Deisher, Amanda J

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.

BACKGROUND: Cardiac radioablation is emerging as a treatment modality for refractory ventricular tachycardia. This study aimed to evaluate the effects of radiation on the coronary arteries in a swine model of proton beam cardiac radioablation. METHODS: Eighteen swine underwent single-fraction 30 to 40 Gy pencil-beam scanning proton therapy targeting the left ventricle and were euthanized 12 to 40 weeks later. Treatment planning was performed without restricting the dose to the coronary arteries. The maximum point dose (D max ) to the epicardial coronary arteries was calculated. In secondary analyses, the mean (D mean ) and minimum dose received by the highest irradiated 0.01 cm 3 (D 0.01 ) of each coronary artery were also calculated. Coronary artery segments were harvested from the D max sites for histological analysis, and the D max was correlated with stenosis severity. RESULTS: Ninety-six coronary arteries were analyzed. No stenoses were observed by computed tomography imaging preirradiation. By histological analysis posteuthanasia, 25/96 (26%) coronaries sampled at their D max sites had ≥75% stenosis. The median D max was 4.7 Gy for the <75% stenosis group and 29.7 Gy for the ≥75% stenosis group ( P <0.001). The AUC-ROC for the association between D max and stenosis ≥75% was 92.2%. A D max value of 20.1 Gy best predicted stenosis ≥75%, with sensitivity 92.3% and specificity 87.1%. The AUC-ROCs for the associations of D mean and D 0.01 with stenosis ≥75% were 84.8% and 91.6%, respectively. In histopathologic analysis, intimal hyperplasia was the most common coronary artery abnormality at the D max sites, and it was present in 61.5% of all arteries and in 93.9% of arteries with D max ≥20 Gy. CONCLUSIONS: In this preclinical model of proton beam cardiac radioablation, coronary stenoses occurred in a dose-dependent manner, with D max showing the closest correlation with stenosis ≥75%. These data provide for the first time a framework for dose constraint considerations for the coronary arteries during treatment planning for cardiac radioablation and thoracic malignancy radiation.

Background: Particle therapy is a noninvasive, catheter-free modality for cardiac ablation. We previously demonstrated the efficacy for creating ablation lesions in the porcine heart. Despite several earlier studies, the exact mechanism of early biophysical effects of proton and photon beam delivery on the myocardium remain incompletely resolved. Methods: Ten normal and 9 infarcted in situ porcine hearts received proton beam irradiation (40 Gy) delivered to the left ventricular myocardium with follow-up for 8 weeks. High-resolution electroanatomical mapping of the left ventricular was performed at baseline and follow-up. Bipolar voltage amplitude, conduction velocity, and connexin-43 were determined within the irradiated and nonirradiated areas. Results: The irradiated area in normal hearts showed a significant reduction of bipolar voltage amplitude (10.1±4.9 mV versus 5.7±3.2, P <0.0001) and conduction velocity (85±26 versus 55±13 cm/s, P =0.03) beginning at 4 weeks after irradiation. In infarcted myocardium after irradiation, bipolar voltage amplitude of the infarct scar (2.0±2.9 versus 0.8±0.7 mV, P =0.008) was significantly reduced as well as the conduction velocity in the infarcted heart (43.7±15.7 versus 26.3±11.4 cm/s, P =0.02). There were no significant changes in bipolar voltage amplitude and conduction velocity in nonirradiated myocardium. Myocytolysis, capillary hyperplasia, and dilation were seen in the irradiated myocardium 8 weeks after irradiation. Active caspase-3 and reduction of connexin-43 expression began in irradiated myocardium 1 week after irradiation and decreased over 8 weeks. Conclusions: Irradiation of the myocardium with proton beams reduce connexin-43 expression, conduction velocity, and bipolar conducted electrogram amplitude in a large porcine model. The changes in biomarkers preceded electrophysiological changes after proton beam therapy.

AbstractBackground and purposeProton therapy is a key modality used in the treatment of ocular melanoma. Traditionally ocular sites are treated using a dedicated eyeline with a passively scattered proton beam and a brass aperture. This work aims to design and characterize a beam‐collimating aperture to treat ocular targets with a gantry‐based spot scanning proton beam.MethodsA plastic aperture system that slides into the gantry nozzle of a spot scanning proton beam was designed and constructed. It consists of an intermediate scraper layer to attenuate stray protons and a 3D‐printed patient‐specific aperture positioned 5.7 cm from the surface of the eye. The aperture system was modeled in TOPAS and Monte Carlo simulations were validated with film measurements. Two different spot configurations were investigated for treatment planning and characterized based on lateral penumbra, central axis (CAX) dose and relative efficiency. Alignment and leakage were investigated through experimental film measurements. Range was verified using a multi‐layer ionization chamber. Reference dose measurements were made with a PinPoint 3D ion chamber. Neutron dose was evaluated through Monte Carlo simulations.ResultsAperture alignment with radiation isocenter was determined to be within 0.31 mm at a gantry angle of 0°. A single‐spot configuration with a 10 mm diameter aperture yielded film‐measured lateral penumbras of 1 mm to 1.25 mm, depending on depth in the spread‐out Bragg peak. TOPAS simulations found that a single spot configuration results in a flat dose distribution for a 10 mm diameter aperture and provides a CAX dose of less than 106% for apertures less than 14 mm in diameter. For larger targets, adding four corner spots to fill in the dose distribution is beneficial. Trade‐offs between lateral penumbra, CAX dose and relative efficiency were characterized for different spot configurations and can be used for future clinical decision‐making. The aperture was experimentally determined to not affect proton beam range, and no concerning leakage radiation or neutron dose was identified. Reference dose measurements with a PinPoint ion chamber were within 2.1% of Monte Carlo calculated doses.ConclusionThe aperture system developed in this work provides a method of treating ocular sites on a gantry‐based spot scanning proton system. Additional work to develop compatible gaze tracking and gating infrastructure is ongoing.

Background: Proton beam therapy offers radiophysical properties that are appealing for noninvasive arrhythmia elimination. This study was conducted to use scanned proton beams for ablation of cardiac tissue, investigate electrophysiological outcomes, and characterize the process of lesion formation in a porcine model using particle therapy. Methods: Twenty-five animals received scanned proton beam irradiation. ECG-gated computed tomography scans were acquired at end-expiration breath hold. Structures (atrioventricular junction or left ventricular myocardium) and organs at risk were contoured. Doses of 30, 40, and 55 Gy were delivered during expiration to the atrioventricular junction (n=5) and left ventricular myocardium (n=20) of intact animals. Results: In this study, procedural success was tracked by pacemaker interrogation in the atrioventricular junction group, time-course magnetic resonance imaging in the left ventricular group, and correlation of lesion outcomes displayed in gross and microscopic pathology. Protein extraction (active caspase-3) was performed to investigate tissue apoptosis. Doses of 40 and 55 Gy caused slowing and interruption of cardiac impulse propagation at the atrioventricular junction. In 40 left ventricular irradiated targets, all lesions were identified on magnetic resonance after 12 weeks, being consistent with outcomes from gross pathology. In the majority of cases, lesion size plateaued between 12 and 16 weeks. Active caspase-3 was seen in lesions 12 and 16 weeks after irradiation but not after 20 weeks. Conclusions: Scanned proton beams can be used as a tool for catheter-free ablation, and time-course of tissue apoptosis was consistent with lesion maturation.

AbstractPurposeThe aim of this work is to describe the clinical implementation of respiratory‐gated spot‐scanning proton therapy (SSPT) for the treatment of thoracic and abdominal moving targets. The experience of our institution is summarized, from initial acceptance and commissioning tests to the development of standard clinical operating procedures for simulation, motion assessment, motion mitigation, treatment planning, and gated SSPT treatment delivery.Materials and methodsA custom respiratory gating interface incorporating the Real‐Time Position Management System (RPM, Varian Medical Systems, Inc., Palo Alto, CA, USA) was developed in‐house for our synchrotron‐based delivery system. To assess gating performance, a motion phantom and radiochromic films were used to compare gated vs nongated delivery. Site‐specific treatment planning protocols and conservative motion cutoffs were developed, allowing for free‐breathing (FB), breath‐holding (BH), or phase‐gating (Ph‐G). Room usage efficiency of BH and Ph‐G treatments was retrospectively evaluated using beam delivery data retrieved from our record and verify system and DICOM files from patient‐specific quality assurance (QA) procedures.ResultsMore than 70 patients were treated using active motion management between the launch of our motion mitigation program in October 2015 and the end date of data collection of this study in January 2018. During acceptance procedures, we found that overall system latency is clinically‐suitable for Ph‐G. Regarding room usage efficiency, the average number of energy layers delivered per minute was <10 for Ph‐G, 10‐15 for BH and ≥15 for FB, making Ph‐G the slowest treatment modality. When comparing to continuous delivery measured during pretreatment QA procedures, the median values of BH treatment time were extended from 6.6 to 9.3 min (+48%). Ph‐G treatments were extended from 7.3 to 13.0 min (+82%).ConclusionsActive motion management has been crucial to the overall success of our SSPT program. Nevertheless, our conservative approach has come with an efficiency cost that is more noticeable in Ph‐G treatments and should be considered in decision‐making.

Purpose Photon counting detectors (PCDs) are being introduced in advanced x‐ray computed tomography (CT) scanners. From a single PCD‐CT acquisition, multiple images can be reconstructed, each based on only a part of the original x‐ray spectrum. In this study, we investigated whether PCD‐CT can be used to estimate stopping power ratios (SPRs) for proton therapy treatment planning, both by comparing to other SPR methods proposed for single energy CT (SECT) and dual energy CT (DECT) as well as to experimental measurements. Methods A previously developed DECT‐based SPR estimation method was adapted to PCD‐CT data, by adjusting the estimation equations to allow for more energy spectra. The method was calibrated directly on noisy data to increase the robustness toward image noise. The new PCD SPR estimation method was tested in theoretical calculations as well as in an experimental setup, using both four and two energy bin PCD‐CT images, and through comparison to two other SPR methods proposed for SECT and DECT. These two methods were also evaluated on PCD‐CT images, full spectrum (one‐bin) or two‐bin images, respectively. In a theoretical framework, we evaluated the effect of patient‐specific tissue variations (density and elemental composition) and image noise on the SPR accuracy; the latter effect was assessed by applying three different noise levels (low, medium, and high noise). SPR estimates derived using real PCD‐CT images were compared to experimentally measured SPRs in nine organic tissue samples, including fat, muscle, and bone tissues. Results For the theoretical calculations, the root‐mean‐square error (RMSE) of the SPR estimation was 0.1% for the new PCD method using both two and four energy bins, compared to 0.2% and 0.7% for the DECT‐ and SECT‐based method, respectively. The PCD method was found to be very robust toward CT image noise, with a RMSE of 2.7% when high noise was added to the CT numbers. Introducing tissue variations, the RMSE only increased to 0.5%; even when adding high image noise to the changed tissues, the RMSE stayed within 3.1%. In the experimental measurements, the RMSE over the nine tissue samples was 0.8% when using two energy bins, and 1.0% for the four‐bin images. Conclusions In all tested cases, the new PCD method produced similar or better results than the SECT‐ and DECT‐based methods, showing an overall improvement of the SPR accuracy. This study thus demonstrated that PCD‐CT scans will be a qualified candidate for SPR estimations.

Background— This study sought to investigate external photon beam radiation for catheter-free ablation of the atrioventricular junction in intact pigs. Methods and Results— Ten pigs were randomized to either sham irradiation or irradiation of the atrioventricular junction (55, 50, 40, and 25 Gy). Animals underwent baseline electrophysiological evaluation, cardiac gated multi-row computed tomographic imaging for beam delivery planning, and intensity-modulated radiation therapy. Doses to the coronary arteries were optimized. Invasive follow-up was conducted ≤4 months after the irradiation. A mean volume of 2.5±0.5 mL was irradiated with target dose. The mean follow-up length after irradiation was 124.8±30.8 days. Out of 7 irradiated animals, complete atrioventricular block was achieved in 6 animals of all 4 dose groups (86%). Using the same targeting margins, ablation lesion size notably increased with the delivered dose because of volumetric effects of isodose lines around the target volume. The mean macroscopically calculated atrial lesion volume for all 4 dose groups was 3.8±1.1 mL, lesions extended anteriorly into the interventricular septum. No short-term side effects were observed. No damage was observed in the tissues of the esophagus, phrenic nerves, or trachea. However, histology revealed in-field beam effects outside of the target volume. Conclusions— Single-fraction doses as low as 25 Gy caused a lesion with interruption of cardiac impulse propagation using this respective target volume. With doses of ≤55 Gy, maximal point-doses to coronary arteries could be kept <7Gy, but target conformity of lesions was not fully achieved using this approach.