Automated Robust Proton Planning Using Dose-Volume Histogram-Based Mimicking of the Photon Reference Dose and Reducing Organ at Risk Dose Optimization

The partial adaptation strategy for online-adaptive proton therapy: A proof of concept study in head and neck cancer patients

BACKGROUND

The accuracy of intensity-modulated proton therapy (IMPT) is greatly affected by anatomy variations that might occur during the treatment course. Online plan adaptations have been proposed as a solution to intervene promptly during a treatment session once the anatomy changes are detected. The implementation of online-adaptive proton therapy (OAPT) is still hindered by time-consuming tasks in the workflow.

PURPOSE

The study introduces the novel concept of partial adaptation and aims at investigating its feasibility as a potential solution to parallelize tasks during an OAPT workflow for saving valuable in-room time.

METHODS

The proof-of-principle simulation study includes datasets from six head and neck cancer (HNC) patients, each consisting of one planning CT (pCT) and three contoured control CTs (cCTs). Robust 3-field normo-fractionated initial IMPT plans were generated on the pCTs with a standardized field configuration, delivering 66 Gy and 54 Gy to the high-risk and low-risk clinical target volume (CTVHigh and CTVLow), respectively. For each cCT, a dose-mimicking-based partial adaptation was applied: two fields were adapted on the current anatomy taking into account the background dose of the first non-adapted field supposedly delivered in the meantime. Fraction doses on the cCTs resulting from partially adapted plans with different first (non-adapted) field assignments were compared against those from non-adapted and fully adapted plans regarding target coverage and organs at risk (OARs) sparing. The robustness of partially adapted plans was also evaluated.

RESULTS

Partially adapted plans showed comparable results to fully adapted plans and were superior to non-adapted plans for both target coverage and OAR sparing. Target coverage degradation in the non-adapted plans (median D98%: 95.9% and 97.5% for CTVLow and CTVHigh, respectively) was recovered by both partial (98.0% and 98.5%) and full adaptation (98.2% and 98.7%) in comparison to the initial plans (98.7% and 98.8%). The initial hotspot dose for the CTVHigh (median D2%: 101.8%) increased in the non-adapted plans (102.9%) and was recovered by the adaptive strategies (partial: 102.5%, full: 101.9%). The near-maximum dose (D0.01cc) to brainstem and spinal cord was within clinical constraints for all investigated dose distributions, but clearly increased for no adaptation and improved in the (both partially and fully) adapted plans with respect to the non-adapted ones. The parotids' median doses (D50) were mainly patient-specific depending on the proximity to the target region, but anyway lower for the partially and fully adapted plans compared to the non-adapted ones. OAR sparing was furthermore improved for the partially adapted plans in comparison to full adaptation. Robustness of the target dose metrics was preserved in all evaluated scenarios.

CONCLUSIONS

For OAPT of HNC patients, partial adaptation is able to generate plans of superior conformity to non-adapted plans and of comparable conformity as fully adapted plans, while having the potential to speed up the online-adaptive workflows. Thus, partial adaptation represents an intermediate approach until fast online adaptation workflows become available. Furthermore, it can be applied in workflows where online treatment verification stops the delivery and triggers an online adaptation for the remaining fraction.

Evaluation of an automated clinical decision system with deep learning dose prediction and NTCP model for prostate cancer proton therapy

Objective.To evaluate the feasibility of using a deep learning dose prediction approach to identify patients who could benefit most from proton therapy based on the normal tissue complication probability (NTCP) model.Approach.Two 3D UNets were established to predict photon and proton doses. A dataset of 95 patients with localized prostate cancer was randomly partitioned into 55, 10, and 30 for training, validation, and testing, respectively. We selected NTCP models for late rectum bleeding and acute urinary urgency of grade 2 or higher to quantify the benefit of proton therapy. Propagated uncertainties of predicted ΔNTCPs resulting from the dose prediction errors were calculated. Patient selection accuracies for a single endpoint and a composite evaluation were assessed under different ΔNTCP thresholds.Main results.Our deep learning-based dose prediction technique can reduce the time spent on plan comparison from approximately 2 days to as little as 5 seconds. The expanded uncertainty of predicted ΔNTCPs for rectum and bladder endpoints propagated from the dose prediction error were 0.0042 and 0.0016, respectively, which is less than one-third of the acceptable tolerance. The averaged selection accuracies for rectum bleeding, urinary urgency, and composite evaluation were 90%, 93.5%, and 93.5%, respectively.Significance.Our study demonstrates that deep learning dose prediction and NTCP evaluation scheme could distinguish the NTCP differences between photon and proton treatment modalities. In addition, the dose prediction uncertainty does not significantly influence the decision accuracy of NTCP-based patient selection for proton therapy. Therefore, automated deep learning dose prediction and NTCP evaluation schemes can potentially be used to screen large patient populations and to avoid unnecessary delays in the start of prostate cancer radiotherapy in the future.

SISS-MCO: large scale sparsity-induced spot selection for fast and fully-automated robust multi-criteria optimisation of proton plans

Objective.Intensity modulated proton therapy (IMPT) is an emerging treatment modality for cancer. However, treatment planning for IMPT is labour-intensive and time-consuming. We have developed a novel approach for multi-criteria optimisation (MCO) of robust IMPT plans (SISS-MCO) that is fully automated and fast, and we compare it for head and neck, cervix, and prostate tumours to a previously published method for automated robust MCO (IPBR-MCO, van de Water 2013).Approach.In both auto-planning approaches, the applied automated MCO of spot weights was performed with wish-list driven prioritised optimisation (Breedveld 2012). In SISS-MCO, spot weight MCO was applied once for every patient after sparsity-induced spot selection (SISS) for pre-selection of the most relevant spots from a large input set of candidate spots. IPBR-MCO had several iterations of spot re-sampling, each followed by MCO of the weights of the current spots.Main results.Compared to the published IPBR-MCO, the novel SISS-MCO resulted in similar or slightly superior plan quality. Optimisation times were reduced by a factor of 6 i.e. from 287 to 47 min. Numbers of spots and energy layers in the final plans were similar.Significance.The novel SISS-MCO automatically generated high-quality robust IMPT plans. Compared to a published algorithm for automated robust IMPT planning, optimisation times were reduced on average by a factor of 6. Moreover, SISS-MCO is a large scale approach; this enables optimisation of more complex wish-lists, and novel research opportunities in proton therapy.

Proton Therapy in Breast Cancer: A Review of Potential Approaches for Patient Selection

Proton radiotherapy may be a compelling technical option for the treatment of breast cancer due to its unique physical property known as the "Bragg peak." This feature offers distinct advantages, promising superior dose conformity within the tumor area and reduced radiation exposure to surrounding healthy tissues, enhancing the potential for better treatment outcomes. However, proton therapy is accompanied by inherent challenges, primarily higher costs and limited accessibility when compared to well-developed photon irradiation. Thus, in clinical practice, it is important for radiation oncologists to carefully select patients before recommendation of proton therapy to ensure the transformation of dosimetric benefits into tangible clinical benefits. Yet, the optimal indications for proton therapy in breast cancer patients remain uncertain. While there is no widely recognized methodology for patient selection, numerous attempts have been made in this direction. In this review, we intended to present an inspiring summarization and discussion about the current practices and exploration on the approaches of this treatment decision-making process in terms of treatment-related side-effects, tumor control, and cost-efficiency, including the normal tissue complication probability (NTCP) model, the tumor control probability (TCP) model, genomic biomarkers, cost-effectiveness analyses (CEAs), and so on. Additionally, we conducted an evaluation of the eligibility criteria in ongoing randomized controlled trials and analyzed their reference value in patient selection. We evaluated the pros and cons of various potential patient selection approaches and proposed possible directions for further optimization and exploration. In summary, while proton therapy holds significant promise in breast cancer treatment, its integration into clinical practice calls for a thoughtful, evidence-driven strategy. By continuously refining the patient selection criteria, we can harness the full potential of proton radiotherapy while ensuring maximum benefit for breast cancer patients.

Artificial general intelligence for radiation oncology

The emergence of artificial general intelligence (AGI) is transforming radiation oncology. As prominent vanguards of AGI, large language models (LLMs) such as GPT-4 and PaLM 2 can process extensive texts and large vision models (LVMs) such as the Segment Anything Model (SAM) can process extensive imaging data to enhance the efficiency and precision of radiation therapy. This paper explores full-spectrum applications of AGI across radiation oncology including initial consultation, simulation, treatment planning, treatment delivery, treatment verification, and patient follow-up. The fusion of vision data with LLMs also creates powerful multimodal models that elucidate nuanced clinical patterns. Together, AGI promises to catalyze a shift towards data-driven, personalized radiation therapy. However, these models should complement human expertise and care. This paper provides an overview of how AGI can transform radiation oncology to elevate the standard of patient care in radiation oncology, with the key insight being AGI's ability to exploit multimodal clinical data at scale.

Automated treatment planning for proton pencil beam scanning using deep learning dose prediction and dose-mimicking optimization

PURPOSE

The purpose of this study is to investigate the use of a deep learning architecture for automated treatment planning for proton pencil beam scanning (PBS).

METHODS

A 3-dimensional (3D) U-Net model has been implemented in a commercial treatment planning system (TPS) that uses contoured regions of interest (ROI) binary masks as model inputs with a predicted dose distribution as the model output. Predicted dose distributions were converted to deliverable PBS treatment plans using a voxel-wise robust dose mimicking optimization algorithm. This model was leveraged to generate machine learning (ML) optimized plans for patients receiving proton PBS irradiation of the chest wall. Model training was carried out on a retrospective set of 48 previously-treated chest wall patient treatment plans. Model evaluation was carried out by generating ML-optimized plans on a hold-out set of 12 contoured chest wall patient CT datasets from previously treated patients. Clinical goal criteria and gamma analysis were used to compare dose distributions of the ML-optimized plans against the clinically approved plans across the test patients.

RESULTS

Statistical analysis of mean clinical goal criteria indicates that compared to the clinical plans, the ML optimization workflow generated robust plans with similar dose to the heart, lungs, and esophagus while achieving superior dosimetric coverage to the PTV chest wall (clinical mean V95 = 97.6% vs. ML mean V95 = 99.1%, p < 0.001) across the 12 test patients.

CONCLUSIONS

ML-based automated treatment plan optimization using the 3D U-Net model can generate treatment plans of similar clinical quality compared to human-driven optimization.

Automated Robust Planning for IMPT in Oropharyngeal Cancer Patients Using Machine Learning

PURPOSE

The aim of this study was to evaluate an automated treatment planning method for robustly optimized intensity modulated proton therapy (IMPT) plans for oropharyngeal carcinoma patients and to compare the results with manually optimized robust IMPT plans.

METHODS AND MATERIALS

An atlas regression forest-based machine learning (ML) model for dose prediction was trained on CT scans, contours, and dose distributions of robust IMPT plans of 88 oropharyngeal cancer (OPC) patients. The ML model was combined with a robust voxel and dose volume histogram-based dose mimicking optimization algorithm for 21 perturbed scenarios to generate a machine-deliverable plan from the predicted dose distribution. Machine learning optimization (MLO) configuration was performed using a cross-validation approach with 3 × 8 tuning patients and comprised of adjustments to the mimicking optimization, to generate higher-quality MLO plans. Independent testing of the MLO algorithm was performed with another 25 patients. Plan quality of clinical and MLO plans was evaluated by the clinical target volume (D98% voxel-wise minimum dose >94%), organ at risk (OAR) doses, and the normal tissue complication probability (NTCP) (sum (Σ) of grade-2 and grade-3 dysphagia and xerostomia).

RESULTS

Adequate robust target coverage was achieved in 24/25 clinical plans and in 23/25 MLO plans in the primary clinical target volume (CTV). In the elective CTV, 22/25 clinical plans and 24/25 MLO plans passed the robust target coverage evaluation threshold. The MLO average Σgrade 2 and Σgrade 3 NTCPs were comparable to the clinical plans (Σgrade 2 NTCPs: clinical 47.5% vs MLO 48.4%, Σgrade 3 NTCPs: clinical 11.9% vs MLO 12.3%). Significant increases in OAR average doses in MLO plans were found in the pharynx constrictor muscles (163 cGy, P = .002) and cervical esophagus (265 cGy, P = .002). The MLO plans were created within 45 minutes.

CONCLUSION

This study showed that automated MLO planning can generate robustly optimized MLO plans with quality comparable to clinical plans in OPC patients.

OpenKBP-Opt: an international and reproducible evaluation of 76 knowledge-based planning pipelines

Objective.To establish an open framework for developing plan optimization models for knowledge-based planning (KBP).Approach.Our framework includes radiotherapy treatment data (i.e. reference plans) for 100 patients with head-and-neck cancer who were treated with intensity-modulated radiotherapy. That data also includes high-quality dose predictions from 19 KBP models that were developed by different research groups using out-of-sample data during the OpenKBP Grand Challenge. The dose predictions were input to four fluence-based dose mimicking models to form 76 unique KBP pipelines that generated 7600 plans (76 pipelines × 100 patients). The predictions and KBP-generated plans were compared to the reference plans via: the dose score, which is the average mean absolute voxel-by-voxel difference in dose; the deviation in dose-volume histogram (DVH) points; and the frequency of clinical planning criteria satisfaction. We also performed a theoretical investigation to justify our dose mimicking models.Main results.The range in rank order correlation of the dose score between predictions and their KBP pipelines was 0.50-0.62, which indicates that the quality of the predictions was generally positively correlated with the quality of the plans. Additionally, compared to the input predictions, the KBP-generated plans performed significantly better (P< 0.05; one-sided Wilcoxon test) on 18 of 23 DVH points. Similarly, each optimization model generated plans that satisfied a higher percentage of criteria than the reference plans, which satisfied 3.5% more criteria than the set of all dose predictions. Lastly, our theoretical investigation demonstrated that the dose mimicking models generated plans that are also optimal for an inverse planning model.Significance.This was the largest international effort to date for evaluating the combination of KBP prediction and optimization models. We found that the best performing models significantly outperformed the reference dose and dose predictions. In the interest of reproducibility, our data and code is freely available.

Robust automated radiation therapy treatment planning using scenario-specific dose prediction and robust dose mimicking

Purpose

We present a framework for robust automated treatment planning using machine learning, comprising scenario‐specific dose prediction and robust dose mimicking.

Methods

The scenario dose prediction pipeline is divided into the prediction of nominal dose from input image and the prediction of scenario dose from nominal dose, each using a deep learning model with U‐net architecture. By using a specially developed dose–volume histogram–based loss function, the predicted scenario doses are ensured sufficient target coverage despite the possibility of the training data being non‐robust. Deliverable plans may then be created by solving a robust dose mimicking problem with the predictions as scenario‐specific reference doses.

Results

Numerical experiments are performed using a data set of 52 intensity‐modulated proton therapy plans for prostate patients. We show that the predicted scenario doses resemble their respective ground truth well, in particular while having target coverage comparable to that of the nominal scenario. The deliverable plans produced by the subsequent robust dose mimicking were showed to be robust against the same scenario set considered for prediction.

Conclusions

We demonstrate the feasibility and merits of the proposed methodology for incorporating robustness into automated treatment planning algorithms.

When your MR linac is down: Can an automated pipeline bail you out of trouble?

PURPOSE

The unique treatment delivery technique provided by magnetic resonance guided radiotherapy (MRgRT) can represent a significant drawback when system fail occurs. This retrospective study proposes and evaluates a pipeline to completely automate the workflow necessary to shift a MRgRT treatment to a traditional radiotherapy linac.

MATERIAL AND METHODS

Patients undergoing treatment during the last MRgRT system failure were retrospectively included in this study. The core of the proposed pipeline was based on a tool able to mimic the original MR linac dose distribution. The so obtained dose distribution (AUTO) has been compared with the distribution obtained in the conventional radiotherapy linac (MAN). Plan comparison has been performed in terms of time required to obtain the final dose distribution, DVH parameters, dosimetric indices and visual analogue scales scoring by radiation oncologists.

RESULTS

AUTO plans generation has been obtained within 10 min for all the considered cases. All AUTO plans were found to be within clinical tolerance, showing a mean target coverage variation of 1.7% with a maximum value of 4.3% and a minimum of 0.6% when compared with MAN plans. The highest OARs mean variation has been found for rectum V60 (6.7%). Dosimetric indices showed no relevant differences, with smaller gradient measure in favour of AUTO plans. Visual analogue scales scoring has confirmed comparable plan quality for AUTO plans.

CONCLUSION

The proposed workflow allows a fast and accurate generation of automatic treatment plans. AUTO plans can be considered equivalent to MAN ones, with limited clinical impact in the worst-case scenario.

Automated robust SBPT planning using EUD-based prediction of SBRT plan for patients with lung cancer

PURPOSE

To evaluate the quality of robust stereotactic body proton therapy (RSBPT) plans generated by one-clicking scripting method for patients with lung cancer.

MATERIALS AND METHODS

Retrospective analysis was performed on fifty lung cancer patients whose plan with robustly stereotactic body radiation therapy (SBRT). Thirty out of fifty patients were used for training to build a regression model, based on robust SBRT reference doses, to predict EUD values of ROIs for robust SBPT planning. Thereafter, robust SBPT plans with both automated EUD-Based mimicking (Automated Robust Proton ARP) and manual (Manual Robust Proton MRP) methods were evaluated in the remaining 20 patients. Plans were compared in terms of dosimetric parameters and planning time.

RESULTS

A statistically significantly improvement in target dose fall off was observed for ARP plans compare to MRP plans (Dose fall off: 135 for MRP and 88 for ARP, p < 0.01), while no differences in target coverage and conformity. A statistically significantly reduce in normal lung tissue were observed for ARP plans compare to MRP plans (Lung [D_mean cGy (RBE)]: MRP: 478 vs. ARP: 351, p < 0.01; Lung [V_{5Gy (RBE)} (%)]: MRP: 16.1 vs. ARP: 12.1, p < 0.01; Lung [V_{20Gy (RBE)} (%)]: MRP: 8.5 vs. ARP: 6.8, p < 0.01). Planning time was reduced for ARP plans compare to MRP plans (optimization time: 12 min for MRP vs. 8 min for ARP; total plan time: 23 min for MRP vs. 18 min for ARP).

CONCLUSION

The automated robust SBPT plans using EUD-Based mimicking of SBRT reference dose improve target dose fall off, reduced the radiation doses to the lungs, reduce planning time, which might be beneficial for patient with lung cancer in clinical.

Head and neck IMPT probabilistic dose accumulation: Feasibility of a 2 mm setup uncertainty setting

OBJECTIVE

To establish optimal robust optimization uncertainty settings for clinical head and neck cancer (HNC) patients undergoing 3D image-guided pencil beam scanning (PBS) proton therapy.

METHODS

We analyzed ten consecutive HNC patients treated with 70 and 54.25 Gy_RBE to the primary and prophylactic clinical target volumes (CTV) respectively using intensity-modulated proton therapy (IMPT). Clinical plans were generated using robust optimization with 5 mm/3% setup/range uncertainties (RayStation v6.1). Additional plans were created for 4, 3, 2 and 1 mm setup and 3% range uncertainty and for 3 mm setup and 3%, 2% and 1% range uncertainty. Systematic and random error distributions were determined for setup and range uncertainties based on our quality assurance program. From these, 25 treatment scenarios were sampled for each plan, each consisting of a systematic setup and range error and daily random setup errors. Fraction doses were calculated on the weekly verification CT closest to the date of treatment as this was considered representative of the daily patient anatomy.

RESULTS

Plans with a 2 mm/3% setup/range uncertainty setting adequately covered the primary and prophylactic CTV (V₉₅ ≥ 99% in 98.8% and 90.8% of the treatment scenarios respectively). The average organ-at-risk dose decreased with 1.1 Gy_RBE/mm setup uncertainty reduction and 0.5 Gy_RBE/1% range uncertainty reduction. Normal tissue complication probabilities decreased by 2.0%/mm setup uncertainty reduction and by 0.9%/1% range uncertainty reduction.

CONCLUSION

The results of this study indicate that margin reduction below 3 mm/3% is possible but requires a larger cohort to substantiate clinical introduction.

Salivary Gland Hypofunction and Xerostomia in Head and Neck Radiation Patients

BACKGROUND

The most manifest long-term consequences of radiation therapy in the head and neck cancer patient are salivary gland hypofunction and a sensation of oral dryness (xerostomia).

METHODS

This critical review addresses the consequences of radiation injury to salivary gland tissue, the clinical management of salivary gland hypofunction and xerostomia, and current and potential strategies to prevent or reduce radiation injury to salivary gland tissue or restore the function of radiation-injured salivary gland tissue.

RESULTS

Salivary gland hypofunction and xerostomia have severe implications for oral functioning, maintenance of oral and general health, and quality of life. Significant progress has been made to spare salivary gland function chiefly due to advances in radiation techniques. Other strategies have also been developed, e.g., radioprotectors, identification and preservation/expansion of salivary stem cells by stimulation with cholinergic muscarinic agonists, and application of new lubricating or stimulatory agents, surgical transfer of submandibular glands, and acupuncture.

CONCLUSION

Many advances to manage salivary gland hypofunction and xerostomia induced by radiation therapy still only offer partial protection since they are often of short duration, lack the protective effects of saliva, or potentially have significant adverse effects. Intensity-modulated radiation therapy (IMRT), and its next step, proton therapy, have the greatest potential as a management strategy for permanently preserving salivary gland function in head and neck cancer patients.Presently, gene transfer to supplement fluid formation and stem cell transfer to increase the regenerative potential in radiation-damaged salivary glands are promising approaches for regaining function and/or regeneration of radiation-damaged salivary gland tissue.

Treatment planning for proton therapy: what is needed in the next 10 years?

Treatment planning is the process where the prescription of the radiation oncologist is translated into a deliverable treatment. With the complexity of contemporary radiotherapy, treatment planning cannot be performed without a computerized treatment planning system. Proton therapy (PT) enables highly conformal treatment plans with a minimum of dose to tissues outside the target volume, but to obtain the most optimal plan for the treatment, there are a multitude of parameters that need to be addressed. In this review areas of ongoing improvements and research in the field of PT treatment planning are identified and discussed. The main focus is on issues of immediate clinical and practical relevance to the PT community highlighting the needs for the near future but also in a longer perspective. We anticipate that the manual tasks performed by treatment planners in the future will involve a high degree of computational thinking, as many issues can be solved much better by e.g. scripting. More accurate and faster dose calculation algorithms are needed, automation for contouring and planning is required and practical tools to handle the variable biological efficiency in PT is urgently demanded just to mention a few of the expected improvements over the coming 10 years.