Biochemical control in intermediate- and high-risk prostate cancer after EBRT with and without brachytherapy boost.

PURPOSE: External beam radiotherapy (EBRT) with or without brachytherapy boost (BTB) has not been compared in prospective studies using guideline-recommended radiation dose and recommended androgen-deprivation therapy (ADT). In this multicenter retrospective analysis, we compared modern-day EBRT with BTB in terms of biochemical control (BC) for intermediate-risk (IR) and high-risk (HR) prostate cancer. METHODS: Patients were treated for primary IR or HR prostate cancer during 1999-2019 at three high-volume centers. Inclusion criteria were prescribed ≥ 76 Gy EQD2 (α/ß = 1.5 Gy) for IR and ≥ 78 Gy EQD2 (α/ß = 1.5 Gy) for HR as EBRT alone or with BTB. All HR patients received ADT and pelvic irradiation, which were optional in IR cases. BC between therapies was compared in survival analyses. RESULTS: Of 2769 initial patients, 1176 met inclusion criteria: 468 HR (260 EBRT, 208 BTB) and 708 IR (539 EBRT, 169 BTB). Median follow-up was 49 and 51 months for HR and IR, respectively. BTB patients with ≥ 113 Gy EQD2Gy experienced a stable, good BC outcome compared with BTB at lower doses. Patients treated with ≥ 113 Gy EQD2Gy also experienced significantly improved BC compared with EBRT (10-year BC failure rates after ≥ 113 Gy BTB and EBRT: respectively 20.4 and 41.8% for HR and 7.5 and 20.8% for IR). CONCLUSIONS: In patients with IR and HR prostate cancer, BTB with ≥ 113 Gy EQD2Gy offered a BC advantage compared with dose-escalated EBRT and lower BTB doses.

Dosimetry in MRgPT: Impact of magnetic fields on TLD dose response during proton irradiation.

BACKGROUND: Proton beam therapy, when integrated with MRI guidance, presents complex dosimetric challenges due to interactions with magnetic fields. Prior research has emphasized the nuanced impact of magnetic fields on dosimetry. For thermoluminescent dosimeters (TLDs) the electron-return effect, alongside small air cavities surrounding the pellets, can lead to nonuniform dose distributions. Future MR-guided proton therapy will require reliable methods for end-to-end tests and dosimetric audits, which so far are often performed using TLDs equipped with phantoms. This implicates the necessity of accounting for these interactions. PURPOSE: This study investigates the influence of magnetic fields on TLDs at two proton energies, using magnetic field strengths of 0, 0.25, and 1 T $1 \,\mathrm{T}$ , aiming to clarify their impact on dose measurement accuracy. METHODS: The study was conducted at a synchrotron-based ion beam therapy beam line, enhanced by a resistive dipole magnet for creating magnetic fields up to 1 T $1 \,\mathrm{T}$ to simulate MR-guided proton therapy. Individual correction factors were applied for TLD measurements. The impact of air gaps on the TLD signal was evaluated using three dedicated TLD holders with air gaps of 0.1, 0.25, and 0.5 mm surrounding the TLD pellets using the highest available proton energy of 252.7 M e V $252.7 \,\mathrm{M}\mathrm{e\mathrm{V}}$ . Additionally, the influence of the magnetic field strength on the TLD response was evaluated for two proton energies of 97.4 M e V $97.4 \,\mathrm{M}\mathrm{e\mathrm{V}}$ and 252.7 M e V $252.7 \,\mathrm{M}\mathrm{e\mathrm{V}}$ . RESULTS: The study found no statistically significant variation in TLD dose response attributable to changes in the air gap or the presence of magnetic fields. A power analysis indicated an upper limit on a potential change in dose-response as small as 1.5%. CONCLUSIONS: The findings suggested that the impact of air gap variations and magnetic field strengths on the TLD response was below the detection threshold of TLD sensitivity. This emphasizes the suitability of TLDs for dose measurement in MR-guided proton therapy, indicating that additional correction factors may not be necessary despite the influence of magnetic fields.

Metastases-directed local therapies (MDT) beyond genuine oligometastatic disease (OMD): Indications, endpoints and the role of imaging.

To further personalise treatment in metastatic cancer, the indications for metastases-directed local therapy (MDT) and the biology of oligometastatic disease (OMD) should be kept conceptually apart. Both need to be vigorously investigated. Tumour growth dynamics - growth rate combined with metastatic seeding efficiency - is the single most important biological feature determining the likelihood of success of MDT in an individual patient, which might even be beneficial in slowly developing polymetastatic disease. This can be reasonably well assessed using appropriate clinical imaging. In the context of considering appropriate indications for MDT, detecting metastases at the edge of image resolution should therefore suggest postponing MDT. While three to five lesions are typically used to define OMD, it could be argued that countability throughout the course of metastatic disease, rather than a specific maximum number of lesions, could serve as a better parameter for guiding MDT. Here we argue that the unit of MDT as a treatment option in metastatic cancer might best be defined not as a single procedure at a single point in time, but as a series of treatments that can be delivered in a single or multiple sessions to different lesions over time. Newly emerging lesions that remain amenable to MDT without triggering the start of a new systemic treatment, a change in systemic therapy, or initiation of best supportive care, would thus not constitute a failure of MDT. This would have implications for defining endpoints in clinical trials and registries: Rather than with any disease progression, failure of MDT would only be declared when there is progression to polymetastatic disease, which then precludes further options for MDT.

Explicitly encoding the cyclic nature of breathing signal allows for accurate breathing motion prediction in radiotherapy with minimal training data.

Background and purpose: Active breathing motion management in radiotherapy consists of motion monitoring, quantification and mitigation. It is impacted by associated latencies of a few 100 ms. Artificial neural networks can successfully predict breathing motion and eliminate latencies. However, they require usually a large dataset for training. The objective of this work was to demonstrate that explicitly encoding the cyclic nature of the breathing signal into the training data enables significant reduction of training datasets which can be obtained from healthy volunteers. Material and methods: Seventy surface scanner breathing signals from 25 healthy volunteers in anterior-posterior direction were used for training and validation (ratio 4:1) of long short-term memory models. The model performance was compared to a model using decomposition into phase, amplitude and a time-dependent baseline. Testing of the models was performed on 55 independent breathing signals in anterior-posterior direction from surface scanner (35 lung, 20 liver) of 30 patients with a mean breathing amplitude of (5.9 ± 6.7) mm. Results: Using the decomposed breathing signal allowed for a reduction of the absolute root-mean square error (RMSE) from 0.34 mm to 0.12 mm during validation. Testing using patient data yielded an average absolute RMSE of the breathing signal of (0.16 ± 0.11) mm with a prediction horizon of 500 ms. Conclusion: It was demonstrated that a motion prediction model can be trained with less than 100 datasets of healthy volunteers if breathing cycle parameters are considered. Applied to 55 patients, the model predicted breathing motion with a high accuracy.

Increasing Quality and Efficiency of the Radiotherapy Treatment Planning Process by Constructing and Implementing a Workflow-Monitoring Application.

PURPOSE: To demonstrate how the efficiency of the treatment planning processes of a university radiation oncology department (2,500 new patients/year) could be improved by constructing and implementing a workflow-monitoring application. METHODS: A web-based application was developed in house, which enhanced the process management tools of the clinic's oncology information system. The application calculates the days left for the next task in the treatment planning process and visualizes the information on a browser-based whiteboard. Workflow monitoring considers tumor types (breast, prostate, lung, etc) and treatment techniques and is backward planned from the planned start of treatment. The effect of introducing this application was analyzed over four phases: (1) baseline data without the workflow-monitoring application, (2) after introducing workflow visualization via a browser-based whiteboard, (3) after upgrading the whiteboard and introducing backend rules, and (4) after updating these rules on the basis of data from the previous phase. RESULTS: Implementing the workflow-monitoring application and the introduced measures significantly reduced delays and, consequently, stress and a negative working atmosphere in the treatment planning process. Most notably, the amount of last-minute physics checks (on the day of the treatment start) could be reduced by 50%. CONCLUSION: The study showed what measures can help organize and prioritize the treatment planning workflow. The increased efficiency is believed to improve the quality and reduce the risk of human error.

On the sensitivity of PROMs during breast radiotherapy.

Purpose: To investigate the sensitivity of patient-reported outcome measures (PROMs) to detect treatment-related side effects in patients with breast cancer undergoing external beam photon radiotherapy. Methods: As part of daily clinical care, an in-house developed PROM tool was used to assess side effects in patients during a) whole-breast irradiation (WBI) to 40 Gy, b) WBI with a sequential boost of 10 Gy, and c) partial-breast irradiation (PBI) to 40 Gy. Results: 414 patients participated in this prospective study between October 2020 and January 2022, with 128 patients (31 %) receiving WBI, 241 (58 %) receiving WBI followed by a sequential boost, and 50 patients (12 %) receiving PBI. Significant differences in the reported toxicities (itching, radiation skin reaction, skin darkening, and tenderness and swelling) were reported between the WBI cohorts with and without boost (p < 0.001, p < 0.001, p < 0.001, and p = 0.002, respectively). The comparison of PBI with WBI (no-boost) yielded significant differences for radiation skin reaction (p < 0.001). Conclusion: The results highlight the high sensitivity of PROMs to detect treatment-related side effects in patients with breast cancer. Thus, PROMs may be a valuable tool for quality control and may support evidence-based learning from real-world data originating from daily routine care.

Impact of log file source and data frequency on accuracy of log file-based patient specific quality assurance.

Performing phantom measurements for patient-specific quality assurance (PSQA) adds a significant amount of time to the adaptive radiotherapy procedure. Log file based PSQA can be used to increase the efficiency of this process. This study compared the dosimetric accuracy of high-frequency linear accelerator (Linac) log files and low-frequency log data stored in the oncology information system (OIS). Thirty patients were included, that were recently treated in the head and neck (HN), brain, and prostate region with volumetric modulated arc therapy (VMAT) and an additional ten patients treated using stereotactic body radiation therapy (SBRT) with 3D-conformal radiotherapy (3D-CRT) technique. Log data containing a single fraction were used to calculate the dose distributions. The dosimetric differences between Linac log files and OIS logs were evaluated with a gamma analysis with 2%/2 mm criterion and dose threshold of 30%. The original treatment plan was used as a reference. Moreover, DVH parameters of D98%, D50%, and D2% of the planning-target volume (PTV) and dose to several organs at risk (OARs) were reported. Significant differences in dose distributions between the two log types and the original dose were observed for PTV D98% and D2% (r < 0.001) for HN cases, PTV D98% (r = 0.005) for brain cases, and PTV D50% (r = 0.015) for prostate cases. No significant differences were found between the two log types with respect to D50%. The root mean square (RMS) error of the leaf positions of the OIS log was approximately twice the RMS error of the Linac log file for VMAT plans, but identical for 3D-CRT plans. The relationship between the gamma pass rate and the RMS error showed a moderate correlation for the Linac log files (r = -0.58, p < 0.001) and strong correlation for OIS logs (r = -0.71, p < 0.001). Furthermore, all doses calculated using Linac log files and OIS log data had a GPR >90% for an RMS error < 3.3 mm. Based on these findings, a tolerance limit of RMS error of 3.3 mm for considering OIS log based PSQA was established. Nevertheless, the OIS log data quality should be improved to achieve adequate PSQA.

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.

EviGUIDE - a tool for evidence-based decision making in image-guided adaptive brachytherapy for cervical cancer.

PURPOSE: To develop a novel decision-support system for radiation oncology that incorporates clinical, treatment and outcome data, as well as outcome models from a large clinical trial on magnetic resonance image-guided adaptive brachytherapy (MR-IGABT) for locally advanced cervical cancer (LACC). METHODS: A system, called EviGUIDE, was developed that combines dosimetric information from the treatment planning system, patient and treatment characteristics, and established tumor control probability (TCP), and normal tissue complication probability (NTCP) models, to predict clinical outcome of radiotherapy treatment of LACC. Six Cox Proportional Hazards models based on data from 1341 patients of the EMBRACE-I study have been integrated. One TCP model for local tumor control, and five NTCP models for OAR morbidities. RESULTS: EviGUIDE incorporates TCP-NTCP graphs to help users visualize the clinical impact of different treatment plans and provides feedback on achievable doses based on a large reference population. It enables holistic assessment of the interplay between multiple clinical endpoints and tumour and treatment variables. Retrospective analysis of 45 patients treated with MR-IGABT showed that there exists a sub-cohort of patients (20%) with increased risk factors, that could greatly benefit from the quantitative and visual feedback. CONCLUSION: A novel digital concept was developed that can enhance clinical decision- making and facilitate personalized treatment. It serves as a proof of concept for a new generation of decision support systems in radiation oncology, which incorporate outcome models and high-quality reference data, and aids the dissemination of evidence-based knowledge about optimal treatment and serve as a blueprint for other sites in radiation oncology.

Clinical Implementation and Evaluation of Auto-Segmentation Tools for Multi-Site Contouring in Radiotherapy.

Background and purpose: Tools for auto-segmentation in radiotherapy are widely available, but guidelines for clinical implementation are missing. The goal was to develop a workflow for performance evaluation of three commercial auto-segmentation tools to select one candidate for clinical implementation. Materials and Methods: One hundred patients with six treatment sites (brain, head-and-neck, thorax, abdomen, and pelvis) were included. Three sets of AI-based contours for organs-at-risk (OAR) generated by three software tools and manually drawn expert contours were blindly rated for contouring accuracy. The dice similarity coefficient (DSC), the Hausdorff distance, and a dose/volume evaluation based on the recalculation of the original treatment plan were assessed. Statistically significant differences were tested using the Kruskal-Wallis test and the post-hoc Dunn Test with Bonferroni correction. Results: The mean DSC scores compared to expert contours for all OARs combined were 0.80 ± 0.10, 0.75 ± 0.10, and 0.74 ± 0.11 for the three software tools. Physicians' rating identified equivalent or superior performance of some AI-based contours in head (eye, lens, optic nerve, brain, chiasm), thorax (e.g., heart and lungs), and pelvis and abdomen (e.g., kidney, femoral head) compared to manual contours. For some OARs, the AI models provided results requiring only minor corrections. Bowel-bag and stomach were not fit for direct use. During the interdisciplinary discussion, the physicians' rating was considered the most relevant. Conclusion: A comprehensive method for evaluation and clinical implementation of commercially available auto-segmentation software was developed. The in-depth analysis yielded clear instructions for clinical use within the radiotherapy department.

Tools for large-scale data analytics of an international multi-center study in radiation oncology for cervical cancer.

PURPOSE: To develop and implement a software that enables centers, treating patients with state-of-the-art radiation oncology, to compare their patient, treatment, and outcome data to a reference cohort, and to assess the quality of their treatment approach. MATERIALS AND METHODS: A comprehensive data dashboard was designed, which al- lowed holistic assessment of institutional treatment approaches. The software was tested in the ongoing EMBRACE-II study for locally advanced cervical cancer. The tool created individualized dashboards and automatic analysis scripts, verified pro- tocol compliance and checked data for inconsistencies. Identified quality assurance (QA) events were analysed. A survey among users was conducted to assess usability. RESULTS: The survey indicated favourable feedback to the prototype and highlighted its value for internal monitoring. Overall, 2302 QA events were identified (0.4% of all collected data). 54% were due to missing or incomplete data, and 46% originated from other causes. At least one QA event was found in 519/1001 (52%) of patients. QA events related to primary study endpoints were found in 16% of patients. Sta- tistical methods demonstrated good performance in detecting anomalies, with precisions ranging from 71% to 100%. Most frequent QA event categories were Treatment Technique (27%), Patient Characteristics (22%), Dose Reporting (17%), Outcome 156 (15%), Outliers (12%), and RT Structures (8%). CONCLUSION: A software tool was developed and tested within a clinical trial in radia- tion oncology. It enabled the quantitative and qualitative comparison of institutional patient and treatment parameters with a large multi-center reference cohort. We demonstrated the value of using statistical methods to automatically detect implau- sible data points and highlighted common pitfalls and uncertainties in radiotherapy for cervical cancer.

Can Generative Adversarial Networks help to overcome the limited data problem in segmentation?

PURPOSE: For image translational tasks, the application of deep learning methods showed that Generative Adversarial Network (GAN) architectures outperform the traditional U-Net networks, when using the same training data size. This study investigates whether this performance boost can also be expected for segmentation tasks with small training dataset size. MATERIALS/METHODS: Two models were trained on varying training dataset sizes ranging from 1-100 patients: a) U-Net and b) U-Net with patch discriminator (conditional GAN). The performance of both models to segment the male pelvis on CT-data was evaluated (Dice similarity coefficient, Hausdorff) with respect to training data size. RESULTS: No significant differences were observed between the U-Net and cGAN when the models were trained with the same training sizes up to 100 patients. The training dataset size had a significant impact on the models' performances, with vast improvements when increasing dataset sizes from 1 to 20 patients. CONCLUSION: When introducing GANs for the segmentation task no significant performance boost was observed in our experiments, even in segmentation models developed on small datasets.

Neural network-assisted automated image registration for MRI-guided adaptive brachytherapy in cervical cancer.

PURPOSE: In image-guided adaptive brachytherapy (IGABT) a quantitative evaluation of the dosimetric changes between fractions due to anatomical variations, can be implemented via rigid registration of images from subsequent fractions based on the applicator as a reference structure. With available treatment planning systems (TPS), this is a manual and time-consuming process. The aim of this retrospective study was to automate this process. A neural network (NN) was trained to predict the applicator structure from MR images. The resulting segmentation was used to automatically register MR-volumes. MATERIAL AND METHODS: DICOM images and plans of 56 patients treated for cervical cancer with high dose-rate (HDR) brachytherapy were used in the study. A 2D and a 3D NN were trained to segment applicator structures on clinical T2-weighted MRI datasets. Different rigid registration algorithms were investigated and compared. To evaluate a fully automatic registration workflow, the NN-predicted applicator segmentations (AS) were used for rigid image registration with the best performing algorithm. The DICE coefficient and mean distance error between dwell positions (MDE) were used to evaluate segmentation and registration performance. RESULTS: The mean DICE coefficient for the predicted AS was 0.70 ±â€¯0.07 and 0.58 ±â€¯0.04 for the 3D NN and 2D NN, respectively. Registration algorithms achieved MDE errors from 8.1 ±â€¯3.7 mm (worst) to 0.7 ±â€¯0.5 mm (best), using ground-truth AS. Using the predicted AS from the 3D NN together with the best registration algorithm, an MDE of 2.7 ±â€¯1.4 mm was achieved. CONCLUSION: Using a combination of deep learning models and state of the art image registration techniques has been demonstrated to be a promising solution for automatic image registration in IGABT. In combination with auto-contouring of organs at risk, the auto-registration workflow from this study could become part of an online-dosimetric interfraction evaluation workflow in the future.

Accelerating and improving radiochromic film calibration by utilizing the dose ratio in photon and proton beams.

PURPOSE: Radiochromic films are versatile 2D dosimeters with high-resolution and near tissue equivalence. To assure high precision and accuracy, a time-consuming calibration process is required. To improve the time efficiency, a novel calibration method utilizing the ratio of the same dose profile measured at different monitor units (MUs) is introduced and tested in a proton and photon beam. METHODS: The calibration procedure employs the dose ratio of film measurements of the same relative profile for different absolute dose values. Hence, the ratio of the dose is constant at any point of the profile, but the ratio of the net optical densities is not constant. The key idea of the method is to optimize the calibration function until the ratio of the calculated doses is constant. The proposed method was tested in the dose range between 0.25-12 and 1-6 Gy in a proton and photon beam, respectively. A radial symmetric profile and a rectangular profile were created, both having a central plateau region of about 3 cm diameter and a dose falloff of about 1.5 cm at larger distances. The dose falloff region was used as input for the optimization method and the central plateau region served as dose reference points. Only the plateau region of the highest dose entered the optimization as an additional objective. The measured data were randomly split into differently sized training and test sets. The optimization was repeated 1000 times with random start value initialization using the same start values for the standard and the gradient method. Finally, a proton plan with four dose levels was created, which were separated spatially, to test the possibility of a full calibration within a single measurement. RESULTS: Parameter estimation was possible with as low as one dose ratio used for optimization in both the photon and the proton case, yet exhibiting a high sensitivity on the dose level. The root mean squared deviation (RMSD) of the dose was less than 1% when the dose ratio was in the order of 20, whereas the median RMSD of all optimizations was 1.7%. Using four dose levels for optimization resulted in a median RMSD of 1% when randomly selecting the dose levels. Having at least one dose ratio of about 20 included in the optimization considerably improved the RMSD of the calibration function. Using six or eight dose levels reduced the sensitivity on the dose level selection and the median RMSD was 0.8%. A full calibration was possible in a single measurement having four dose levels in one plan but spatially separated. CONCLUSIONS: The number of measurements required to obtain an EBT3 film calibration function could be reduced using the proposed dose ratio method while maintaining the same accuracy as with the standard method.

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.

Technical Note: Dose prediction for radiation therapy using feature-based losses and One Cycle Learning.

PURPOSE: To present the technical details of the runner-up model in the open knowledge-based planning (OpenKBP) challenge for the dose-volume histogram (DVH) stream. The model was designed to ensure simple and reproducible training, without the necessity of costly advanced generative adversarial network (GAN) techniques. METHODS: The model was developed based on the OpenKBP challenge dataset, consisting of 200 and 40 head-and-neck patients for training and validation, respectively. The final model is a U-Net with additional ResNet blocks between up- and down convolutions. The results were obtained by training the model with AdamW with the One Cycle scheduler. The loss function is a combination of the L1 loss with a feature loss, which uses a pretrained video classifier as a feature extractor. The performance was evaluated on another 100 patients in the OpenKBP test dataset. The DVH metrics of the test data were evaluated, where D 0.1 c c , and D mean were calculated for the organs at risk (OARs) and D 1 % , D 95 % , and D 99 % were computed for the target structures. DVH metric differences between predicted and true dose are reported in percentage. RESULTS: The model achieved 2nd and 4th place in the DVH and dose stream of the OpenKBP challenge, respectively. The dose and DVH score were 2.62 ± 1.10 and 1.52 ± 1.06, respectively. Mean dose differences for the different structures and DVH parameters were within ±1%. CONCLUSION: This straightforward approach produced excellent results. It incorporated One Cycle Learning, ResNet, and feature-based losses, which are common computer vision techniques.

An MR-only acquisition and artificial intelligence based image-processing protocol for photon and proton therapy using a low field MR.

OBJECTIVE: Recent developments on synthetically generated CTs (sCT), hybrid MRI linacs and MR-only simulations underlined the clinical feasibility and acceptance of MR guided radiation therapy. However, considering clinical application of open and low field MR with a limited field of view can result in truncation of the patient's anatomy which further affects the MR to sCT conversion. In this study an acquisition protocol and subsequent MR image stitching is proposed to overcome the limited field of view restriction of open MR scanners, for MR-only photon and proton therapy. MATERIAL AND METHODS: 12 prostate cancer patients scanned with an open 0.35T scanner were included. To obtain the full body contour an enhanced imaging protocol including two repeated scans after bilateral table movement was introduced. All required structures (patient contour, target and organ at risk) were delineated on a post-processed combined transversal image set (stitched MRI). The postprocessed MR was converted into a sCT by a pretrained neural network generator. Inversely planned photon and proton plans (VMAT and SFUD) were designed using the sCT and recalculated for rigidly and deformably registered CT images and compared based on D2%, D50%, V70Gy for organs at risk and based on D2%, D50%, D98% for the CTV and PTV. The stitched MRI and the untruncated MRI were compared to the CT, and the maximum surface distance was calculated. The sCT was evaluated with respect to delineation accuracy by comparing on stitched MRI and sCT using the DICE coefficient for femoral bones and the whole body. RESULTS: Maximum surface distance analysis revealed uncertainties in lateral direction of 1-3mm on average. DICE coefficient analysis confirms good performance of the sCT conversion, i.e. 92%, 93%, and 100% were obtained for femoral bone left and right and whole body. Dose comparison resulted in uncertainties below 1% between deformed CT and sCT and below 2% between rigidly registered CT and sCT in the CTV for photon and proton treatment plans. DISCUSSION: A newly developed acquisition protocol for open MR scanners and subsequent Sct generation revealed good acceptance for photon and proton therapy. Moreover, this protocol tackles the restriction of the limited FOVs and expands the capacities towards MR guided proton therapy with horizontal beam lines.

Latent space manipulation for high-resolution medical image synthesis via the StyleGAN.

INTRODUCTION: This paper explores the potential of the StyleGAN model as an high-resolution image generator for synthetic medical images. The possibility to generate sample patient images of different modalities can be helpful for training deep learning algorithms as e.g. a data augmentation technique. METHODS: The StyleGAN model was trained on Computed Tomography (CT) and T2- weighted Magnetic Resonance (MR) images from 100 patients with pelvic malignancies. The resulting model was investigated with regards to three features: Image Modality, Sex, and Longitudinal Slice Position. Further, the style transfer feature of the StyleGAN was used to move images between the modalities. The root-mean-squard error (RMSE) and the Mean Absolute Error (MAE) were used to quantify errors for MR and CT, respectively. RESULTS: We demonstrate how these features can be transformed by manipulating the latent style vectors, and attempt to quantify how the errors change as we move through the latent style space. The best results were achieved by using the style transfer feature of the StyleGAN (58.7 HU MAE for MR to CT and 0.339 RMSE for CT to MR). Slices below and above an initial central slice can be predicted with an error below 75 HU MAE and 0.3 RMSE within 4cm for CT and MR, respectively. DISCUSSION: The StyleGAN is a promising model to use for generating synthetic medical images for MR and CT modalities as well as for 3D volumes.

Investigating conditional GAN performance with different generator architectures, an ensemble model, and different MR scanners for MR-sCT conversion.

Recent developments in magnetic resonance (MR) to synthetic computed tomography (sCT) conversion have shown that treatment planning is possible without an initial planning CT. Promising conversion results have been demonstrated recently using conditional generative adversarial networks (cGANs). However, the performance is generally only tested on images from one MR scanner, which neglects the potential of neural networks to find general high-level abstract features. In this study, we explored the generalizability of the generator models, trained on a single field strength scanner, to data acquired with higher field strengths. T2-weighted 0.35T MRIs and CTs from 51 patients treated for prostate (40) and cervical cancer (11) were included. 25 of them were used to train four different generators (SE-ResNet, DenseNet, U-Net, and Embedded Net). Further, an ensemble model was created from the four network outputs. The models were validated on 16 patients from a 0.35T MR scanner. Further, the trained models were tested on the Gold Atlas dataset, containing T2-weighted MR scans of different field strengths; 1.5T(7) and 3T(12), and 10 patients from the 0.35T scanner. The sCTs were dosimetrically compared using clinical VMAT plans for all test patients. For the same scanner (0.35T), the results from the different models were comparable on the test set, with only minor differences in the mean absolute error (MAE) (35-51HU body). Similar results were obtained for conversions of 3T GE Signa and the 3T GE Discovery images (40-62HU MAE) for three of the models. However, larger differences were observed for the 1.5T images (48-65HU MAE). The overall best model was found to be the ensemble model. All dose differences were below 1%. This study shows that it is possible to generalize models trained on images of one scanner to other scanners and different field strengths. The best metric results were achieved by the combination of all networks.

Basic Properties of a New Polymer Gel for 3D-Dosimetry at High Dose-Rates Typical for FFF Irradiation Based on Dithiothreitol and Methacrylic Acid (MAGADIT): Sensitivity, Range, Reproducibility, Accuracy, Dose Rate Effect and Impact of Oxygen Scavenger.

The photon induced radical-initiated polymerization in polymer gels can be used for high-resolution tissue equivalent dosimeters in quality control of radiation therapy. The dose (D) distribution in radiation therapy can be measured as a change of the physical measurement parameter T2 using T2-weighted magnetic resonance imaging. The detection by T2 is relying on the local change of the molecular mobility due to local polymerization initiated by radicals generated by the ionizing radiation. The dosimetric signals R2 = 1/T2 of many of the current polymer gels are dose-rate dependent, which reduces the reliability of the gel for clinical use. A novel gel dosimeter, based on methacrylic acid, gelatin and the newly added dithiothreitol (MAGADIT) as an oxygen-scavenger was analyzed for basic properties, such as sensitivity, reproducibility, accuracy and dose-rate dependence. Dithiothreitol features no toxic classification with a difference to THPC and offers a stronger negative redox-potential than ascorbic acid. Polymer gels with three different concentration levels of dithiothreitol were irradiated with a preclinical research X-ray unit and MR-scanned (T2) for quantitative dosimetry after calibration. The polymer gel with the lowest concentration of the oxygen scavenger was about factor 3 more sensitive to dose as compared to the gel with the highest concentration. The dose sensitivity (α = ∆R2/∆D) of MAGADIT gels was significantly dependent on the applied dose rate D ˙ (≈48% reduction between D ˙ = 0.6 Gy/min and D ˙ = 4 Gy/min). However, this undesirable dose-rate effect reduced between 4-8 Gy/min (≈23%) and almost disappeared in the high dose-rate range (8 ≤   D ˙ ≤   12 Gy/min) used in flattening-filter-free (FFF) irradiations. The dose response varied for different samples within one manufacturing batch within 3%-6% (reproducibility). The accuracy ranged between 3.5% and 7.9%. The impact of the dose rate on the spatial integrity is demonstrated in the example of a linear accelerator (LINAC) small sized 5 × 10 mm2 10 MV photon field. For MAGADIT the maximum shift in the flanks in this field is limited to about 0.8 mm at a FFF dose rate of 15 Gy/min. Dose rate sensitive polymer gels likely perform better at high dose rates; MAGADIT exhibits a slightly improved performance compared to the reference normoxic polymer gel methacrylic and ascorbic acid in gelatin initiated by copper (MAGIC) using ascorbic acid.