We consider both artificial phantom geometries and clinically realistic lung cases for three different pencil beam energies.

RESULTS

For artificial phantom cases, the trained LSTM model achieved a 98.57% γ-index pass rate ([1%, 3 mm]) in comparison to MC simulations for a pencil beam with initial energy 104.25 MeV. For a lung patient case, we observe pass rates of 98.56%, 97.74%, and 94.51% for an initial energy of 67.85, 104.25, and 134.68 MeV, respectively. Applying the LSTM dose calculation on patient cases that were fully excluded from the training process yields an average γ-index pass rate of 97.85%.

CONCLUSIONS

LSTM networks are well suited for proton dose calculation tasks. Further research, especially regarding model generalization and computational performance in comparison to established dose calculation methods, is warranted.

DeepDose: a robust deep learning-based dose engine for abdominal tumours in a 1.5 T MRI radiotherapy system

We present a robust deep learning-based framework for dose calculations of abdominal tumours in a 1.5 T MRI radiotherapy system. For a set of patient plans, a convolutional neural network is trained on the dose of individual multi-leaf-collimator segments following the DeepDose framework. It can then be used to predict the dose distribution per segment for a set of patient anatomies. The network was trained using data from three anatomical sites of the abdomen: prostate, rectal and oligometastatic tumours. A total of 216 patient fractions were used, previously treated in our clinic with fixed-beam IMRT using the Elekta MR-linac. For the purpose of training, 176 fractions were used with random gantry angles assigned to each segment, while 20 fractions were used for the validation of the network. The ground truth data were calculated with a Monte Carlo dose engine at 1% statistical uncertainty per segment. For a total of 20 independent abdominal test fractions with the clinical angles, the network was able to accurately predict the dose distributions, achieving 99.4% ± 0.6% for the whole plan prediction at the 3%/3 mm gamma test. The average dose difference and standard deviation per segment was 0.3% ± 0.7%. Additional dose prediction on one cervical and one pancreatic case yielded high dose agreement of 99.9% and 99.8% respectively for the 3%/3 mm criterion. Overall, we show that our deep learning-based dose engine calculates highly accurate dose distributions for a variety of abdominal tumour sites treated on the MR-linac, in terms of performance and generality.

Technical Challenges of Real-Time Adaptive MR-Guided Radiotherapy

In the past few years, radiotherapy (RT) has experienced a major technological innovation with the development of hybrid machines combining magnetic resonance (MR) imaging and linear accelerators. This new technology for MR-guided cancer treatment has the potential to revolutionize the field of adaptive RT due to the opportunity to provide high-resolution, real-time MR imaging before and during treatment application. However, from a technical point of view, several challenges remain which need to be tackled to ensure safe and robust real-time adaptive MR-guided RT delivery. In this manuscript, several technical challenges to MR-guided RT are discussed. Starting with magnetic field strength tradeoffs, the potential and limitations for purely MR-based RT workflows are discussed. Furthermore, the current status of real-time 3D MR imaging and its potential for real-time RT are summarized. Finally, the potential of quantitative MR imaging for future biological RT adaptation is highlighted.