Evaluation of a workflow for cone-beam CT-guided online adaptive palliative radiotherapy planned using diagnostic CT scans.

PURPOSE: Single-visit radiotherapy (RT) is beneficial for patients requiring pain control and can limit interruptions to systemic treatments. However, the requirement for a dedicated planning CT (pCT)-scan can result in treatment delays. We developed a workflow involving preplanning on available diagnostic CT (dCT) imaging, followed by online plan adaption using a cone-beam CT (CBCT)-scan prior to RT-delivery, in order to account for any changes in anatomy and target position. METHODS: Patients previously treated with palliative RT for bone metastases were selected from our hospital database. Patient dCT-images were deformed to treatment CBCTs in the Ethos platform (Varian Medical Systems) and a synthetic CT (sCT) generated. Treatment quality was analyzed by comparing a coverage of the V95% of the planning/clinical target volume and different organ-at-risk (OAR) doses between adapted and initial clinical treatment plans. Doses were recalculated on the CBCT and sCT in a separate treatment planning system. Adapted plan doses were measured on-couch using an anthropomorphic phantom with a Gafchromic EBT3 dosimetric film and compared to dose calculations. RESULTS: All adapted treatment plans met the clinical goals for target and OARs and outperformed the original treatment plans calculated on the (daily) sCT. Differences in V95% of the target volume coverage between the initial and adapted treatments were <0.2%. Dose recalculations on CBCT and sCT were comparable, and the average gamma pass rate (3%/2 mm) of dosimetric measurements was 98.8%. CONCLUSIONS: Online daily adaptive RT using dCTs instead of a dedicated pCT is feasible using the Ethos platform. This workflow has now been implemented clinically.

Investigating the potential of deep learning for patient-specific quality assurance of salivary gland contours using EORTC-1219-DAHANCA-29 clinical trial data.

INTRODUCTION: Manual quality assurance (QA) of radiotherapy contours for clinical trials is time and labor intensive and subject to inter-observer variability. Therefore, we investigated whether deep-learning (DL) can provide an automated solution to salivary gland contour QA. MATERIAL AND METHODS: DL-models were trained to generate contours for parotid (PG) and submandibular glands (SMG). Sørensen-Dice coefficient (SDC) and Hausdorff distance (HD) were used to assess agreement between DL and clinical contours and thresholds were defined to highlight cases as potentially sub-optimal. 3 types of deliberate errors (expansion, contraction and displacement) were gradually applied to a test set, to confirm that SDC and HD were suitable QA metrics. DL-based QA was performed on 62 patients from the EORTC-1219-DAHANCA-29 trial. All highlighted contours were visually inspected. RESULTS: Increasing the magnitude of all 3 types of errors resulted in progressively severe deterioration/increase in average SDC/HD. 19/124 clinical PG contours were highlighted as potentially sub-optimal, of which 5 (26%) were actually deemed clinically sub-optimal. 2/19 non-highlighted contours were false negatives (11%). 15/69 clinical SMG contours were highlighted, with 7 (47%) deemed clinically sub-optimal and 2/15 non-highlighted contours were false negatives (13%). For most incorrectly highlighted contours causes for low agreement could be identified. CONCLUSION: Automated DL-based contour QA is feasible but some visual inspection remains essential. The substantial number of false positives were caused by sub-optimal performance of the DL-model. Improvements to the model will increase the extent of automation and reliability, facilitating the adoption of DL-based contour QA in clinical trials and routine practice.

Optimizing SABR delivery for synchronous multiple lung tumors using volumetric-modulated arc therapy.

BACKGROUND: Volumetric-modulated arc therapy (VMAT) delivery for stereotactic ablative radiotherapy (SABR) of multiple lung tumors allows for faster treatments. We report on clinical outcomes and describe a general approach for treatment planning. MATERIAL AND METHODS: Patients undergoing multi iso-center VMAT-based SABR for ≥2 lung lesions between 2009 and 2014 were identified from the VU University Medical Center and London Health Sciences Centre. Patients were eligible if the start date of the SABR treatment for the different lesions was within a time range of 30 days. SABR was delivered using separate iso-centers for lesions at a substantial distance from each other. Tumors were either treated with a single fraction of 34 Gy, or using three risk-adapted dose-fractionation schemes, namely three fractions of 18 Gy, five fractions of 11 Gy, or eight fractions of 7.5 Gy, depending on the tumor size and the location. Multivariable analysis was performed to assess factors predictive of clinical outcomes. RESULTS: Of 84 patients (188 lesions) identified, 46% were treated for multiple metastases and 54% for multiple primary NSCLC. About 97% were treated for two or three lesions, and 56% had bilateral disease. After a median follow-up of 28 months, median overall survival (OS) for primary tumors was 27.6 months, and not reached for metastatic lesions (p = .028). Grade ≥3 toxicity was observed in 2% of patients. Multivariable analysis showed that grade 2 or higher radiation pneumonitis (n = 9) was best predicted by a total lung V35Gy of ≥6.5% (in 2Gy/fraction equivalent) (p = .007). CONCLUSION: Severe toxicity was uncommon following SABR using VMAT for up to three lung tumors. Further investigations of planning parameters are needed in patients presenting with more lesions.

Roll and pitch set-up errors during volumetric modulated arc delivery: can adapting gantry and collimator angles compensate?

PURPOSE: The purpose of this work was to investigate whether adapting gantry and collimator angles can compensate for roll and pitch setup errors during volumetric modulated arc therapy (VMAT) delivery. METHODS: Previously delivered clinical plans for locally advanced head-and-neck (H&N) cancer (n = 5), localized prostate cancer (n = 2), and whole brain with simultaneous integrated boost to 5 metastases (WB + 5M, n = 1) were used for this study. Known rigid rotations were introduced in the planning CT scans. To compensate for these, in-house software was used to adapt gantry and collimator angles in the plan. Doses to planning target volumes (PTV) and critical organs at risk (OAR) were calculated with and without compensation and compared with the original clinical plan. Measurements in the sagittal plane in a polystyrene phantom using radiochromic film were compared by gamma (γ) evaluation for 2 H&N cancer patients. RESULTS: For H&N plans, the introduction of 2°-roll and 3°-pitch rotations reduced mean PTV coverage from 98.7 to 96.3%. This improved to 98.1% with gantry and collimator compensation. For prostate plans respective figures were 98.4, 97.5, and 98.4%. For WB + 5M, compensation worked less well, especially for smaller volumes and volumes farther from the isocenter. Mean comparative γ evaluation (3%, 1 mm) between original and pitched plans resulted in 86% γ < 1. The corrected plan restored the mean comparison to 96% γ < 1. CONCLUSION: Preliminary data suggest that adapting gantry and collimator angles is a promising way to correct roll and pitch set-up errors of < 3° during VMAT for H&N and prostate cancer.

Is there a preferred IMRT technique for left-breast irradiation?

Not all clinics have breath-hold radiotherapy available for left-breast irradiation. However intensity-modulated radiotherapy (IMRT) has also been advocated as a means of lowering heart doses. There is currently no large-scale, long-term follow-up data after breast IMRT and, since dose distributions may differ from classic tangent-based radiotherapy, caution is needed to avoid unexpected worsening of the late toxicity profile. We compared four IMRT techniques for free-breathing left-breast irradiation. Consistent with the aforementioned concerns, our goal in planning was to prioritize organ at risk (OAR) sparing in a way that mimicked tangent-based radiotherapy. Ten simultaneous integrated boost treatment plans (PTVelective = 15 × 2.67 Gy; PTVboost = 15 × 3.35 Gy) were created using 1) hybrid-IMRT (H-IMRT), 2) full IMRT (F-IMRT), and 3) volumetric-modulated arc therapy with two partial arcs (2ARC) and 4) six partial arcs (6ARC). Reduction in OAR mean and low dose was prioritized. End-points included OAR sparing (e.g., heart, left anterior descending artery [LAD+3 mm], lungs, and contralateral breast) and PTV coverage/dose homogeneity. Under these conditions we found the following: 1) H-IMRT provided the best mean and low dose OAR sparing, PTVelective coverage (mean V95% = 98%), PTVboost coverage (V95% = 98%), and PTV homogeneity. However, it delivered most intermediate-high dose to the heart, LAD+3 mm and ipsilateral lung; 2) 6ARC had the best intermediate-high dose sparing, followed by F-IMRT, but this was at the expense of more dose in the contralateral lung and breast and worse PTV coverage (PTVelective mean V95% = 96%/97% and PTVboost mean V95% = 91%/96% for 6ARC/F-IMRT). When trying to spare mean and low dose to OARs, the preferred IMRT technique for left-breast irradiation without breath-hold was H-IMRT. This is currently the standard solution in our institution for left-breast radiotherapy under free-breathing and breath-hold conditions.

Changes in non-surgical management of stage III non-small cell lung cancer at a single institution between 2003 and 2010.

BACKGROUND: Concurrent chemo-radiotherapy (CON-CRT) is recommended for selected patients with stage III non-small cell lung cancer (NSCLC), but utilization varies. We assessed the response to national guidelines introduced in 2004 and the impact on outcomes. MATERIAL AND METHODS: Retrospective study of stage III NSCLC patients treated with radical intent non-surgical treatment during 2003-2010 in a university medical center characterized by multidisciplinary assessment, routine use of four-dimensional computed tomography for radiotherapy planning, and rapid implementation of radiotherapy advances. RESULTS: Between 2003 and 2010, 319/435 (73%) patients with stage III NSCLC received (chemo) radiotherapy. The number receiving CON-CRT in successive two-year periods increased from 13/48 (27%) - 40/80 (50%) - 63/90 (70%), to 74/101 (73%). Median overall survival (OS) from start of radiotherapy was 18.6 months for CON-CRT (190/319) and 17.4 months for sequential (SEQ), typically hypofractionated, CRT (90/319) (p = 0.78). Eleven months OS with radiotherapy alone (39/319) was significantly shorter (p = 0.006). OS did not differ between the four periods (p = 0.87). CON-CRT was not over-represented in the 16% of patients dying within five months of starting radiotherapy. CONCLUSIONS: Between 2003 and 2010, CON-CRT for stage III NSCLC was rapidly and safely increased. However, OS did not increase and, as practiced, did not differ between CON- or SEQ-CRT.

Spine Stereotactic Body Radiation Therapy Without Immobilization: Detailed Analysis of Intrafraction Motion Using High-Frequency kV Imaging During Irradiation.

PURPOSE: Spine stereotactic body radiation therapy (SBRT) requires high positioning accuracy and a stable patient to maximize target coverage and reduce excessive irradiation to organs at risk. Positional verification during spine SBRT delivery helps to ensure accurate positioning for all patients. We report our experience with noninvasive 3-dimensional target position monitoring during volumetric modulated arc therapy of spine metastases in nonimmobilized patients positioned using only a thin mattress and simple arm and knee supports. METHODS AND MATERIALS: Fluoroscopic planar kV images were acquired at 7 frames/s using the on-board imaging system during volumetric modulated arc therapy spine SBRT. Template matching and triangulation were used to track the target in vertical, longitudinal, and lateral directions. If the tracking trace deviated >1 mm from the planned position in ≥1 direction, treatment was manually interrupted and 6-dimensional cone beam computed tomography (CBCT)-based couch correction was performed. Tracking data were used to retrospectively analyze the target position. Positional data, agreement with CBCT, correlation between position of the couch and direction of any positional correction, and treatment times were analyzed. RESULTS: In total, 175 fractions were analyzed. Delivery was interrupted 83 times in 66 fractions for a deviation >1 mm. In 97% of cases the difference between tracking data and subsequent clinical shift performed after the CBCT match was ≤0.5 mm. Lateral/longitudinal shift performed after intervention correlated with the couch roll/pitch at the start of treatment (correlation coefficient, -0.63/0.53). Mean (SD; range) time between start of first imaging and end of the last arc was 15.2 minutes (5.1; 7.6-36.3). CONCLUSIONS: Spine tracking during irradiation can be used to prompt an intervention CBCT scan and repositioning so that a spine SBRT target deviates by ≤1 mm from the planned position, even in nonimmobilized patients. kV tracking and CBCT are in good agreement. The data support verification CBCT after all 6 degrees-of-freedom positional corrections in nonimmobilized spine SBRT patients.

Technical note: Dosimetry and FLASH potential of UHDR proton PBS for small lung tumors: Bragg-peak-based delivery versus transmission beam and IMPT.

BACKGROUND: High-energy transmission beams (TBs) are currently the main delivery method for proton pencil beam scanning ultrahigh dose-rate (UHDR) FLASH radiotherapy. TBs place the Bragg-peaks behind the target, outside the patient, making delivery practical and achievement of high dose-rates more likely. However, they lead to higher integral dose compared to conventional intensity-modulated proton therapy (IMPT), in which Bragg-peaks are placed within the tumor. It is hypothesized that, when energy changes are not required and high beam currents are possible, Bragg-peak-based beams can not only achieve more conformal dose distributions than TBs, but also have more FLASH-potential. PURPOSE: This works aims to verify this hypothesis by taking three different Bragg-peak-based delivery techniques and comparing them with TB and IMPT-plans in terms of dosimetry and FLASH-potential for single-fraction lung stereotactic body radiotherapy (SBRT). METHODS: For a peripherally located lung target of various sizes, five different proton plans were made using "matRad" and inhouse-developed algorithms for spot/energy-layer/beam reduction and minimum monitor unit maximization: (1) IMPT-plan, reference for dosimetry, (2) TB-plan, reference for FLASH-amount, (3) pristine Bragg-peak plan (non-depth-modulated Bragg-peaks), (4) Bragg-peak plan using generic ridge filter, and (5) Bragg-peak plan using 3D range-modulated ridge filter. RESULTS: Bragg-peak-based plans are able to achieve sufficient plan quality and high dose-rates. IMPT-plans resulted in lowest OAR-dose and integral dose (also after a FLASH sparing-effect of 30%) compared to both TB-plans and Bragg-peak-based plans. Bragg-peak-based plans vary only slightly between themselves and generally achieve lower integral dose than TB-plans. However, TB-plans nearly always resulted in lower mean lung dose than Bragg-peak-based plans and due to a higher amount of FLASH-dose for TB-plans, this difference increased after including a FLASH sparing-effect. CONCLUSION: This work indicates that there is no benefit in using Bragg-peak-based beams instead of TBs for peripherally located, UHDR stereotactic lung radiotherapy, if lung dose is the priority.

Comparable cell survival between high dose rate flattening filter free and conventional dose rate irradiation.

PURPOSE: Investigation of clonogenic cell survival and cell proliferation following single dose and fractionated delivery of high dose rate flattening filter free (FFF) irradiation compared to conventional dose rates. MATERIAL AND METHODS: The human astrocytoma D384, glioma T98 and lung carcinoma SW1573 cell lines were irradiated using either a single dose (0-12 Gy) or a fractionated protocol of 5 daily fractions of 2 Gy (D384) or 3 Gy (SW1573). Cells were irradiated inside a phantom using fixed gantry beams of a linear accelerator. A sliding window technique created homogeneous dose distributions over the surface of the cell cultures. Irradiations using standard beams (6 MV, 600 MU/min.) and high dose rate FFF beams (10 MV, 2400 MU/min.) were compared. Cell survival was determined by clonogenic assay. In the fractionated irradiation set-up, the number of clonogenic cells was estimated by including tumor cell proliferation during the overall treatment time in the analysis. RESULTS: All cell lines showed equal cell survival following irradiation using either the FFF beams or conventional flattened (FF) beams. This was observed after single dose exposure (0-12 Gy) as well as after fractionated irradiation (p = 0.08 for D384 and 0.20 for SW1373 cell lines). CONCLUSION: FFF irradiation with a dose rate of 2400 MU/min and four times higher dose per pulse compared to irradiation with FF beams did not change cell survival for three human cancer cell lines up to a fraction dose of 12 Gy compared to irradiation using FF beams.

Impact of imperfection in medical imaging data on deep learning-based segmentation performance: An experimental study using synthesized data.

BACKGROUND: Clinical data used to train deep learning models are often not clean data. They can contain imperfections in both the imaging data and the corresponding segmentations. PURPOSE: This study investigates the influence of data imperfections on the performance of deep learning models for parotid gland segmentation. This was done in a controlled manner by using synthesized data. The insights this study provides may be used to make deep learning models better and more reliable. METHODS: The data were synthesized by using the clinical segmentations, creating a pseudo ground-truth in the process. Three kinds of imperfections were simulated: incorrect segmentations, low image contrast, and artifacts in the imaging data. The severity of each imperfection was varied in five levels. Models resulting from training sets from each of the five levels were cross-evaluated with test sets from each of the five levels. RESULTS: Using synthesized data led to almost perfect parotid gland segmentation when no error was added. Lowering the quality of the parotid gland segmentations used for training substantially lowered the model performance. Additionally, lowering the image quality of the training data by decreasing the contrast or introducing artifacts made the resulting models more robust to data containing those respective kinds of data imperfection. CONCLUSION: This study demonstrated the importance of good-quality segmentations for deep learning training and it shows that using low-quality imaging data for training can enhance the robustness of the resulting models.

Dose distribution prediction for head-and-neck cancer radiotherapy using a generative adversarial network: influence of input data.

Purpose: A three-dimensional deep generative adversarial network (GAN) was used to predict dose distributions for locally advanced head and neck cancer radiotherapy. Given the labor- and time-intensive nature of manual planning target volume (PTV) and organ-at-risk (OAR) segmentation, we investigated whether dose distributions could be predicted without the need for fully segmented datasets. Materials and methods: GANs were trained/validated/tested using 320/30/35 previously segmented CT datasets and treatment plans. The following input combinations were used to train and test the models: CT-scan only (C); CT+PTVboost/elective (CP); CT+PTVs+OARs+body structure (CPOB); PTVs+OARs+body structure (POB); PTVs+body structure (PB). Mean absolute errors (MAEs) for the predicted dose distribution and mean doses to individual OARs (individual salivary glands, individual swallowing structures) were analyzed. Results: For the five models listed, MAEs were 7.3 Gy, 3.5 Gy, 3.4 Gy, 3.4 Gy, and 3.5 Gy, respectively, without significant differences among CP-CPOB, CP-POB, CP-PB, among CPOB-POB. Dose volume histograms showed that all four models that included PTV contours predicted dose distributions that had a high level of agreement with clinical treatment plans. The best model CPOB and the worst model PB (except model C) predicted mean dose to within ±3 Gy of the clinical dose, for 82.6%/88.6%/82.9% and 71.4%/67.1%/72.2% of all OARs, parotid glands (PG), and submandibular glands (SMG), respectively. The R2 values (0.17/0.96/0.97/0.95/0.95) of OAR mean doses for each model also indicated that except for model C, the predictions correlated highly with the clinical dose distributions. Interestingly model C could reasonably predict the dose in eight patients, but on average, it performed inadequately. Conclusion: We demonstrated the influence of the CT scan, and PTV and OAR contours on dose prediction. Model CP was not statistically different from model CPOB and represents the minimum data statistically required to adequately predict the clinical dose distribution in a group of patients.

Single Ultra-High Dose Rate Proton Transmission Beam for Whole Breast FLASH-Irradiation: Quantification of FLASH-Dose and Relation with Beam Parameters.

Healthy tissue-sparing effects of FLASH (≥40 Gy/s, ≥4-8 Gy/fraction) radiotherapy (RT) make it potentially useful for whole breast irradiation (WBI), since there is often a lot of normal tissue within the planning target volume (PTV). We investigated WBI plan quality and determined FLASH-dose for various machine settings using ultra-high dose rate (UHDR) proton transmission beams (TBs). While five-fraction WBI is commonplace, a potential FLASH-effect might facilitate shorter treatments, so hypothetical 2- and 1-fraction schedules were also analyzed. Using one tangential 250 MeV TB delivering 5 × 5.7 Gy, 2 × 9.74 Gy or 1 × 14.32 Gy, we evaluated: (1) spots with equal monitor units (MUs) in a uniform square grid with variable spacing; (2) spot MUs optimized with a minimum MU-threshold; and (3) splitting the optimized TB into two sub-beams: one delivering spots above an MU-threshold, i.e., at UHDRs; the other delivering the remaining spots necessary to improve plan quality. Scenarios 1-3 were planned for a test case, and scenario 3 was also planned for three other patients. Dose rates were calculated using the pencil beam scanning dose rate and the sliding-window dose rate. Various machine parameters were considered: minimum spot irradiation time (minST): 2 ms/1 ms/0.5 ms; maximum nozzle current (maxN): 200 nA/400 nA/800 nA; two gantry-current (GC) techniques: energy-layer and spot-based. For the test case (PTV = 819 cc) we found: (1) a 7 mm grid achieved the best balance between plan quality and FLASH-dose for equal-MU spots; (2) near the target boundary, lower-MU spots are necessary for homogeneity but decrease FLASH-dose; (3) the non-split beam achieved >95% FLASH for favorable (not clinically available) machine parameters (SB GC, low minST, high maxN), but <5% for clinically available settings (EB GC, minST = 2 ms, maxN = 200 nA); and (4) splitting gave better plan quality and higher FLASH-dose (~50%) for available settings. The clinical cases achieved ~50% (PTV = 1047 cc) or >95% (PTV = 477/677 cc) FLASH after splitting. A single UHDR-TB for WBI can achieve acceptable plan quality. Current machine parameters limit FLASH-dose, which can be partially overcome using beam-splitting. WBI FLASH-RT is technically feasible.

Deep learning-based markerless lung tumor tracking in stereotactic radiotherapy using Siamese networks.

BACKGROUND: Radiotherapy (RT) is involved in about 50% of all cancer patients, making it a very important treatment modality. The most common type of RT is external beam RT, which consists of delivering the radiation to the tumor from outside the body. One novel treatment delivery method is volumetric modulated arc therapy (VMAT), where the gantry continuously rotates around the patient during the radiation delivery. PURPOSE: Accurate tumor position monitoring during stereotactic body radiotherapy (SBRT) for lung tumors can help to ensure that the tumor is only irradiated when it is inside the planning target volume. This can maximize tumor control and reduce uncertainty margins, lowering organ-at-risk dose. Conventional tracking methods are prone to errors, or have a low tracking rate, especially for small tumors that are in close vicinity to bony structures. METHODS: We investigated patient-specific deep Siamese networks for real-time tumor tracking, during VMAT. Due to lack of ground truth tumor locations in the kilovoltage (kV) images, each patient-specific model was trained on synthetic data (DRRs), generated from the 4D planning CT scans, and evaluated on clinical data (x-rays). Since there are no annotated datasets with kV images, we evaluated the model on a 3D printed anthropomorphic phantom but also on six patients by computing the correlation coefficient with the breathing-related vertical displacement of the surface-mounted marker (RPM). For each patient/phantom, we used 80% of DRRs for training and 20% for validation. RESULTS: The proposed Siamese model outperformed the conventional benchmark template matching-based method (RTR): (1) when evaluating both methods on the 3D phantom, the Siamese model obtained a 0.57-0.79-mm mean absolute distance to the ground truth tumor locations, compared to 1.04-1.56 mm obtained by RTR; (2) on patient data, the Siamese-determined longitudinal tumor position had a correlation coefficient of 0.71-0.98 with the RPM, compared to 0.07-0.85 for RTR; (3) the Siamese model had a 100% tracking rate, compared to 62%-82% for RTR. CONCLUSIONS: Based on these results, we argue that Siamese-based real-time 2D markerless tumor tracking during radiation delivery is possible. Further investigation and development of 3D tracking is warranted.

Same-day adaptive palliative radiotherapy without prior CT simulation: Early outcomes in the FAST-METS study.

BACKGROUND AND PURPOSE: Standard palliative radiotherapy workflows involve waiting times or multiple clinic visits. We developed and implemented a rapid palliative workflow using diagnostic imaging (dCT) for pre-planning, with subsequent on-couch target and plan adaptation based on a synthetic computed tomography (CT) obtained from cone-beam CT imaging (CBCT). MATERIALS AND METHODS: Patients with painful bone metastases and recent diagnostic imaging were eligible for inclusion in this prospective, ethics-approved study. The workflow consisted of 1) telephone consultation with a radiation oncologist (RO); 2) pre-planning on the dCT using planning templates and mostly intensity-modulated radiotherapy; 3) RO consultation on the day of treatment; 4) CBCT scan with on-couch adaptation of the target and treatment plan; 5) delivery of either scheduled or adapted treatment plan. Primary outcomes were dosimetric data and treatment times; secondary outcome was patient satisfaction. RESULTS: 47 patients were enrolled between December 2021 and October 2022. In all treatments, adapted treatment plans were chosen due to significant improvements in target coverage (PTV/CTV V95%, p-value < 0.005) compared to the original treatment plan calculated on daily anatomy. Most patients were satisfied with the workflow. The average treatment time, including consultation and on-couch adaptive treatment, was 85 minutes. On-couch adaptation took on average 30 min. but was longer in cases where the automated deformable image registration failed to correctly propagate the targets. CONCLUSION: A fast treatment workflow for patients referred for painful bone metastases was implemented successfully using online adaptive radiotherapy, without a dedicated CT simulation. Patients were generally satisfied with the palliative radiotherapy workflow.

Influence of Beam Angle on Normal Tissue Complication Probability of Knowledge-Based Head and Neck Cancer Proton Planning.

Knowledge-based planning solutions have brought significant improvements in treatment planning. However, the performance of a proton-specific knowledge-based planning model in creating knowledge-based plans (KBPs) with beam angles differing from those used to train the model remains unexplored. We used a previously validated RapidPlanPT model and scripting to create nine KBPs, one with default and eight with altered beam angles, for 10 recent oropharynx cancer patients. The altered-angle plans were compared against the default-angle ones in terms of grade 2 dysphagia and xerostomia normal tissue complication probability (NTCP), mean doses of several organs at risk, and dose homogeneity index (HI). As KBP could be suboptimal, a proof of principle automatic iterative optimizer (AIO) was added with the aim of reducing the plan NTCP. There were no statistically significant differences in NTCP or HI between default- and altered-angle KBPs, and the altered-angle plans showed a <1% reduction in NTCP. AIO was able to reduce the sum of grade 2 NTCPs in 66/90 cases with mean a reduction of 3.5 ± 1.8%. While the altered-angle plans saw greater benefit from AIO, both default- and altered-angle plans could be improved, indicating that the KBP model alone was not completely optimal to achieve the lowest NTCP. Overall, the data showed that the model was robust to the various beam arrangements within the range described in this analysis.

Single-fraction 34 Gy Lung Stereotactic Body Radiation Therapy Using Proton Transmission Beams: FLASH-dose Calculations and the Influence of Different Dose-rate Methods and Dose/Dose-rate Thresholds.

Purpose: Research suggests that in addition to the dose-rate, a dose threshold is also important for the reduction in normal tissue toxicity with similar tumor control after ultrahigh dose-rate radiation therapy (UHDR-RT). In this analysis we aimed to identify factors that might limit the ability to achieve this "FLASH"-effect in a scenario attractive for UHDR-RT (high fractional beam dose, small target, few organs-at-risk): single-fraction 34 Gy lung stereotactic body radiation therapy. Methods and Materials: Clinical volumetric-modulated arc therapy (VMAT) plans, intensity modulated proton therapy (IMPT) plans and transmission beam (TB) plans were compared for 6 small and 1 large lung lesion. The TB-plan dose-rate was calculated using 4 methods and the FLASH-percentage (percentage of dose delivered at dose-rates ≥40/100 Gy/s and ≥4/8 Gy) was determined for various variables: a minimum spot time (minST) of 0.5/2 ms, maximum nozzle current (maxN) of 200/40 0nA, and 2 gantry current (GC) techniques (energy-layer based, spot-based [SB]). Results: Based on absolute doses 5-beam TB and VMAT-plans are similar, but TB-plans have higher rib, skin, and ipsilateral lung dose than IMPT. Dose-rate calculation methods not considering scanning achieve FLASH-percentages between ∼30% to 80%, while methods considering scanning often achieve <30%. FLASH-percentages increase for lower minST/higher maxN and when using SB GC instead of energy-layer based GC, often approaching the percentage of dose exceeding the dose-threshold. For the small lesions average beam irradiation times (including scanning) varied between 0.06 to 0.31 seconds and total irradiation times between 0.28 to 1.57 seconds, for the large lesion beam times were between 0.16 to 1.47 seconds with total irradiation times of 1.09 to 5.89 seconds. Conclusions: In a theoretically advantageous scenario for FLASH we found that TB-plan dosimetry was similar to that of VMAT, but inferior to that of IMPT, and that decreasing minST or using SB GC increase the estimated amount of FLASH. For the appropriate machine/delivery parameters high enough dose-rates can be achieved regardless of calculation method, meaning that a possible FLASH dose-threshold will likely be the primary limiting factor.

Fast, Automated, Knowledge-Based Treatment Planning for Selecting Patients for Proton Therapy Based on Normal Tissue Complication Probabilities.

Purpose: Selecting patients who will benefit from proton therapy is laborious and subjective. We demonstrate a novel automated solution for creating high-quality knowledge-based plans (KBPs) using proton and photon beams to identify patients for proton treatment based on their normal tissue complication probabilities (NTCP). Methods and Materials: Two previously validated RapidPlan PT models for locally advanced head and neck cancer were used in combination with scripting to automatically create proton and photon KBPs for 72 patients with recent oropharynx cancer. NTCPs were calculated for each patient based on the KBPs, and patient selection was simulated according to the current Dutch national protocol. Results: The photon/proton KBP exhibited good correlation between predicted and achieved organ-at-risk mean doses, with a ≤5 Gy difference in 208/196 out of 215 structures relevant for the head and neck cancer NTCP model. The proton KBPs yielded on average 7.1/6.1/7.6 Gy lower dose to salivary/swallowing structures/oral cavity than the photon KBPs. This reduced average grade 2/3 dysphagia and xerostomia by 7.1/3.3 and 5.5/2.0 percentage points, resulting in 16 of 72 patients (22%) being indicated for proton treatment. The entire automated process took <30 minutes per patient. Conclusions: Automated support for decision making using KBP is feasible and fast. The planning solution has potential to speed up the planning and patient-selection process significantly without major compromises to the plan quality.

Deep Learning for Automated Elective Lymph Node Level Segmentation for Head and Neck Cancer Radiotherapy.

Depending on the clinical situation, different combinations of lymph node (LN) levels define the elective LN target volume in head-and-neck cancer (HNC) radiotherapy. The accurate auto-contouring of individual LN levels could reduce the burden and variability of manual segmentation and be used regardless of the primary tumor location. We evaluated three deep learning approaches for the segmenting individual LN levels I−V, which were manually contoured on CT scans from 70 HNC patients. The networks were trained and evaluated using five-fold cross-validation and ensemble learning for 60 patients with (1) 3D patch-based UNets, (2) multi-view (MV) voxel classification networks and (3) sequential UNet+MV. The performances were evaluated using Dice similarity coefficients (DSC) for automated and manual segmentations for individual levels, and the planning target volumes were extrapolated from the combined levels I−V and II−IV, both for the cross-validation and for an independent test set of 10 patients. The median DSC were 0.80, 0.66 and 0.82 for UNet, MV and UNet+MV, respectively. Overall, UNet+MV significantly (p < 0.0001) outperformed other arrangements and yielded DSC = 0.87, 0.85, 0.86, 0.82, 0.77, 0.77 for the combined and individual level I−V structures, respectively. Both PTVs were also significantly (p < 0.0001) more accurate with UNet+MV, with DSC = 0.91 and 0.90, respectively. The accurate segmentation of individual LN levels I−V can be achieved using an ensemble of UNets. UNet+MV can further refine this result.

The markerless lung target tracking AAPM Grand Challenge (MATCH) results.

PURPOSE: Lung stereotactic ablative body radiotherapy (SABR) is a radiation therapy success story with level 1 evidence demonstrating its efficacy. To provide real-time respiratory motion management for lung SABR, several commercial and preclinical markerless lung target tracking (MLTT) approaches have been developed. However, these approaches have yet to be benchmarked using a common measurement methodology. This knowledge gap motivated the MArkerless lung target Tracking CHallenge (MATCH). The aim was to localize lung targets accurately and precisely in a retrospective in silico study and a prospective experimental study. METHODS: MATCH was an American Association of Physicists in Medicine sponsored Grand Challenge. Common materials for the in silico and experimental studies were the experiment setup including an anthropomorphic thorax phantom with two targets within the lungs, and a lung SABR planning protocol. The phantom was moved rigidly with patient-measured lung target motion traces, which also acted as ground truth motion. In the retrospective in silico study a volumetric modulated arc therapy treatment was simulated and a dataset consisting of treatment planning data and intra-treatment kilovoltage (kV) and megavoltage (MV) images for four blinded lung motion traces was provided to the participants. The participants used their MLTT approach to localize the moving target based on the dataset. In the experimental study, the participants received the phantom experiment setup and five patient-measured lung motion traces. The participants used their MLTT approach to localize the moving target during an experimental SABR phantom treatment. The challenge was open to any participant, and participants could complete either one or both parts of the challenge. For both the in silico and experimental studies the MLTT results were analyzed and ranked using the prospectively defined metric of the percentage of the tracked target position being within 2 mm of the ground truth. RESULTS: A total of 30 institutions registered and 15 result submissions were received, four for the in silico study and 11 for the experimental study. The participating MLTT approaches were: Accuray CyberKnife (2), Accuray Radixact (2), BrainLab Vero, C-RAD, and preclinical MLTT (5) on a conventional linear accelerator (Varian TrueBeam). For the in silico study the percentage of the 3D tracking error within 2 mm ranged from 50% to 92%. For the experimental study, the percentage of the 3D tracking error within 2 mm ranged from 39% to 96%. CONCLUSIONS: A common methodology for measuring the accuracy of MLTT approaches has been developed and used to benchmark preclinical and commercial approaches retrospectively and prospectively. Several MLTT approaches were able to track the target with sub-millimeter accuracy and precision. The study outcome paves the way for broader clinical implementation of MLTT. MATCH is live, with datasets and analysis software being available online at https://www.aapm.org/GrandChallenge/MATCH/ to support future research.