Tryggestad, Erik

Abstract Objective . Phase gating is a critical technique to mitigate tumor motion during radiotherapy, particularly in spot-scanned particle therapy where internal motion can interfere with dynamic spot scanning patterns and, simultaneously, introducing substantial range uncertainties. However, the current commercial state of the art in real-time phase prediction is challenged by patient-specific breathing variability as well as detection and delivery system latencies. This leads to suboptimal efficiency, mis-timed radiation delivery and requires frequent manual intervention. This study aims to improve phase prediction accuracy using deep learning (DL)-based time series forecasting to enable more accurate dose delivery. Approach . Retrospective breathing waveforms from 69 proton therapy patients, sampled at 30 Hz, were labeled with inspiratory peaks and assigned subjective regularity scores (four levels). DL models with various architectures were trained using waveform amplitude to predict current or future breathing phases. Model performance was evaluated using mean squared error, phase binning accuracy, and timing deviation for radiation on/off events. Bayesian optimization was used for hyperparameter tuning. Results were compared between models and against a commercial algorithm currently in clinical use. Main Results . The curated waveform dataset included 165 242 s for training, 24 057 s for validation, and 30 322 s for testing, with an additional 40 604 s from separate patients for extended validation. The long short-term memory and temporal fusion transformer models significantly outperformed the commercial algorithm, improving phase prediction accuracy by nearly 20% and reducing timing deviations across all regularity levels. Significance . DL-based time series forecasting may substantially improve breathing phase prediction accuracy over current clinically available methods, offering a more precise and reliable approach to phase-gated radiation delivery.

Abstract Background Deep learning auto segmentation (DLAS) tools are widely used for radiation therapy planning, yet limited information exists on how patient positioning impacts their performance, particularly in head and neck (HN) computed tomography (CT) imaging. Purpose This study investigates the impact of abnormal neck tilt on the accuracy of DLAS‐generated contours in HN CT scans. We hypothesize that abnormal positioning degrades auto segmentation performance, particularly for organs at risk (OARs) influenced by cervical spine orientation. Methods A total of 35 HN CT scans were retrospectively analyzed. Neck tilt was quantified using principal component analysis of the spinal cord contours from C1 to C4, with patients stratified into normal and abnormal tilt groups based on percentile thresholds. Seven FDA‐cleared DLAS tools were evaluated for OAR segmentation accuracy across brainstem, parotid glands, submandibular glands, brachial plexus, and optic nerves. Gold‐standard contours were curated from manually created clinical contours and compared to DLAS output using Sorenson's volumetric dice similarity coefficient (DSC), surface Dice similarity coefficient with a 2 mm threshold (sDSC), mean distance to agreement (MDA) and difference in mean dose. Statistical differences were assessed using Wilcoxon rank‐sum tests ( p < 0.05). Results Among the evaluated OARs, the parotid glands showed consistent and statistically significant degradation in contouring accuracy for patients with abnormal neck tilt across all metrics (DSC, sDSC, MDA) and across all seven DLAS tools. Changes in contouring accuracy resulted in significant differences in both absolute and signed difference in mean dose. No consistent differences were observed for other structures across multiple DLAS tools. Conclusions Abnormal neck tilt is associated with reduced DLAS accuracy and greater dose variability for parotid gland segmentation in HN radiotherapy planning. These findings underscore the need for enhanced patient quality assurance strategies for patients with abnormal positioning when using clinical DLAS workflows. Future work should explore automated detection of anatomical outliers and site‐specific model retraining to ensure accurate contouring delineation for patients with atypical positioning or anatomy.

Abstract Dose uncertainty induced by respiratory motion remains a major concern for treating thoracic and abdominal lesions using particle beams. This Task Group report reviews the impact of tumor motion and dosimetric considerations in particle radiotherapy, current motion‐management techniques, and limitations for different particle‐beam delivery modes (i.e., passive scattering, uniform scanning, and pencil‐beam scanning). Furthermore, the report provides guidance and risk analysis for quality assurance of the motion‐management procedures to ensure consistency and accuracy, and discusses future development and emerging motion‐management strategies. This report supplements previously published AAPM report TG76, and considers aspects of motion management that are crucial to the accurate and safe delivery of particle‐beam therapy. To that end, this report produces general recommendations for commissioning and facility‐specific dosimetric characterization, motion assessment, treatment planning, active and passive motion‐management techniques, image guidance and related decision‐making, monitoring throughout therapy, and recommendations for vendors. Key among these recommendations are that: (1) facilities should perform thorough planning studies (using retrospective data) and develop standard operating procedures that address all aspects of therapy for any treatment site involving respiratory motion; (2) a risk‐based methodology should be adopted for quality management and ongoing process improvement.

PurposeThe presence of respiratory motion during radiation treatment leads to degradation of the expected dose distribution, both for target coverage and healthy tissue sparing, particularly for techniques like pencil beam scanning proton therapy which have dynamic delivery systems. While tools exist to estimate this degraded four‐dimensional (4D) dose, they typically have one or more deficiencies such as not including the particular effects from a dynamic delivery, using analytical dose calculations, and/or using nonphysical dose‐accumulation methods. This work presents a clinically useful 4D‐dose calculator that addresses each of these shortcomings.MethodsTo quickly compute the 4D dose, the three main tasks of the calculator were run on graphics processing units (GPUs). These tasks were (a) simulating the delivery of the plan using measured delivery parameters to distribute the plan amongst 4DCT phases characterizing the patient breathing, (b) using an in‐house Monte Carlo simulation (MC) dose calculator to determine the dose delivered to each breathing phase, and (c) accumulating the doses from the various breathing phases onto a single phase for evaluation. The accumulation was performed by individually transferring the energy and mass of dose‐grid subvoxels, a technique that models the transfer of dose in a more physically realistic manner. The calculator was run on three test cases, with lung, esophagus, and liver targets, respectively, to assess the various uncertainties in the beam delivery simulation as well as to characterize the dose‐accumulation technique.ResultsFour‐dimensional doses were successfully computed for the three test cases with computation times ranging from 4–6 min on a server with eight NVIDIA Titan X graphics cards; the most time‐consuming component was the MC dose engine. The subvoxel‐based dose‐accumulation technique produced stable 4D‐dose distributions at subvoxel scales of 0.5–1.0 mm without impairing the total computation time. The uncertainties in the beam delivery simulation led to moderate variations of the dose–volume histograms for these cases; the variations were reduced by implementing repainting or phase‐gating motion mitigation techniques in the calculator.ConclusionsA MC‐based and GPU‐accelerated 4D‐dose calculator was developed to estimate the effects of respiratory motion on pencil beam scanning proton therapy treatments. After future validation, the calculator could be used to assess treatment plans and its quick runtime would make it easily usable in a future 4D‐robust optimization system.

PurposeTo accurately model the beam delivery time (BDT) for a synchrotron‐based proton spot scanning system using experimentally determined beam parameters.MethodsA model to simulate the proton spot delivery sequences was constructed, and BDT was calculated by summing times for layer switch, spot switch, and spot delivery. Test plans were designed to isolate and quantify the relevant beam parameters in the operation cycle of the proton beam therapy delivery system. These parameters included the layer switch time, magnet preparation and verification time, average beam scanning speeds in x‐ and y‐directions, proton spill rate, and maximum charge and maximum extraction time for each spill. The experimentally determined parameters, as well as the nominal values initially provided by the vendor, served as inputs to the model to predict BDTs for 602 clinical proton beam deliveries. The calculated BDTs (TBDT) were compared with the BDTs recorded in the treatment delivery log files (TLog): ∆t = TLog−TBDT.ResultsThe experimentally determined average layer switch time for all 97 energies was 1.91 s (ranging from 1.9 to 2.0 s for beam energies from 71.3 to 228.8 MeV), average magnet preparation and verification time was 1.93 ms, the average scanning speeds were 5.9 m/s in x‐direction and 19.3 m/s in y‐direction, the proton spill rate was 8.7 MU/s, and the maximum proton charge available for one acceleration is 2.0 ± 0.4 nC. Some of the measured parameters differed from the nominal values provided by the vendor. The calculated BDTs using experimentally determined parameters matched the recorded BDTs of 602 beam deliveries (∆t = −0.49 ± 1.44 s), which were significantly more accurate than BDTs calculated using nominal timing parameters (∆t = −7.48 ± 6.97 s).ConclusionsAn accurate model for BDT prediction was achieved by using the experimentally determined proton beam therapy delivery parameters, which may be useful in modeling the interplay effect and patient throughput. The model may provide guidance on how to effectively reduce BDT and may be used to identifying deteriorating machine performance.