Automatic IMRT planning via static field fluence prediction (AIP-SFFP): a deep learning algorithm for real-time prostate treatment planning

Deep reinforcement learning in radiation therapy planning optimization: A comprehensive review

PURPOSE

The formulation and optimization of radiation therapy plans are complex and time-consuming processes that heavily rely on the expertise of medical physicists. Consequently, there is an urgent need for automated optimization methods. Recent advancements in reinforcement learning, particularly deep reinforcement learning (DRL), show great promise for automating radiotherapy planning. This review summarizes the current state of DRL applications in this field, evaluates their effectiveness, and identifies challenges and future directions.

METHODS

A systematic search was conducted in Google Scholar, PubMed, IEEE Xplore, and Scopus using keywords such as "deep reinforcement learning", "radiation therapy", and "treatment planning". The extracted data were synthesized for an overview and critical analysis.

RESULTS

The application of deep reinforcement learning in radiation therapy plan optimization can generally be divided into three categories: optimizing treatment planning parameters, directly optimizing machine parameters, and adaptive radiotherapy. From the perspective of disease sites, DRL has been applied to cervical cancer, prostate cancer, vestibular schwannoma, and lung cancer. Regarding types of radiation therapy, it has been used in HDRBT, IMRT, SBRT, VMAT, GK, and Cyberknife.

CONCLUSIONS

Deep reinforcement learning technology has played a significant role in advancing the automated optimization of radiation therapy plans. However, there is still a considerable gap before it can be widely applied in clinical settings due to three main reasons: inefficiency, limited methods for quality assessment, and poor interpretability. To address these challenges, significant research opportunities exist in the future, such as constructing evaluators, parallelized training, and exploring continuous action spaces.

A joint ESTRO and AAPM guideline for development, clinical validation and reporting of artificial intelligence models in radiation therapy

BACKGROUND AND PURPOSE

Artificial Intelligence (AI) models in radiation therapy are being developed with increasing pace. Despite this, the radiation therapy community has not widely adopted these models in clinical practice. A cohesive guideline on how to develop, report and clinically validate AI algorithms might help bridge this gap.

METHODS AND MATERIALS

A Delphi process with all co-authors was followed to determine which topics should be addressed in this comprehensive guideline. Separate sections of the guideline, including Statements, were written by subgroups of the authors and discussed with the whole group at several meetings. Statements were formulated and scored as highly recommended or recommended.

RESULTS

The following topics were found most relevant: Decision making, image analysis, volume segmentation, treatment planning, patient specific quality assurance of treatment delivery, adaptive treatment, outcome prediction, training, validation and testing of AI model parameters, model availability for others to verify, model quality assurance/updates and upgrades, ethics. Key references were given together with an outlook on current hurdles and possibilities to overcome these. 19 Statements were formulated.

CONCLUSION

A cohesive guideline has been written which addresses main topics regarding AI in radiation therapy. It will help to guide development, as well as transparent and consistent reporting and validation of new AI tools and facilitate adoption.

Automatic IMRT treatment planning through fluence prediction and plan fine-tuning for nasopharyngeal carcinoma

BACKGROUND

At present, the implementation of intensity-modulated radiation therapy (IMRT) treatment planning for geometrically complex nasopharyngeal carcinoma (NPC) through manual trial-and-error fashion presents challenges to the improvement of planning efficiency and the obtaining of high-consistency plan quality. This paper aims to propose an automatic IMRT plan generation method through fluence prediction and further plan fine-tuning for patients with NPC and evaluates the planning efficiency and plan quality.

METHODS

A total of 38 patients with NPC treated with nine-beam IMRT were enrolled in this study and automatically re-planned with the proposed method. A trained deep learning model was employed to generate static field fluence maps for each patient with 3D computed tomography images and structure contours as input. Automatic IMRT treatment planning was achieved by using its generated dose with slight tightening for further plan fine-tuning. Lastly, the plan quality was compared between automatic plans and clinical plans.

RESULTS

The average time for automatic plan generation was less than 4 min, including fluence maps prediction with a python script and automated plan tuning with a C# script. Compared with clinical plans, automatic plans showed better conformity and homogeneity for planning target volumes (PTVs) except for the conformity of PTV-1. Meanwhile, the dosimetric metrics for most organs at risk (OARs) were ameliorated in the automatic plan, especially Dmax of the brainstem and spinal cord, and Dmean of the left and right parotid glands significantly decreased (P < 0.05).

CONCLUSION

We have successfully implemented an automatic IMRT plan generation method for patients with NPC. This method shows high planning efficiency and comparable or superior plan quality than clinical plans. The qualitative results before and after the plan fine-tuning indicates that further optimization using dose objectives generated by predicted fluence maps is crucial to obtain high-quality automatic plans.

"Under the hood": artificial intelligence in personalized radiotherapy

This review presents and discusses the ways in which artificial intelligence (AI) tools currently intervene, or could potentially intervene in the future, to enhance the diverse tasks involved in the radiotherapy workflow. The radiotherapy framework is presented on 2 different levels for the personalization of the treatment, distinct in tasks and methodologies. The first level is the clinically well-established anatomy-based workflow, known as adaptive radiation therapy. The second level is referred to as biology-driven workflow, explored in the research literature and recently appearing in some preliminary clinical trials for personalized radiation treatments. A 2-fold role for AI is defined according to these 2 different levels. In the anatomy-based workflow, the role of AI is to streamline and improve the tasks in terms of time and variability reductions compared to conventional methodologies. The biology-driven workflow instead fully relies on AI, which introduces decision-making tools opening uncharted frontiers that were in the past deemed challenging to explore. These methodologies are referred to as radiomics and dosiomics, handling imaging and dosimetric information, or multiomics, when complemented by clinical and biological parameters (ie, biomarkers). The review explicitly highlights the methodologies that are currently incorporated into clinical practice or still in research, with the aim of presenting the AI's growing role in personalized radiotherapy.

Single patient learning for adaptive radiotherapy dose prediction

BACKGROUND

Throughout a patient's course of radiation therapy, maintaining accuracy of their initial treatment plan over time is challenging due to anatomical changes-for example, stemming from patient weight loss or tumor shrinkage. Online adaptation of their RT plan to these changes is crucial, but hindered by manual and time-consuming processes. While deep learning (DL) based solutions have shown promise in streamlining adaptive radiation therapy (ART) workflows, they often require large and extensive datasets to train population-based models.

PURPOSE

This study extends our prior research by introducing a minimalist approach to patient-specific adaptive dose prediction. In contrast to our prior method, which involved fine-tuning a pre-trained population model, this new method trains a model from scratch using only a patient's initial treatment data. This patient-specific dose predictor aims to enhance clinical accessibility, thereby empowering physicians and treatment planners to make more informed, quantitative decisions in ART. We hypothesize that patient-specific DL models will provide more accurate adaptive dose predictions for their respective patients compared to a population-based DL model.

METHODS

We selected 33 patients to train an adaptive population-based (AP) model. Ten additional patients were selected, and their respective initial RT data served as single samples for training patient-specific (PS) models. These 10 patients contained an additional 26 ART plans that were withheld as the test dataset to evaluate AP versus PS model dose prediction performance. We assessed model performance using Mean Absolute Percent Error (MAPE) by comparing predicted doses to the originally delivered ground truth doses. We used the Wilcoxon signed-rank test to determine statistically significant differences in terms of MAPE between the AP and PS model results across the test dataset. Furthermore, we calculated differences between predicted and ground truth mean doses for segmented structures and determined statistical significance in the differences for each of them.

RESULTS

The average MAPE across AP and PS model dose predictions was 5.759% and 4.069%, respectively. The Wilcoxon signed-rank test yielded two-tailed p-value = 2.9802 × 10 - 8 $2.9802\ \times \ {10}^{ - 8}$ , indicating that the MAPE differences between the AP and PS model dose predictions are statistically significant, and 95% confidence interval = [-2.1610, -1.0130], indicating 95% confidence that the MAPE difference between the AP and PS models for a population lies in this range. Out of 24 total segmented structures, the comparison of mean dose differences for 12 structures indicated statistical significance with two-tailed p-values < 0.05.

CONCLUSION

Our study demonstrates the potential of patient-specific deep learning models in application to ART. Notably, our method streamlines the training process by minimizing the size of the required training dataset, as only a single patient's initial treatment data is required. External institutions considering the implementation of such a technology could package such a model so that it only requires the upload of a reference treatment plan for model training and deployment. Our single patient learning strategy demonstrates promise in ART due to its minimal dataset requirement and its utility in personalization of cancer treatment.

Deep learning based MLC aperture and monitor unit prediction as a warm start for breast VMAT optimisation

Objective. Automated treatment planning today is focussed on non-exact, two-step procedures. Firstly, dose-volume histograms (DVHs) or 3D dose distributions are predicted from the patient anatomy. Secondly, these are converted in multi-leaf collimator (MLC) apertures and monitor units (MUs) using a generic optimisation to obtain the final treatment plan. In contrast, we present a method to predict volumetric modulated arc therapy (VMAT) MLC apertures and MUs directly from patient anatomy using deep learning. The predicted plan is then provided as initialisation to the optimiser for fine-tuning.Approach. 148 patients (training: 101; validation: 23; test: 24), treated for right breast cancer, are replanned to obtain a homogeneous database of 3-arc VMAT plans (PTVBreast: 45.57 Gy; PTVBoost: 55.86 Gy) according to the clinical protocol, using RapidPlanTMwith automatic optimisation and extended convergence mode (clinical workflow). Projections of the CT and contours are created along the beam's eye view of all control points and given as input to a U-net type convolutional neural networks (CNN). The output are the MLC aperture and MU for all control points, from which a DICOM RTplan is built. This is imported and further optimised in the treatment planning system using automatic optimisation without convergence mode, with clinical PTV objectives and organs-at-risk (OAR) objectives based on the DVHs calculated from the imported plan (CNN workflow).Main results. Mean dose differences between the clinical and CNN workflow over the test set are 0.2 ± 0.5 Gy atD95%and 0.6 ± 0.4 Gy atD0.035ccof PTVBreastand -0.4 ± 0.3 Gy atD95%and 0.7 ± 0.3 Gy atD0.035ccof PTVBoost. For the OAR, they are -0.2 ± 0.2 Gy forDmean,heartand 0.04 ± 0.8 Gy forDmean,ipsilateral lung. The mean computation time is 60 and 25 min respectively.Significance. VMAT optimisation can be initialised by MLC apertures and MUs, directly predicted from patient anatomy using a CNN, reducing planning time with more than half while maintaining clinically acceptable plans. This procedure puts the planner in a supervising role over an AI-based treatment planning workflow.

Generating deliverable DICOM RT treatment plans for prostate VMAT by predicting MLC motion sequences with an encoder-decoder network

BACKGROUND

Deep learning-based auto-planning is an active research field; however, for some tasks a treatment planning system (TPS) is still required.

PURPOSE

To introduce a deep learning-based model generating deliverable DICOM RT treatment plans that can be directly irradiated by a linear accelerator (LINAC). The model was based on an encoder-decoder network and can predict multileaf collimator (MLC) motion sequences for prostate VMAT radiotherapy.

METHODS

A total of 619 treatment plans from 460 patients treated for prostate cancer with single-arc VMAT were included in this study. An encoder-decoder network was trained using 465 clinical treatment plans and validated on 77 plans. The performance was analyzed on a separate test set of 77 treatment plans. Separate L1 losses were computed for the leaf and jaw positions as well as the monitor units, with the leaf loss being weighted by a factor of 100 before being added to the other losses. The generated treatment plans were recalculated in a treatment planning system and the dose-volume metrics and gamma passing rates were compared to the original dose.

RESULTS

All generated treatment plans showed good agreement with the original data, with an average gamma passing rate (3%/3 mm) of 91.9 ± 7.1%. However, the coverage of the PTVs. was slightly lower for the generated plans (D98%  = 92.9 ± 2.6%) in comparison to the original plans (D98%  = 95.7 ± 2.2%). There was no significant difference in mean dose to the bladder between the predicted and original plan (Dmean of 28.0 ± 13.5 vs. 28.1 ± 13.3% of prescribed dose) or rectum (Dmean of 42.3 ± 7.4 vs. 42.6 ± 7.5%). The maximum dose to bladder was only slightly higher in the predicted plans (D2% of 100.7 ± 5.3 vs. 99.8 ± 4.0%) and for the rectum it was even lower (D2% of 100.5 ± 3.7 vs. 100.1 ± 4.3).

CONCLUSIONS

The deep learning-based model could predict MLC motion sequences in prostate VMAT plans, eliminating the need for sequencing inside a TPS, thus revolutionizing autonomous treatment planning workflows. This research completes the loop in deep learning-based treatment planning processes, enabling more efficient workflows for real-time or online adaptive radiotherapy.

Simultaneous dose distribution and fluence prediction for nasopharyngeal carcinoma IMRT

BACKGROUND

Current intensity-modulated radiation therapy (IMRT) treatment planning is still a manual and time/resource consuming task, knowledge-based planning methods with appropriate predictions have been shown to enhance the plan quality consistency and improve planning efficiency. This study aims to develop a novel prediction framework to simultaneously predict dose distribution and fluence for nasopharyngeal carcinoma treated with IMRT, the predicted dose information and fluence can be used as the dose objectives and initial solution for an automatic IMRT plan optimization scheme, respectively.

METHODS

We proposed a shared encoder network to simultaneously generate dose distribution and fluence maps. The same inputs (three-dimensional contours and CT images) were used for both dose distribution and fluence prediction. The model was trained with datasets of 340 nasopharyngeal carcinoma patients (260 cases for training, 40 cases for validation, 40 cases for testing) treated with nine-beam IMRT. The predicted fluence was then imported back to treatment planning system to generate the final deliverable plan. Predicted fluence accuracy was quantitatively evaluated within projected planning target volumes in beams-eye-view with 5 mm margin. The comparison between predicted doses, predicted fluence generated doses and ground truth doses were also conducted inside patient body.

RESULTS

The proposed network successfully predicted similar dose distribution and fluence maps compared with ground truth. The quantitative evaluation showed that the pixel-based mean absolute error between predicted fluence and ground truth fluence was 0.53% ± 0.13%. The structural similarity index also showed high fluence similarity with values of 0.96 ± 0.02. Meanwhile, the difference in the clinical dose indices for most structures between predicted dose, predicted fluence generated dose and ground truth dose were less than 1 Gy. As a comparison, the predicted dose achieved better target dose coverage and dose hot spot than predicted fluence generated dose compared with ground truth dose.

CONCLUSION

We proposed an approach to predict 3D dose distribution and fluence maps simultaneously for nasopharyngeal carcinoma patients. Hence, the proposed method can be potentially integrated in a fast automatic plan generation scheme by using predicted dose as dose objectives and predicted fluence as a warm start.

Implementation and evaluation of an intelligent automatic treatment planning robot for prostate cancer stereotactic body radiation therapy

PURPOSE

We previously developed a virtual treatment planner (VTP), an artificial intelligence robot, operating a treatment planning system (TPS). Using deep reinforcement learning guided by human knowledge, we trained the VTP to autonomously adjust relevant parameters in treatment plan optimization, similar to a human planner, to generate high-quality plans for prostate cancer stereotactic body radiation therapy (SBRT). This study describes the clinical implementation and evaluation of VTP.

MATERIALS AND METHODS

We integrate VTP with Eclipse TPS using scripting Application Programming Interface. VTP observes dose-volume histograms of relevant structures, decides how to adjust dosimetric constraints, including doses, volumes, and weighting factors, and applies the adjustments to the TPS interface to launch the optimization engine. This process continues until a high-quality plan is achieved. We evaluated VTP's performance using the prostate SBRT case from the 2016 American Association of Medical Dosimetrist/Radiosurgery Society plan study with its plan scoring system, and compared to human-generated plans submitted to the challenge. Using the same scoring system, we also compared the plan quality of 36 prostate SBRT cases (20 planned with IMRT and 16 planned with VMAT) treated at our institution for both VTP and human-generated plans.

RESULTS

In the plan study case, VTP achieved a score of 142.1/150.0, ranking the third in the competition (median 134.6). For the clinical cases, VTP achieved 110.6 ± 6.5 for 20 IMRT plans and 126.2 ± 4.7 for 16 VMAT plans, similar to scores of human-generated plans with 110.4 ± 7.0 for IMRT plans and 125.4 ± 4.4 for VMAT plans. The workflow, plan quality and planning time of VTP were reviewed to be satisfactory by experienced physicists.

CONCLUSION

We successfully implemented VTP to operate a TPS for autonomous human-like treatment planning for prostate SBRT.

Online adaptive planning methods for intensity-modulated radiotherapy

Online adaptive radiation therapy aims at adapting a patient's treatment plan to their current anatomy to account for inter-fraction variations before daily treatment delivery. As this process needs to be accomplished while the patient is immobilized on the treatment couch, it requires time-efficient adaptive planning methods to generate a quality daily treatment plan rapidly. The conventional planning methods do not meet the time requirement of online adaptive radiation therapy because they often involve excessive human intervention, significantly prolonging the planning phase. This article reviews the planning strategies employed by current commercial online adaptive radiation therapy systems, research on online adaptive planning, and artificial intelligence's potential application to online adaptive planning.

Toward real-time automatic treatment planning (RTTP) with a one-step 3D fluence map prediction method and (nonorthogonal) convolution technique

PURPOSE

To establish and evaluate a (quasi) real-time automated treatment planning (RTTP) strategy utilizing a one-step full 3D fluence map prediction model based on a nonorthogonal convolution operation for rectal cancer radiotherapy.

METHODS

The RTTP approach directly extracts 3D projections from volumetric CT and anatomical data according to the beam incident direction. A 3D deep learning model with a nonorthogonal convolution operation was established that takes projections in cone beam space as input, extracts the features along and around the ray-trace path, and outputs a predicted fluence map (PFM) for each beam. The PFM is then converted to the MLC sequence with deliverable MUs to generate the final treatment plan. A total of 314 rectal adenocarcinoma patients with 2198 projection data samples were used in model training and validation. An extra 20 patients were used to test the feasibility of the RTTP method by comparing the plan quality, efficiency, deliverability performance, and physician blinded review results with the manual plans.

RESULTS

Overall, the RTTP plans met the clinical dose criteria for target coverage, conformity, homogeneity, and organ-at-risk dose sparing. Compared to manual plans, the RTTP plans showed increases in PTV D_1% by only 2.33% (p < 0.001) and a decrease in PTV D_99% by 0.45% (p < 0.05). The RTTP plans showed a dose increase in the bladder, with a V₅₀ of 14.01 ± 11.75% vs. 10.74 ± 8.51%, respectively, and no significant increases in the femoral head with the mean dose. The planning efficiency was improved in RTTP planning, with 39 s vs. 944 s in fluence map generation; the deliverability performance was saved by 1.91% (p < 0.001) in total MU. According to the blinded plan review by our physician, 55% of RTTP plans can be directly used in clinical radiotherapy treatment.

CONCLUSION

The quasi RTTP method improves the planning efficiency and deliverability performance while maintaining a plan quality close to that of the optimized manual plans in rectal radiotherapy.

Safety and potential increased risk of toxicity of radiotherapy combined immunotherapy strategy

Accumulating interest has emerged in exploring the toxicity profiles of the combination strategy of radiotherapy (RT) and immune checkpoint inhibitors (ICIs). Much remains unknown regarding safety and the potential increased risk of toxicity of a combined treatment. ICI prolongs survival but can induce immune-related adverse events as well. To increase awareness of adverse effect and support immediate and successful management, we go over the literature on the safety of RT combined immunotherapy strategy. Representative evidence relevant to RT combined with ICI in the brain, lung, head and neck, and pelvic malignance was reviewed respectively. Given radiation doses and fractionation, the irradiated volume, the timing of RT, and ICI would significantly affect the safety and efficiency of ICI+RT combination therapy, and no consensus had been reached about how to arrange RT delivery in the combined contexture, we went over the available literature and tried to address these challenges including the timing of RT, optimal dose and fractionations, RT target and target volume, and potential biomarkers to predict toxicity. We found even though RT+ICI combination therapy might augment toxicities, the majority of patients experienced grade 4 or 5 AE are relatively rare and no significant difference could be found between combination group and monotherapy group. Sometimes the acute toxicity with ICI is much less predictable and often life threatening and in some can give rise to permanent effects. Clinicians across disciplines should be aware of these uncommon lethal complications induced by ICI+RT. Early recognition is the key to successful treatment, reversibility of organ dysfunction, and in some cases even prevention of fatal outcome. If recognized early, managed properly, and no fatal AE occurs, the development of irAE indicates a good prognosis. It should be noted that nothing is known about potential late effects because very few studies have 5-year follow-up. The nature of irAE is the attack of activated immune cells on normal tissues. The nature of RT-induced AE is the DNA damage on normal tissue, which is related with the dose delivered and volume irradiated and the tolerance of surrounding normal tissues. The immune-modulating effect of SBRT may augment the damage on normal tissues. To maximize the antitumor immune response, 8-12 Gy/fraction is preferred when conducting RT. The available clinical evidence suggest RT of this dose/fractionated strategy combined with ICI have a tolerable AE profile, which need further validation by more clinical trials in the future. The combination strategy of RT with anti-PD1/PDL1 anti-body is supposed to be concurrent or RT followed by anti-PD1/PDL1 antibody. Although RT and ipilimumab combination sequence is controversial, ipilimumab prior to or concurrent with RT might be proper, which need more clinical validation. Under the concept of immunological dose painting, SBRT work as a trigger of immune response. It has been observed that SBRT of partially radiated tumors combined with ICI could induce similar tumor control compared with total tumor irradiation. The side effects of RT may be mitigated potentially due to the reduction of irradiated volume. The antitumor efficiency and safety profile of immunological RT dose painting+ICI deserve further investigation. Clinical predictive factors for irAE risk remain unclear, and more investigation deserves to be conducted about the irAE prediction.

Multimodal Deep Learning-Based Prognostication in Glioma Patients: A Systematic Review

Malignant brain tumors pose a substantial burden on morbidity and mortality. As clinical data collection improves, along with the capacity to analyze it, novel predictive clinical tools may improve prognosis prediction. Deep learning (DL) holds promise for integrating clinical data of various modalities. A systematic review of the DL-based prognostication of gliomas was performed using the Embase (Elsevier), PubMed MEDLINE (National library of Medicine), and Scopus (Elsevier) databases, in accordance with PRISMA guidelines. All included studies focused on the prognostication of gliomas, and predicted overall survival (13 studies, 81%), overall survival as well as genotype (2 studies, 12.5%), and response to immunotherapy (1 study, 6.2%). Multimodal analyses were varied, with 6 studies (37.5%) combining MRI with clinical data; 6 studies (37.5%) integrating MRI with histologic, clinical, and biomarker data; 3 studies (18.8%) combining MRI with genomic data; and 1 study (6.2%) combining histologic imaging with clinical data. Studies that compared multimodal models to unimodal-only models demonstrated improved predictive performance. The risk of bias was mixed, most commonly due to inconsistent methodological reporting. Overall, the use of multimodal data in DL assessments of gliomas leads to a more accurate overall survival prediction. However, due to data limitations and a lack of transparency in model and code reporting, the full extent of multimodal DL as a resource for brain tumor patients has not yet been realized.

Evaluating the plan quality of a general head-and-neck knowledge-based planning model versus separate unilateral/bilateral models

The implementation of knowledge-based planning (KBP) continues to grow in radiotherapy clinics. KBP guides radiation treatment design by generating clinically acceptable plans in a timely and resource-efficient manner. The role of multiple KBP models tailored for variations within a disease site remains undefined in part because of the substantial effort and number of training cases required to create a high-quality KBP model. In this study, our aim was to explore whether site-specific KBP models lead to clinically meaningful differences in plan quality for head-and-neck (HN) patients when compared to a general model. One KBP model was created from prior volumetric-modulated arc therapy (VMAT) cases that treated unilateral HN lymph nodes while another model was created from VMAT cases that treated bilateral HN nodes. Thirty cases from each model (60 cases total) were randomly selected to create a third, general model. These models were applied to 60 HN test cases - 30 unilateral and 30 bilateral - to generate 180 VMAT plans in Eclipse. Clinically relevant dose metrics were compared between models. Paired-sample t-tests were used for statistical analysis, with the threshold for statistical significance set a priori at 0.007, taking into consideration multiple hypothesis testing to avoid type I error. For unilateral test cases, the unilateral model-generated plans had significantly lower spinal cord maximum doses (12.1 Gy vs 19.3 Gy, p < 0.001) and oral cavity mean doses (20.8 Gy vs 23.0 Gy, p < 0.001), compared with the bilateral model-generated plans. The unilateral and general models generated comparable plans for unilateral HN test cases. For bilateral test cases, the bilateral model created plans had significantly lower brainstem maximum doses (10.8 Gy vs 12.2 Gy, p < 0.001) and parotid mean doses (24.0 Gy vs 25.5 Gy, p < 0.001) when compared to the unilateral model. Right parotid mean doses were lower for bilateral model plans compared to general model plans (23.8 Gy vs 24.4 Gy). The general model created plans with significantly lower brainstem maximum doses (10.3 Gy vs 10.8 Gy) and oral cavity mean doses (35.3 Gy vs 36.7 Gy) when compared with bilateral model-generated plans. The general model outperformed the bilateral model in several dose metrics but they were not deemed clinically significant. For both case sets, the unilateral and general model created plans had higher monitor units when compared to the bilateral model, likely due to more stringent constraint settings. All other dose metrics were comparable. This study demonstrates that a balanced general HN model created using carefully curated treatment plans can produce high quality plans comparable to dedicated unilateral and bilateral models.

Using Artificial Intelligence for Optimization of the Processes and Resource Utilization in Radiotherapy

The radiotherapy (RT) process from planning to treatment delivery is a multistep, complex operation involving numerous levels of human-machine interaction and requiring high precision. These steps are labor-intensive and time-consuming and require meticulous coordination between professionals with diverse expertise. We reviewed and summarized the current status and prospects of artificial intelligence and machine learning relevant to the various steps in RT treatment planning and delivery workflow specifically in low- and middle-income countries (LMICs). We also searched the PubMed database using the search terms (Artificial Intelligence OR Machine Learning OR Deep Learning OR Automation OR knowledge-based planning AND Radiotherapy) AND (list of Low- and Middle-Income Countries as defined by the World Bank at the time of writing this review). The search yielded a total of 90 results, of which results with first authors from the LMICs were chosen. The reference lists of retrieved articles were also reviewed to search for more studies. No language restrictions were imposed. A total of 20 research items with unique study objectives conducted with the aim of enhancing RT processes were examined in detail. Artificial intelligence and machine learning can improve the overall efficiency of RT processes by reducing human intervention, aiding decision making, and efficiently executing lengthy, repetitive tasks. This improvement could permit the radiation oncologist to redistribute resources and focus on responsibilities such as patient counseling, education, and research, especially in resource-constrained LMICs.

Input feature design and its impact on the performance of deep learning models for predicting fluence maps in intensity-modulated radiation therapy

Objective. Deep learning (DL) models for fluence map prediction (FMP) have great potential to reduce treatment planning time in intensity-modulated radiation therapy (IMRT) by avoiding the lengthy inverse optimization process. This study aims to improve the rigor of input feature design in a DL-FMP model by examining how different designs of input features influence model prediction performance.Approach. This study included 231 head-and-neck intensity-modulated radiation therapy patients. Three input feature designs were investigated. The first design (D1) assumed that information of all critical structures from all beam angles should be combined to predict fluence maps. The second design (D2) assumed that local anatomical information was sufficient for predicting radiation intensity of a beamlet at a respective beam angle. The third design (D3) assumed the need for both local anatomical information and inter-beam modulation to predict radiation intensity values of the beamlets that intersect at a voxel. For each input design, we tailored the DL model accordingly. All models were trained using the same set of ground truth plans (GT plans). The plans generated by DL models (DL plans) were analyzed using key dose-volume metrics. One-way ANOVA with multiple comparisons correction (Bonferroni method) was performed (significance level = 0.05).Main results. For PTV-related metrics, all DL plans had significantly higher maximum dose (p < 0.001), conformity index (p < 0.001), and heterogeneity index (p < 0.001) compared to GT plans, with D2 being the worst performer. Meanwhile, except for cord+5 mm (p < 0.001), DL plans of all designs resulted in OAR dose metrics that are comparable to those of GT plans.Significance. Local anatomical information contains most of the information that DL models need to predict fluence maps for clinically acceptable OAR sparing. Input features from beam angles are needed to achieve the best PTV coverage. These results provide valuable insights for further improvement of DL-FMP models and DL models in general.

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.

Fast VMAT planning for prostate radiotherapy: dosimetric validation of a deep learning-based initial segment generation method

Objective. To develop and evaluate a deep learning based fast volumetric modulated arc therapy (VMAT) plan generation method for prostate radiotherapy. Approach. A customized 3D U-Net was trained and validated to predict initial segments at 90 evenly distributed control points of an arc, linked to our research treatment planning system (TPS) for segment shape optimization (SSO) and segment weight optimization (SWO). For 27 test patients, the VMAT plans generated based on the deep learning prediction (VMATDL) were compared with VMAT plans generated with a previously validated automated treatment planning method (VMATref). For all test cases, the deep learning prediction accuracy, plan dosimetric quality, and the planning efficiency were quantified and analyzed. Main results. For all 27 test cases, the resulting plans were clinically acceptable. The V 95% for the PTV2 was greater than 99%, and the V 107% was below 0.2%. Statistically significant difference in target coverage was not observed between the VMATref and VMATDL plans (P = 0.3243 > 0.05). The dose sparing effect to the OARs between the two groups of plans was similar. Small differences were only observed for the Dmean of rectum and anus. Compared to the VMATref, the VMATDL reduced 29.3% of the optimization time on average. Significance. A fully automated VMAT plan generation method may result in significant improvement in prostate treatment planning efficiency. Due to the clinically acceptable dosimetric quality and high efficiency, it could potentially be used for clinical planning application and real-time adaptive therapy application after further validation.

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.

Automatic treatment planning for cervical cancer radiation therapy using direct three-dimensional patient anatomy match

Purpose

Current knowledge‐based planning methods for radiation therapy mainly use low‐dimensional features extracted from contoured structures to identify geometrically similar patients. Here, we propose a knowledge‐based treatment planning method where the anatomical similarity is quantified by the rigid registration of the three‐dimensional (3D) planning target volume (PTV) and organs at risks (OARs) between an incoming patient and database patients.

Methods

A database that contains PTV and OARs contours from 81 cervical cancer radiation therapy patients was established. To identify the anatomically similar patients, the PTV of the new patient was registered to each PTV in the database and the Dice similarity coefficients were calculated for the PTV, rectum, and bladder between the new patient and database patients. Then the top 20 patients in the PTV match and top 3 patients in the subsequent bladder or rectum match were selected. The best dose–volume histogram parameters from the top three patients were applied as the dose constraints to the automatic plan optimization. A fast Fourier transform algorithm was developed to accelerate the 3D PTV registration process run through the database. The entire treatment planning process was automated using in‐house customized Pinnacle scripts. The automatic plans were generated for 20 patients using leave‐one‐out scheme and were evaluated against the corresponding clinical plans.

Results

The automatic plans significantly reduced rectum and bladder V50Gy by 11.79% ± 5.2% (p < 0.01) and 2.85% ± 3.16% (p < 0.01), respectively. The dose parameters achieved for the PTV and other OARs were comparable to those in the clinical plans. The entire planning process, including both dose prediction and inverse optimization, costs about 6 min.

Conclusions

The direct 3D contour match method utilizes the full spatial information of the PTV and OARs of interest and provides an intuitive measurement for patient plan anatomy similarity. The proposed automatic planning method can generate plans with better quality and higher efficiency.

Artificial Intelligence in Radiation Therapy

Artificial intelligence (AI) has great potential to transform the clinical workflow of radiotherapy. Since the introduction of deep neural networks, many AI-based methods have been proposed to address challenges in different aspects of radiotherapy. Commercial vendors have started to release AI-based tools that can be readily integrated to the established clinical workflow. To show the recent progress in AI-aided radiotherapy, we have reviewed AI-based studies in five major aspects of radiotherapy including image reconstruction, image registration, image segmentation, image synthesis, and automatic treatment planning. In each section, we summarized and categorized the recently published methods, followed by a discussion of the challenges, concerns, and future development. Given the rapid development of AI-aided radiotherapy, the efficiency and effectiveness of radiotherapy in the future could be substantially improved through intelligent automation of various aspects of radiotherapy.

Assessing the robustness of artificial intelligence powered planning tools in radiotherapy clinical settings-a phantom simulation approach

Background

Artificial intelligence (AI) based radiotherapy treatment planning tools have gained interest in automating the treatment planning process. It is essential to understand their overall robustness in various clinical scenarios. This is an existing gap between many AI based tools and their actual clinical deployment. This study works to fill the gap for AI based treatment planning by investigating a clinical robustness assessment (CRA) tool for the AI based planning methods using a phantom simulation approach.

Methods

A cylindrical phantom was created in the treatment planning system (TPS) with the axial dimension of 30 cm by 18 cm. Key structures involved in pancreas stereotactic body radiation therapy (SBRT) including PTV25, PTV33, C-Loop, stomach, bowel and liver were created within the phantom. Several simulation scenarios were created to mimic multiple scenarios of anatomical changes, including displacement, expansion, rotation and combination of three. The goal of treatment planning was to deliver 25 Gy to PTV25 and 33 Gy to PTV33 in 5 fractions in simultaneous integral boost (SIB) manner while limiting luminal organ-at-risk (OAR) max dose to be under 29 Gy. A previously developed deep learning based AI treatment planning tool for pancreas SBRT was identified as the validation object. For each scenario, the anatomy information was fed into the AI tool and the final fluence map associated to the plan was generated, which was subsequently sent to TPS for leaf sequencing and dose calculation. The final auto plan's quality was analyzed against the treatment planning constraint. The final plans' quality was further analyzed to evaluate potential correlation with anatomical changes using the Manhattan plot.

Results

A total of 32 scenarios were simulated in this study. For all scenarios, the mean PTV25 V25Gy of the AI based auto plans was 96.7% while mean PTV33 V33Gy was 82.2%. Large variation (16.3%) in PTV33 V33Gy was observed due to anatomical variations, a.k.a. proximity of luminal structure to PTV33. Mean max dose was 28.55, 27.68 and 24.63 Gy for C-Loop, bowel and stomach, respectively. Using D_0.03cc as max dose surrogate, the value was 28.03, 27.12 and 23.84 Gy for C-Loop, bowel and stomach, respectively. Max dose constraint of 29 Gy was achieved for 81.3% cases for C-Loop and stomach, and 78.1% for bowel. Using D_0.03cc as max dose surrogate, the passing rate was 90.6% for C-Loop, and 81.3% for bowel and stomach. Manhattan plot revealed high correlation between the OAR over dose and the minimal distance between the PTV33 and OAR.

Conclusions

The results showed promising robustness of the pancreas SBRT AI tool, providing important evidence of its readiness for clinical implementation. The established workflow could guide the process of assuring clinical readiness of future AI based treatment planning tools.

Artificial intelligence applications in intensity modulated radiation treatment planning: an overview

Artificial intelligence (AI) refers to methods that improve and automate challenging human tasks by systematically capturing and applying relevant knowledge in these tasks. Over the past decades, a number of approaches have been developed to address different types and needs of system intelligence ranging from search strategies to knowledge representation and inference to robotic planning. In the context of radiation treatment planning, multiple AI approaches may be adopted to improve the planning quality and efficiency. For example, knowledge representation and inference methods may improve dose prescription by integrating and reasoning about the domain knowledge described in many clinical guidelines and clinical trials reports. In this review, we will focus on the most studied AI approach in intensity modulated radiation therapy (IMRT)/volumetric modulated arc therapy (VMAT)-machine learning (ML) and describe our recent efforts in applying ML to improve the quality, consistency, and efficiency of IMRT/VMAT planning. With the available high-quality data, we can build models to accurately predict critical variables for each step of the planning process and thus automate and improve its outcomes. Specific to the IMRT/VMAT planning process, we can build models for each of the four critical components in the process: dose-volume histogram (DVH), Dose, Fluence, and Human Planner. These models can be divided into two general groups. The first group focuses on encoding prior experience and knowledge through ML and more recently deep learning (DL) from prior clinical plans and using these models to predict the optimal DVH (DVH prediction model), or 3D dose distribution (dose prediction model), or fluence map (fluence map model). The goal of these models is to reduce or remove the trial-and-error process and guarantee consistently high-quality plans. The second group of models focuses on mimicking human planners' decision-making process (planning strategy model) during the iterative adjustments/guidance of the optimization engine. Each critical step of the IMRT/VMAT treatment planning process can be improved and automated by AI methods. As more training data becomes available and more sophisticated models are developed, we can expect that the AI methods in treatment planning will continue to improve accuracy, efficiency, and robustness.

Insights of an AI agent via analysis of prediction errors: a case study of fluence map prediction for radiation therapy planning

Purpose.We have previously reported an artificial intelligence (AI) agent that automatically generates intensity-modulated radiation therapy (IMRT) plans via fluence map prediction, by-passing inverse planning. This AI agent achieved clinically comparable quality for prostate cases, but its performance on head-and-neck patients leaves room for improvement. This study aims to collect insights of the deep-learning-based (DL-based) fluence map prediction model by systematically analyzing its prediction errors.Methods.From the modeling perspective, the DL model's output is the fluence maps of IMRT plans. However, from the clinical planning perspective, the plan quality evaluation should be based on the clinical dosimetric criteria such as dose-volume histograms. To account for the complex and non-intuitive relationships between fluence map prediction errors and the corresponding dose distribution changes, we propose a novel error analysis approach that systematically examines plan dosimetric changes that are induced by varying amounts of fluence prediction errors. We investigated four decomposition modes of model prediction errors. The two spatial domain decompositions are based on fluence intensity and fluence gradient. The two frequency domain decompositions are based on Fourier-space banded frequency rings and Fourier-space truncated low-frequency disks. The decomposed error was analyzed for its impact on the resulting plans' dosimetric metrics. The analysis was conducted on 15 test cases spared from the 200 training and 16 validation cases used to train the model.Results.Most planning target volume metrics were significantly correlated with most error decompositions. The Fourier space disk radii had the largest Spearman's coefficients. The low-frequency region within a disk of ∼20% Fourier space contained most of errors that impact overall plan quality.Conclusions.This study demonstrates the feasibility of using fluence map prediction error analysis to understand the AI agent's performance. Such insights will help fine-tune the DL models in architecture design and loss function selection.

Knowledge-based radiation treatment planning: A data-driven method survey

This paper surveys the data-driven dose prediction methods investigated for knowledge-based planning (KBP) in the last decade. These methods were classified into two major categories-traditional KBP methods and deep-learning (DL) methods-according to their techniques of utilizing previous knowledge. Traditional KBP methods include studies that require geometric or anatomical features to either find the best-matched case(s) from a repository of prior treatment plans or to build dose prediction models. DL methods include studies that train neural networks to make dose predictions. A comprehensive review of each category is presented, highlighting key features, methods, and their advancements over the years. We separated the cited works according to the framework and cancer site in each category. Finally, we briefly discuss the performance of both traditional KBP methods and DL methods, then discuss future trends of both data-driven KBP methods to dose prediction.

A geometry-guided deep learning technique for CBCT reconstruction

Purpose.Although deep learning (DL) technique has been successfully used for computed tomography (CT) reconstruction, its implementation on cone-beam CT (CBCT) reconstruction is extremely challenging due to memory limitations. In this study, a novel DL technique is developed to resolve the memory issue, and its feasibility is demonstrated for CBCT reconstruction from sparsely sampled projection data.Methods.The novel geometry-guided deep learning (GDL) technique is composed of a GDL reconstruction module and a post-processing module. The GDL reconstruction module learns and performs projection-to-image domain transformation by replacing the traditional single fully connected layer with an array of small fully connected layers in the network architecture based on the projection geometry. The DL post-processing module further improves image quality after reconstruction. We demonstrated the feasibility and advantage of the model by comparing ground truth CBCT with CBCT images reconstructed using (1) GDL reconstruction module only, (2) GDL reconstruction module with DL post-processing module, (3) Feldkamp, Davis, and Kress (FDK) only, (4) FDK with DL post-processing module, (5) ray-tracing only, and (6) ray-tracing with DL post-processing module. The differences are quantified by peak-signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and root-mean-square error (RMSE).Results.CBCT images reconstructed with GDL show improvements in quantitative scores of PSNR, SSIM, and RMSE. Reconstruction time per image for all reconstruction methods are comparable. Compared to current DL methods using large fully connected layers, the estimated memory requirement using GDL is four orders of magnitude less, making DL CBCT reconstruction feasible.Conclusion.With much lower memory requirement compared to other existing networks, the GDL technique is demonstrated to be the first DL technique that can rapidly and accurately reconstruct CBCT images from sparsely sampled data.

A hierarchical deep reinforcement learning framework for intelligent automatic treatment planning of prostate cancer intensity modulated radiation therapy

Purpose.We have previously proposed an intelligent automatic treatment planning (IATP) framework that builds a virtual treatment planner network (VTPN) to operate a treatment planning system (TPS) to generate high-quality radiation therapy (RT) treatment plans. While the potential of IATP in automating RT treatment planning has been demonstrated, its poor scalability caused by an almost linear growth of network size with the number of treatment planning parameters (TPPs) is a bottleneck, preventing its application in complicate, but clinically relevant treatment planning problems. The decision-making behavior of the trained network is hard to understand. Motivated by the decision-making process of a human planner, this study proposes a hierarchical IATP framework.Methods and materials.The hierarchical VTPN (HieVTPN) consists of three networks, i.e. Structure-Net, Parameter-Net, and Action-Net. When interacting with a TPS, the networks are employed in a sequential order in each step to decide the structure to adjust, the TPP to adjust for the selected structure, and the specific adjustment manner for the parameter, respectively. We developed an end-to-end hierarchical deep reinforcement learning scheme to simultaneously train the three networks. We then evaluated the effectiveness of the proposed framework in the treatment planning problems for prostate cancer intensity modulated RT (IMRT) and stereotactic body RT (SBRT). We benchmarked the performance of our approach by comparing plans made by VTPN of a parallel architecture, and the human plans submitted for competition in the 2016 American Association of Medical Dosimetrist (AAMD)/Radiosurgery Society (RSS) Plan Study. We analyzed scalability of the network size with respect to the number of TPPs. Numerical experiments were also performed to understand the rationale of the decision-making behaviors of the trained HieVTPN.Results.Both HieVTPNs for prostate IMRT and SBRT were trained successfully using 10 training patient cases and 5 validation cases. For IMRT, HieVTPN was able to generate high-quality plans for 59 testing patient cases that were not included in training process, achieving an average plan score of 8.62 (±0.83), with 9 being the maximal score. The score was comparable to that of the VTPN, 8.45 (±0.48). For SBRT planning, HieVTPN achieved an average plan score of 139.07 on five testing patient cases compared to the score of 132.21 averaged over the human plans summited for competition in AAMD/RSS plan study. Different from VTPN with network size linearly scaling with the number of TPPs, the network size of HieVTPN is almost independent of the number of TPPs. It was also observed that the decision-making behaviors of HieVTPN were understandable and generally agreed with the human experience.Conclusions.With the scalability and explainability, the hierarchical IATP framework is more favorable than the previous framework in terms of handling treatment planning problems involving a large number of TPPs.

Automation in radiotherapy treatment planning: Examples of use in clinical practice and future trends for a complete automated workflow

Modern radiotherapy treatment planning is a complex and time-consuming process that requires the skills of experienced users to obtain quality plans. Since the early 2000s, the automation of this planning process has become an important research topic in radiotherapy. Today, the first commercial automated treatment planning solutions are available and implemented in a growing number of clinical radiotherapy departments. It should be noted that these various commercial solutions are based on very different methods, implying a daily practice that varies from one center to another. It is likely that this change in planning practices is still in its infancy. Indeed, the rise of artificial intelligence methods, based in particular on deep learning, has recently revived research interest in this subject. The numerous articles currently being published announce a lasting and profound transformation of radiotherapy planning practices in the years to come. From this perspective, an evolution of initial training for clinical teams and the drafting of new quality assurance recommendations is desirable.

An artificial intelligence-driven agent for real-time head-and-neck IMRT plan generation using conditional generative adversarial network (cGAN)

PURPOSE

To develop an artificial intelligence (AI) agent for fully automated rapid head-and-neck intensity-modulated radiation therapy (IMRT) plan generation without time-consuming dose-volume-based inverse planning.

METHODS

This AI agent was trained via implementing a conditional generative adversarial network (cGAN) architecture. The generator, PyraNet, is a novel deep learning network that implements 28 classic ResNet blocks in pyramid-like concatenations. The discriminator is a customized four-layer DenseNet. The AI agent first generates multiple customized two-dimensional projections at nine template beam angles from a patient's three-dimensional computed tomography (CT) volume and structures. These projections are then stacked as four-dimensional inputs of PyraNet, from which nine radiation fluence maps of the corresponding template beam angles are generated simultaneously. Finally, the predicted fluence maps are automatically postprocessed by Gaussian deconvolution operations and imported into a commercial treatment planning system (TPS) for plan integrity check and visualization. The AI agent was built and tested upon 231 oropharyngeal IMRT plans from a TPS plan library. 200/16/15 plans were assigned for training/validation/testing, respectively. Only the primary plans in the sequential boost regime were studied. All plans were normalized to 44 Gy prescription (2 Gy/fx). A customized Harr wavelet loss was adopted for fluence map comparison during the training of the PyraNet. For test cases, isodose distributions in AI plans and TPS plans were qualitatively evaluated for overall dose distributions. Key dosimetric metrics were compared by Wilcoxon signed-rank tests with a significance level of 0.05.

RESULTS

All 15 AI plans were successfully generated. Isodose gradients outside of PTV in AI plans were comparable to those of the TPS plans. After PTV coverage normalization, D_mean of left parotid (D_AI  = 23.1 ± 2.4 Gy; D_TPS  = 23.1 ± 2.0 Gy), right parotid (D_AI  = 23.8 ± 3.0 Gy; D_TPS  = 23.9 ± 2.3 Gy), and oral cavity (D_AI  = 24.7 ± 6.0 Gy; D_TPS  = 23.9 ± 4.3 Gy) in the AI plans and the TPS plans were comparable without statistical significance. AI plans achieved comparable results for maximum dose at 0.01cc of brainstem (D_AI  = 15.0 ± 2.1 Gy; D_TPS  = 15.5 ± 2.7 Gy) and cord + 5mm (D_AI  = 27.5 ± 2.3 Gy; D_TPS  = 25.8 ± 1.9 Gy) without clinically relevant differences, but body D_max results (D_AI  = 121.1 ± 3.9 Gy; D_TPS  = 109.0 ± 0.9 Gy) were higher than the TPS plan results. The AI agent needed ~3 s for predicting fluence maps of an IMRT plan.

CONCLUSIONS

With rapid and fully automated execution, the developed AI agent can generate complex head-and-neck IMRT plans with acceptable dosimetry quality. This approach holds great potential for clinical applications in preplanning decision-making and real-time planning.

DVHnet: A deep learning-based prediction of patient-specific dose volume histograms for radiotherapy planning

PURPOSE

To develop a deep learning method to predict patient-specific dose volume histograms (DVHs) for radiotherapy planning.

METHODS

Patient data included 180 cases with nasopharyngeal cancer, of which 153 cases were used for training and 27 for testing. A network (named "DVHnet") based on a convolutional neural network (CNN) was designed for directly predicting DVHs of organs at risk (OARs). Two-channel images with contoured structures were generated as the inputs for training the model. A one-dimensional array consisting of 256 continuous volume percentages on a DVH curve for each slice was calculated as the corresponding output. The combined DVH was then calculated. Sixteen OARs were modeled in the study. Prediction accuracy was evaluated against the corresponding DVH curve of ground truth (GT) plans. A global DVH analysis and critical dosimetry metrics for each OAR were calculated for quantitative evaluation. The performance of DVHnet also was evaluated against two baselines: DosemapNet (developed by our research group) and commercial RapidPlan software.

RESULTS

The predicted mean difference in average dose of all OARs using DVHnet was 0.30 ± 0.95 Gy. And the predicted differences in D2% and D50 can be control within 2.32 and 0.69 Gy. For most OARs, there were no obvious differences between the dosimetric metrics of the predicted and GT values for both DVHnet and DosemapNet (P ≥ 0.05). Only the predicted D2% of the optic organs for DVHnet, and of brain stem PRV for DosemapNet displayed statistically significant differences. Except for the optic organs, DVHnet performs better than or comparably with RapidPlan. The mean difference in proportion of points of interest was 3.59% ± 7.78%.

CONCLUSIONS

A deep learning network model was developed to automatically extract useful features for accurate prediction of patient-specific DVH curves directly. The performance of DVHnet was comparable to DosemapNet and RapidPlan.

A Review on Application of Deep Learning Algorithms in External Beam Radiotherapy Automated Treatment Planning

Treatment planning plays an important role in the process of radiotherapy (RT). The quality of the treatment plan directly and significantly affects patient treatment outcomes. In the past decades, technological advances in computer and software have promoted the development of RT treatment planning systems with sophisticated dose calculation and optimization algorithms. Treatment planners now have greater flexibility in designing highly complex RT treatment plans in order to mitigate the damage to healthy tissues better while maximizing radiation dose to tumor targets. Nevertheless, treatment planning is still largely a time-inefficient and labor-intensive process in current clinical practice. Artificial intelligence, including machine learning (ML) and deep learning (DL), has been recently used to automate RT treatment planning and has gained enormous attention in the RT community due to its great promises in improving treatment planning quality and efficiency. In this article, we reviewed the historical advancement, strengths, and weaknesses of various DL-based automated RT treatment planning techniques. We have also discussed the challenges, issues, and potential research directions of DL-based automated RT treatment planning techniques.