Radiotherapy Workflow Research Articles

Intentional deep overfit learning (IDOL): A novel deep learning strategy for adaptive radiation therapy.

Applications of deep learning (DL) are essential to realizing an effective adaptive radiotherapy (ART) workflow. Despite the promise demonstrated by DL approaches in several critical ART tasks, there remain unsolved challenges to achieve satisfactory generalizability of a trained model in a clinical setting. Foremost among these is the difficulty of collecting a task-specific training dataset with high-quality, consistent annotations for supervised learning applications. In this study, we propose a tailored DL framework for patient-specific performance that leverages the behavior of a model intentionally overfitted to a patient-specific training dataset augmented from the prior information available in an ART workflow-an approach we term Intentional Deep Overfit Learning (IDOL). Implementing the IDOL framework in any task in radiotherapy consists of two training stages: (1) training a generalized model with a diverse training dataset of patients, just as in the conventional DL approach, and (2) intentionally overfitting this general model to a small training dataset-specific the patient of interest ( ) generated through perturbations and augmentations of the available task- and patient-specific prior information to establish a personalized IDOL model. The IDOL framework itself is task-agnostic and is, thus, widely applicable to many components of the ART workflow, three of which we use as a proof of concept here: the autocontouring task on replanning CTs for traditional ART, the MRI super-resolution (SR) task for MRI-guided ART, and the synthetic CT (sCT) reconstruction task for MRI-only ART. In the replanning CT autocontouring task, the accuracy measured by the Dice similarity coefficient improves from 0.847 with the general model to 0.935 by adopting the IDOL model. In the case of MRI SR, the mean absolute error (MAE) is improved by 40% using the IDOL framework over the conventional model. Finally, in the sCT reconstruction task, the MAE is reduced from 68 to 22 HU by utilizing the IDOL framework. In this study, we propose a novel IDOL framework for ART and demonstrate its feasibility using three ART tasks. We expect the IDOL framework to be especially useful in creating personally tailored models in situations with limited availability of training data but existing prior information, which is usually true in the medical setting in general and is especially true in ART.

Machine-assisted interpolation algorithm for semi-automated segmentation of highly deformable organs.

Accurate and robust auto-segmentation of highly deformable organs (HDOs), for example, stomach or bowel, remains an outstanding problem due to these organs' frequent and large anatomical variations. Yet, time-consuming manual segmentation of these organs presents a particular challenge to time-limited modern radiotherapy techniques such as on-line adaptive radiotherapy and high-dose-rate brachytherapy. We propose a machine-assisted interpolation (MAI) that uses prior information in the form of sparse manual delineations to facilitate rapid, accurate segmentation of the stomach from low field magnetic resonance images (MRI) and the bowel from computed tomography (CT) images. Stomach MR images from 116 patients undergoing 0.35T MRI-guided abdominal radiotherapy and bowel CT images from 120 patients undergoing high dose rate pelvic brachytherapy treatment were collected. For each patient volume, the manual delineation of the HDO was extracted from every 8th slice. These manually drawn contours were first interpolated to obtain an initial estimate of the HDO contour. A two-channel 64×64pixel patch-based convolutional neural network (CNN) was trained to localize the position of the organ's boundary on each slice within a five-pixel wide road using the image and interpolated contour estimate. This boundary prediction was then input, in conjunction with the image, to an organ closing CNN which output the final organ segmentation. A Dense-UNet architecture was used for both networks. The MAI algorithm was separately trained for the stomach segmentation and the bowel segmentation. Algorithm performance was compared against linear interpolation (LI) alone and against fully automated segmentation (FAS) using a Dense-UNet trained on the same datasets. The Dice Similarity Coefficient (DSC) and mean surface distance (MSD) metrics were used to compare the predictions from the three methods. Statistically significance was tested using Student's t test. For the stomach segmentation, the mean DSC from MAI (0.91 ± 0.02) was 5.0% and 10.0% higher as compared to LI and FAS, respectively. The average MSD from MAI (0.77 ± 0.25mm) was 0.54 and 3.19mm lower compared to the two other methods. Only 7% of MAI stomach predictions resulted in a DSC<0.8, as compared to 30% and 28% for LI and FAS, respectively. For the bowel segmentation, the mean DSC of MAI (0.90 ± 0.04) was 6% and 18% higher, and the average MSD of MAI (0.93 ± 0.48mm) was 0.42 and 4.9mm lower as compared to LI and FAS. Sixteen percent of the predicted contour from MAI resulted in a DSC<0.8, as compared to 46% and 60% for FAS and LI, respectively. All comparisons between MAI and the baseline methods were found to be statistically significant (p-value<0.001). The proposed MAI algorithm significantly outperformed LI in terms of accuracy and robustness for both stomach segmentation from low-field MRIs and bowel segmentation from CT images. At this time, FAS methods for HDOs still require significant manual editing. Therefore, we believe that the MAI algorithm has the potential to expedite the process of HDO delineation within the radiation therapy workflow.

The impact of gadolinium-based MR contrast on radiotherapy planning for oropharyngeal treatment on the MR Linac.

Gadolinium-based contrast agents (GBCAs) may add value to magnetic resonance (MR)-only radiotherapy (RT) workflows including on hybrid machines such as the MR Linac. The impact of GBCAs on RT dose distributions however have not been well studied. This work used retrospective GBCA-enhanced datasets to assess the dosimetric effect of GBCAs on head and neck plans. Ten patients with oropharyngeal squamous cell carcinoma receiving RT from November 2018 to April 2020 were included in this study. RT planning included contrast-enhanced computed tomography (CT) and MR scans. A contrast agent "contour" was defined by delineating GBCA-enhanced regions using an agreed window/level threshold, transferred to the planning CT and given a standardized electron density (ED) of 1.149 in the Monaco treatment planning system (Elekta AB). Four plans were per patient calculated and compared using two methods: (1) optimized without contrast (Plan A) then recalculated with ED (Plan B), and (2) optimized with contrast ED (Plan C) then without (Plan D). For target parameters minimum and maximum doses to 1cc of PTVs, D95 values, and percent dose differences were calculated. Dose differences for organs-at-risk (OARs) were calculated as a percentage of the clinical tolerance value. For the purposes of this study,±2% over the whole treatment course was considered to be a clinically acceptable dose deviation. Wilcoxon-signed rank tests were used to determine any dose differences within and between the two methods of optimization and recalculation (p<0.05). Pearson's correlations were used to establish the relationship between gadolinium uptake volume in a structure (i.e., proportion of structure covered by a density override) and the resulting dose difference. The median percent dose differences for key reportable dosimetric parameters between non-contrast and simulated contrast plans were<1.2% over all fractions over all patients for reportable target parameters (mean 0.34%, range 0.22%-1.02%). The percent dose differences for maximum dose to 1cc of both PTV1 and PTV2 were significantly different after application of density override (p<0.05) but only in method 2 (Plan C vs. Plan D). For D95 PTV1, there was a statistically significant effect of density override (p<0.01), however only in method 1 (Plan A vs. Plan B). There were no significant differences between calculation methods of the impact of contrast in most target parameters with the exception of D95 PTV1 (p<0.01) and for D95 PTV2 (p<0.05). The median percent dose differences for reportable OAR parameters as a percentage of clinical planning tolerances were<2.0% over a full treatment course (mean 0.65%, range 0.27%-1.62%). There were no significant differences in dose to OARs within or between methods for contrast impact assessment. Dose differences to targets and OARs in oropharyngeal cancer treatment due to the presence of GBCA were minimal, and this work suggests that prospective in vivo evaluations of impact may not be necessary in this clinical site. Accounting for GBCAs may not be needed in daily adaptive workflows on the MR Linac.

A convolutional neural network for estimating cone-beam CT intensity deviations from virtual CT projections

Extending cone-beam CT (CBCT) use toward dose accumulation and adaptive radiotherapy (ART) necessitates more accurate HU reproduction since cone-beam geometries are heavily degraded by photon scatter. This study proposes a novel method which aims to demonstrate how deep learning based on phantom data can be used effectively for CBCT intensity correction in patient images. Four anthropomorphic phantoms were scanned on a CBCT and conventional fan-beam CT system. Intensity correction is performed by estimating the cone-beam intensity deviations from prior information contained in the CT. Residual projections were extracted by subtraction of raw cone-beam projections from virtual CT projections. An improved version of U-net is utilized to train on a total of 2001 projection pairs. Once trained, the network could estimate intensity deviations from input patient head and neck raw projections. The results from our novel method showed that corrected CBCT images improved the (contrast-to-noise ratio) with respect to uncorrected reconstructions by a factor of 2.08. The mean absolute error and structural similarity index improved from 318 HU to 74 HU and 0.750 to 0.812 respectively. Visual assessment based on line-profile measurements and difference image analysis indicate the proposed method reduced noise and the presence of beam-hardening artefacts compared to uncorrected and manufacturer reconstructions. Projection domain intensity correction for cone-beam acquisitions of patients was shown to be feasible using a convolutional neural network trained on phantom data. The method shows promise for further improvements which may eventually facilitate dose monitoring and ART in the clinical radiotherapy workflow.

Treatment Planning System for Clinical Translation of Electron FLASH Radiotherapy

Electron ultrahigh dose rate FLASH radiotherapy (UHDR/eFLASH-RT) is currently being extensively tested and modeled, and was studied in several rodent and large animal experiments and one clinical cancer patient. These studies suggest prospective dosimetry assessments are necessary to ensure safe delivery and widespread clinical translation of eFLASH-RT would clearly benefit from prior dose prediction and a treatment planning process that accounts for anatomical heterogeneity and complex geometries. In this study, a GAMOS Monte Carlo beam model was developed and implemented in a clinical treatment planning system for a modified UHDR electron LINAC to demonstrate eFLASH-RT planning of patients.The gantry head was modelled without scattering foils or targets, representative of the LINAC modifications. Centered at the nominal energy of 10 MeV, the energy spectrum (σE) was varied to match the central axis percent depth dose (PDD) profiles of an open field measured with radiographic film. The beam's emittance cone angle (θcone) was determined by matching calculated and film measured lateral profiles, at several depths. The beam model and its configuration in the treatment planning system were validated with film measured lateral profiles and PDD's of the open field and nominal fields for clinical applicators. A GAMOS forward dose calculation was conducted for a mouse whole brain treatment and an eFLASH-RT plan was compared to a conventional electron plan for a human patient with a rib metastasis.Our eFLASH beam model agrees best with measurements at σE = 0.5MeV and θcone = 3.9o ± 0.2o. The model and its treatment planning system configuration were validated to clinically acceptable accuracy (the absolute average error was within 1.5% for in water lateral, 3% for in air lateral, and 2% for PDD's). The forward dose calculation showed dose was delivered to the entire mouse brain with adequate conformality (90% of the tumor volume receiving 90% of the prescribed dose). The comparison plans for the patient undergoing treatment for a rib metastasis demonstrated the capability to achieve an acceptable eFLASH-RT plan with routine accessories (90% of the tumor receiving 95% and 90% of the prescribed dose for eFLASH and conventional, respectively) in relatively complex geometry (right posterior oblique treatment with patient specific applicator cutout).To the best of our knowledge, this study presents the first functional beam model commissioned in a clinical TPS for eFLASH-RT, enabling treatment planning and evaluation with minimal deviation from conventional radiotherapy workflow. By producing plans with the quality comparable to conventional plans, specifically for human patients involving complex geometries and tissue heterogeneity, the clinical translation of eFLASH-RT can be expedited. The open-source methods (described in https://github.com/mr3536/eFLASHBeamModeltoTPS) can model other eFLASH irradiators with different beam characteristics.

Resource planning principles for the radiotherapy process using simulations applied to a longer vacation period use case.

Radiotherapy (RT) resources need to be used wisely to balance workload and patient throughput. There are no known strategies on how to plan resource use around longer vacation periods to avoid patient waiting times. We created a simulation model over the RT workflow to evaluate different scenarios for this purpose. The simulation model mimics a large modern RT department in Sweden. It was based on real data on patient referral patterns and resource use extracted from clinical systems (3666 treatment courses). Workshops with managers and staff were held to investigate nine different scenarios for the summer vacation period including one scenario to validate the model. Different capacity reductions, vacation period lengths and timing of the vacation periods between the preparatory part of the RT workflow and the treatment part were evaluated. For an eight-week period, resource utilization was predicted to be high both before and after the vacation period regardless of timing. However, more patients would be waiting with completed preparations with simultaneous vacation periods than when the preparatory part started one-two weeks prior to the treatment part. With shorter vacation periods, treatment would require overtime during the vacation period with higher levels of patients waiting compared to an eight-week scenario. Our proposed strategy aided managers to identify a preferred scenario for the summer vacation period with the preparatory part starting one-two weeks prior to the treatment part for an eight-week vacation period. This can help other RT departments to plan for similar situations.

Technical overview of magnetic resonance fingerprinting and its applications in radiation therapy

Magnetic resonance fingerprinting (MRF) is an emerging imaging technique for rapid and simultaneous quantification of multiple tissue properties. The technique has been developed for quantitative imaging of different organs. The obtained quantitative measures have the potential to improve multiple steps of a typical radiotherapy workflow and potentially further improve integration of magnetic resonance imaging guided clinical decision making. In this review paper, we first provide a technical overview of the MRF method from data acquisition to postprocessing, along with recent development in advanced reconstruction methods. We further discuss critical aspects that could influence its usage in radiation therapy, such as accuracy and precision, repeatability and reproducibility, geometric distortion, and motion robustness. Finally, future directions for MRF application in radiation therapy are discussed.

Artificial intelligence for assisting cancer diagnosis and treatment in the era of precision medicine.

Over the past decade, artificial intelligence (AI) has contributed substantially to the resolution of various medical problems, including cancer. Deep learning (DL), a subfield of AI, is characterized by its ability to perform automated feature extraction and has great power in the assimilation and evaluation of large amounts of complicated data. On the basis of a large quantity of medical data and novel computational technologies, AI, especially DL, has been applied in various aspects of oncology research and has the potential to enhance cancer diagnosis and treatment. These applications range from early cancer detection, diagnosis, classification and grading, molecular characterization of tumors, prediction of patient outcomes and treatment responses, personalized treatment, automatic radiotherapy workflows, novel anti‐cancer drug discovery, and clinical trials. In this review, we introduced the general principle of AI, summarized major areas of its application for cancer diagnosis and treatment, and discussed its future directions and remaining challenges. As the adoption of AI in clinical use is increasing, we anticipate the arrival of AI‐powered cancer care.

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

Background and purposeWhole bladder radiotherapy is challenging due to inter- and intrafraction size and shape changes. To account for these changes, currently a Library of Plans (LoP) technique is often applied, but daily adaptive radiotherapy is also increasingly becoming available. The aim of this study was to compare LoP with two magnetic resonance imaging guided radiotherapy (MRgRT) strategies by comparing target coverage and volume of healthy tissue inside the planning target volume (PTV) for whole bladder treatments.Methods and materialsData from 25 MRgRT lymph node oligometastases treatments (125 fractions) were used, with three MRI scans acquired at each fraction at 0, 15 and 30 min. Bladders were delineated and used to evaluate three strategies: 1) LoP with two plans for a 15 min fraction, 2) MRgRT15min for a 15 min fraction and 3) MRgRT30min for a 30 min fraction. The volumes of healthy tissue inside and bladder outside the PTV were analyzed on the simulated post-treatment images.ResultsMRgRT30min had 120% and 121% more healthy tissue inside the PTV than LoP and MRgRT15min. For LoP slightly more target outside the PTV was found than for MRgRT30min and MRgRT15min, with median 0% (range 0–23%) compared to 0% (0–20%) and 0% (0–10%), respectively.ConclusionsTaking into account both target coverage and volume of healthy tissue inside the PTV, MRgRT15min performed better than LoP and MRgRT30min for whole bladder treatments. A 15 min daily adaptive radiotherapy workflow is needed to potentially benefit from replanning compared to LoP.

Computational ECG mapping and respiratory gating to optimize stereotactic ablative radiotherapy workflow for refractory ventricular tachycardia.

BackgroundStereotactic ablative radiotherapy (SAbR) is an emerging therapy for refractory ventricular tachycardia (VT). However, the current workflow is complicated, and the precision and safety in patients with significant cardiorespiratory motion and VT targets near the stomach may be suboptimal.ObjectiveWe hypothesized that automated 12-lead electrocardiogram (ECG) mapping and respiratory-gated therapy may improve the ease and precision of SAbR planning and facilitate safe radiation delivery in patients with refractory VT.MethodsConsecutive patients with refractory VT were studied at 2 hospitals. VT exit sites were localized using a 3-D computational ECG algorithm noninvasively and compared to available prior invasive mapping. Radiotherapy (25 Gy) was delivered at end-expiration when cardiac respiratory motion was ≥0.6 cm or targets were ≤2 cm from the stomach.ResultsIn 6 patients (ejection fraction 29% ± 13%), 4.2 ± 2.3 VT morphologies per patient were mapped. Overall, 7 out of 7 computational ECG mappings (100%) colocalized to the identical cardiac segment when prior invasive electrophysiology study was available. Respiratory gating was associated with smaller planning target volumes compared to nongated volumes (71 ± 7 vs 153 ± 35 cc, P < .01). In 2 patients with inferior wall VT targets close to the stomach (6 mm proximity) or significant respiratory motion (22 mm excursion), no GI complications were observed at 9- and 12-month follow-up. Implantable cardioverter-defibrillator shocks decreased from 23 ± 12 shocks/patient to 0.67 ± 1.0 (P < .001) post-SAbR at 6.0 ± 4.9 months follow-up.ConclusionsA workflow including computational ECG mapping and protocol-guided respiratory gating is feasible, is safe, and may improve the ease of SAbR planning. Studies to validate this workflow in larger populations are required.

Increasing Demand on Human Capital and Resource Utilization in Radiation Therapy: The Past Decade

To quantify the change resource utilization in radiation therapy in the context of advancing technologies and techniques over the last decade. Prospectively, the time to complete radiation therapy workflow tasks was captured between January 1, 2020, and December 31, 2020. The institutional task workflows are specific to each technique and broadly organized into 4 categories: 3-dimenstional conformal radiation therapy, intensity modulated radiation therapy, volumetric modulated arc therapy simple, and volumetric modulated arc therapy complex. These discipline-specific task times were used to quantify a resource utilization factor, which is the median time taken to complete all tasks for each category divided by the median time for 3-dimensional conformal radiation therapy treatments. Retrospectively, all plans treated between January 1, 2012, and December 31, 2019, were quantified and categorized. The resource factor was applied to determine resource utilization. For context, institutional staffing levels were captured across the same decade for medical dosimetrists, medical physicists, and radiation oncologists. This analysis includes 30,229 patient plans in the retrospective data set and 4747 patient plans in the prospective data set. This analysis demonstrates that over this period, patient numbers increased by approximately 45%, whereas time-based human resources increased by almost 150%. The resource allocation factors for 3-dimenstional conformal radiation therapy, intensity modulated radiation therapy, volumetric modulated arc therapy simple, and volumetric arc therapy complex were 1.0, 2.4, 2.9, and 4.3, respectively. Across the 3 disciplines, staffing levels increased from 15 to 17 (13%) for medical dosimetrists, from 10 to 13 (30%) for medical physicists, and from 16 to 23 (44%) for radiation oncologists. This work demonstrates the increase in resource utilization due to the introduction of advanced technologies and changes in radiation therapy techniques over the past decade. Human resource utilization is the predominant factor and should be considered with increasing patient volume for operational planning.

Single-isocenter stereotactic non-coplanar arc treatment of 200\xa0patients with brain metastases: multileaf collimator size and setup uncertainties

PurposeThe purpose of this study was to evaluate our 2 years’ experience with single-isocenter, non-coplanar, volumetric modulated arc therapy (VMAT) for brain metastasis (BM) stereotactic radiosurgery (SRS).MethodsA total of 202 patients treated with the VMAT SRS solution were analyzed retrospectively. Plan quality was assessed for 5 mm (120) and 2.5 mm (high-definition, HD) central leaf width multileaf collimators (MLCs). For BMs at varying distances from the plan isocenter, the geometric offset from the ideal position for two image-guided radiotherapy workflows was calculated. In the workflow with ExacTrac (BrainLAB, München, Germany; W‑ET), patient positioning errors were corrected at each couch rotation. In the workflow without ExacTrac (W-noET), only the initial patient setup correction was considered. The dose variation due to rotational errors was simulated for multiple-BM plans with the HD-MLC.ResultsPlan conformity and quality assurance were equivalent for plans delivered with the two MLCs while the HD-MLC plans provided better healthy brain tissue (BmP) sparing. 95% of the BMs had residual intrafractional setup errors ≤ 2 mm for W‑ET and 68% for W‑noET. For small BM (≤1 cc) situated >3 cm from the plan isocenter, the dose received by 95% of the BM decreased in median (interquartile range) by 6.3% (2.8–8.8%) for a 1-degree rotational error.ConclusionThis study indicates that the HD-MLC is advantageous compared to the 120-MLC for sparing healthy brain tissue. When a 2-mm margin is applied, W‑noET is sufficient to ensure coverage of BM situated ≤ 3 cm of the plan isocenter, while for BM further away, W‑ET is recommended.

Prospects for daily online adaptive radiotherapy via ethos for prostate cancer patients without nodal involvement using unedited CBCT auto-segmentation.

Purpose Implementing new online adaptive radiation therapy technologies is challenging because extra clinical resources are required particularly expert contour review. Here, we provide the first assessment of Varian's Ethos™ adaptive platform for prostate cancer using no manual edits after auto‐segmentation to minimize this impact on clinical efficiency.MethodsTwenty‐five prostate patients previously treated at our clinic were re‐planned using an Ethos™ emulator. Clinical target volumes (CTV) included intact prostate and proximal seminal vesicles. The following clinical margins were used: 3 mm posterior, 5 mm left/right/anterior, and 7 mm superior/inferior. Adapted plans were calculated for 10 fractions per patient using Ethos's auto‐segmentation and auto‐planning workflow without manual contouring edits. Doses and auto‐segmented structures were exported to our clinical treatment planning system where contours were modified as needed for all 250 CTVs and organs‐at‐risk. Dose metrics from adapted plans were compared to unadapted plans to evaluate CTV and OAR dose changes.ResultsOverall 96% of fractions required auto‐segmentation edits, although corrections were generally minor (<10% of the volume for 70% of CTVs, 88% of bladders, and 90% of rectums). However, for one patient the auto‐segmented CTV failed to include the superior portion of prostate that extended into the bladder at all 10 fractions resulting in under‐contouring of the CTV by 31.3% ± 6.7%. For the 24 patients with minor auto‐segmentation corrections, adaptation improved CTV D98% by 2.9% ± 5.3%. For non‐adapted fractions where bladder or rectum V90% exceeded clinical thresholds, adaptation reduced them by 13.1% ± 1.0% and 6.5% ± 7.3%, respectively.Conclusion For most patients, Ethos's online adaptive radiation therapy workflow improved CTV D98% and reduced normal tissue dose when structures would otherwise exceed clinical thresholds, even without time‐consuming manual edits. However, for one in 25 patients, large contour edits were required and thus scrutiny of the daily auto‐segmentation is necessary and not all patients will be good candidates for adaptation.

Investigation of the clinical inter-observer bias in prostate fiducial marker image registration between CT and MR images

Background and purposeInter-modality image registration between computed tomography (CT) and magnetic resonance (MR) images is associated with systematic uncertainties and the magnitude of these uncertainties is not well documented. The purpose of this study was to investigate the potential uncertainty of gold fiducial marker (GFM) registration for localized prostate cancer and to estimate the inter-observer bias in a clinical setting.Methods Four experienced observers registered CT and MR images for 42 prostate cancer patients. Manual GFM identification was followed by a landmark-based registration. The absolute difference between observers in GFM identification and the displacement of the clinical target volume (CTV) was investigated. The CTV center of mass (CoM) vector displacements, DICE-index and Hausdorff distances for the observer registrations were compared against a clinical baseline registration. The time allocated for the manual registrations was compared.ResultsAbsolute difference in GFM identification between observers ranged from 0.0 to 3.0 mm. The maximum CTV CoM displacement from the clinical baseline was 3.1 mm. Displacements larger than or equal to 1 mm, 2 mm and 3 mm were 46%, 18% and 4%, respectively. No statistically significant difference was detected between observers in terms of CTV displacement. Median DICE-index and Hausdorff distance for the CTV, with their respective ranges were 0.94 [0.70–1.00] and 2.5 mm [0.7–8.7].ConclusionsRegistration of CT and MR images using GFMs for localized prostate cancer patients was subject to inter-observer bias on an individual patient level. A CTV displacement as large as 3 mm occurred for individual patients. These results show that GFM registration in a clinical setting is associated with uncertainties, which motivates the removal of inter-modality registrations in the radiotherapy workflow and a transition to an MRI-only workflow for localized prostate cancer.

Multi-parametric MRI (mpMRI) for treatment response assessment of radiation therapy.

Magnetic resonance imaging (MRI) plays an important role in the modern radiation therapy (RT) workflow. In comparison with computed tomography (CT) imaging, which is the dominant imaging modality in RT, MRI possesses excellent soft-tissue contrast for radiographic evaluation. Based on quantitative models, MRI can be used to assess tissue functional and physiological information. With the developments of scanner design, acquisition strategy, advanced data analysis, and modeling, multiparametric MRI (mpMRI), a combination of morphologic and functional imaging modalities, has been increasingly adopted for disease detection, localization, and characterization. Integration of mpMRI techniques into RT enriches the opportunities to individualize RT. In particular, RT response assessment using mpMRI allows for accurate characterization of both tissue anatomical and biochemical changes to support decision-making in monotherapy of radiation treatment and/or systematic cancer management. In recent years, accumulating evidence have, indeed, demonstrated the potentials of mpMRI in RT response assessment regarding patient stratification, trial benchmarking, early treatment intervention, and outcome modeling. Clinical application of mpMRI for treatment response assessment in routine radiation oncology workflow, however, is more complex than implementing an additional imaging protocol; mpMRI requires additional focus on optimal study design, practice standardization, and unified statistical reporting strategy to realize its full potential in the context of RT. In this article, the mpMRI theories, including image mechanism, protocol design, and data analysis, will be reviewed with a focus on the radiation oncology field. Representative works will be discussed to demonstrate how mpMRI can be used for RT response assessment. Additionally, issues and limits of current works, as well as challenges and potential future research directions, will also be discussed.

Technical Note: On the feasibility of performing dosimetry in target and organ at risk using polymer dosimetry gel and thermoluminescence detectors in an anthropomorphic, deformable, and multimodal pelvis phantom.

To assess the feasibility of performing dose measurements in the target (prostate) and an adjacent organ at risk (rectum) using polymer dosimetry gel and thermoluminescence detectors (TLDs) in an anthropomorphic, deformable, and multimodal pelvis phantom (ADAM PETer). The 3D printed prostate organ surrogate of the ADAM PETer phantom was filled with polymer dosimetry gel. Nine TLD600 (LiF:Mg,Ti) were installed in 3×3 rows on a specifically designed 3D-printed TLD holder. The TLD holder was inserted into the rectum at the level of the prostate and fixed by a partially inflated endorectal balloon. Computed tomography (CT) images were taken and treatment planning was performed. A prescribed dose of 4.5Gy was delivered to the planning target volume (PTV). The doses measured by the dosimetry gel in the prostate and the TLDs in the rectum ("measured dose") were compared to the doses calculated by the treatment planning system ("planned dose") on a voxel-by-voxel basis. In the prostate organ surrogate, the 3D-γ-index was 97.7% for the 3% dose difference and 3mm distance to agreement criterium. In the center of the prostate organ surrogate, measured and planned doses showed only minor deviations (<0.1Gy, corresponding to a percentage error of 2.22%). On the edges of the prostate, slight differences between planned and measured doses were detected with a maximum deviation of 0.24Gy, corresponding to 5.3% of the prescribed dose. The difference between planned and measured doses in the TLDs was on average 0.08Gy (range: 0.02-0.21Gy), corresponding to 1.78% of the prescribed dose (range: 0.44%-4.67%). The present study demonstrates the feasibility of using polymer dosimetry gel and TLDs for 3D and 1D dose measurements in the prostate and the rectum organ surrogates in an anthropomorphic, deformable and multimodal phantom. The described methodology might offer new perspectives for end-to-end tests in image-guided adaptive radiotherapy workflows.

PO-1670 Clinical implementation of MRI-only radiotherapy workflow for prostate cancer with a standard linac

PO-1670 Clinical implementation of MRI-only radiotherapy workflow for prostate cancer with a standard linac

Automated treatment planning of prostate stereotactic body radiotherapy with focal boosting on a fast-rotating O-ring linac: Plan quality comparison with C-arm linacs.

PurposeThe integration of auto‐segmentation and automated treatment planning methods on a fast‐rotating O‐ring linac may improve the time efficiency of online adaptive radiotherapy workflows. This study investigates whether automated treatment planning of prostate SBRT with focal boosting on the O‐ring linac could generate plans that are of similar quality as those obtained through manual planning on clinical C‐arm linacs.MethodsFor 20 men with prostate cancer, reference treatment plans were generated on a TrueBeam STx C‐arm linac with HD120 MLC and a TrueBeam C‐arm linac with Millennium 120 MLC using 6 MV flattened dual arc VMAT. Manual planning on the Halcyon fast‐rotating O‐ring linac was performed using 6 MV FFF dual arc VMAT (HA2‐DL10) and triple arc VMAT (HA3‐DL10) to investigate the performance of the dual‐layer MLC system. Automated planning was performed for triple arc VMAT on the Halcyon linac (ET3‐DL10) using the automated planning algorithms of Ethos Treatment Planning. The prescribed dose was 35 Gy to the prostate and 30 Gy to the seminal vesicles in five fractions. The iso‐toxic focal boost to the intraprostatic tumor nodule(s) was aimed to receive up to 50 Gy. Plan deliverability was verified using portal image dosimetry measurements.ResultsCompared to the C‐arm linacs, ET3‐DL10 shows increased seminal vesicles PTV coverage (D99%) and reduced high‐dose spillage to the bladder (V37Gy) and urethra (D0.035cc) but this came at the cost of increased high‐dose spillage to the rectum (V38Gy) and a higher intermediate dose spillage (D2cm). No statistically significant differences were found when benchmarking HA2‐DL10 and HA3‐DL10 with the C‐arm linacs. All plans passed the patient‐specific QA tolerance limit.ConclusionsAutomated planning of prostate SBRT with focal boosting on the fast‐rotating O‐ring linac is feasible and achieves similar plan quality as those obtained on clinical C‐arm linacs using manual planning.

Fully automated segmentation of brain tumor from multiparametric MRI using 3D context deep supervised U-Net.

Owing to histologic complexities of brain tumors, its diagnosis requires the use of multimodalities to obtain valuable structural information so that brain tumor subregions can be properly delineated. In current clinical workflow, physicians typically perform slice-by-slice delineation of brain tumor subregions, which is a time-consuming process and also more susceptible to intra- and inter-rater variabilities possibly leading to misclassification. To deal with this issue, this study aims to develop an automatic segmentation of brain tumor in MR images using deep learning. In this study, we develop a context deep-supervised U-Net to segment brain tumor subregions. A context block which aggregates multiscale contextual information for dense segmentation was proposed. This approach enlarges the effective receptive field of convolutional neural networks, which, in turn, improves the segmentation accuracy of brain tumor subregions. We performed the fivefold cross-validation on the Brain Tumor Segmentation Challenge (BraTS) 2020 training dataset. The BraTS 2020 testing datasets were obtained via BraTS online website as a hold-out test. For BraTS, the evaluation system divides the tumor into three regions: whole tumor (WT), tumor core (TC), and enhancing tumor (ET). The performance of our proposed method was compared against two state-of-the-arts CNN networks in terms of segmentation accuracy via Dice similarity coefficient (DSC) and Hausdorff distance (HD). The tumor volumes generated by our proposed method were compared with manually contoured volumes via Bland-Altman plots and Pearson analysis. The proposed method achieved the segmentation results with a DSC of 0.923±0.047, 0.893±0.176, and 0.846±0.165 and a 95% HD95 of 3.946±7.041, 3.981±6.670, and 10.128±51.136mm on WT, TC, and ET, respectively. Experimental results demonstrate that our method achieved comparable to significantly (p<0.05) better segmentation accuracies than other two state-of-the-arts CNN networks. Pearson correlation analysis showed a high positive correlation between the tumor volumes generated by proposed method and manual contour. Overall qualitative and quantitative results of this work demonstrate the potential of translating proposed technique into clinical practice for segmenting brain tumor subregions, and further facilitate brain tumor radiotherapy workflow.

Volumetric modulated arc therapy dose prediction and deliverable treatment plan generation for prostate cancer patients using a densely connected deep learning model.

Radiation therapy treatment planning is a manual, time-consuming task that might be accelerated using machine learning algorithms. In this study, we aimed to evaluate if a triplet-based deep learning model can predict volumetric modulated arc therapy (VMAT) dose distributions for prostate cancer patients. A modified U-Net was trained on triplets, a combination of three consecutive image slices and corresponding segmentations, from 160 patients, and compared to a baseline U-Net. Dose predictions from 17 test patients were transformed into deliverable treatment plans using a novel planning workflow. The model achieved a mean absolute dose error of 1.3%, 1.9%, 1.0% and ≤ 2.6% for clinical target volume (CTV) CTV_D100%, planning target volume (PTV) PTV_D98%, PTV_D95% and organs at risk (OAR) respectively, when compared to the clinical ground truth (GT) dose distributions. All predicted distributions were successfully transformed into deliverable treatment plans and tested on a phantom, resulting in a passing rate of 100% (global gamma, 3%, 2mm, 15% cutoff). The dose difference between deliverable treatment plans and GT dose distributions was within 4.4%. The difference between the baseline model and our improved model was statistically significant (p<0.05) for CVT_D100%, PTV_D98% and PTV_D95%. Triplet-based training improved VMAT dose distribution predictions when compared to 2D. Dose predictions were successfully transformed into deliverable treatment plans using our proposed treatment planning procedure. Our method may automate parts of the workflow for external beam prostate radiation therapy and improve the overall treatment speed and plan quality.