Proton Treatment Research Articles

Developing a comparative photon-proton planning service in Victoria: the experience at Peter MacCallum Cancer Centre.

Proton-beam therapy (PBT) is a cutting-edge radiation therapy modality that is currently not available in Australia. Comparative photon-proton (CPP) planning is required for the medical treatment overseas programme (MTOP) and will be required for access to PBT in Australia in the future. Comparative planning brings professional development benefits to all members of the radiation therapy team. This service was also created to support future proposals for a PBT facility in Victoria. We report our experience developing an in-house CPP service at Peter MacCallum Cancer Centre. A set of resources to support CPP planning was established. Training of relevant staff was undertaken after which an in-house training programme was developed. A standard protocol for PBT planning parameters was established. All CPP plans were reviewed. Future goals for the CPP planning programme were described. In total, 62 cases were comparatively planned over 54 months. Of these, 60% were paediatric cases, 14% were adolescents and young adults (15-25 years) and 26% were adults. The vast majority (over 75%) of patients comparatively planned required irradiation to the central nervous system including brain and cranio-spinal irradiation. A variety of proton plans were reviewed by international PBT experts to confirm their deliverability. Our team at Peter MacCallum Cancer Centre has gained significant experience in CPP planning and will continue to develop this further. Local expertise will help support decentralisation of patient selection for proton treatments in the near future and the PBT business case in Victoria.

Proton-FLASH: effects of ultra-high dose rate irradiation on an in-vivo mouse ear model

FLASH-radiotherapy may provide significant sparing of healthy tissue through ultra-high dose rates in protons, electrons, and x-rays while maintaining the tumor control. Key factors for the FLASH effect might be oxygen depletion, the immune system, and the irradiated blood volume, but none could be fully confirmed yet. Therefore, further investigations are necessary. We investigated the protective (tissue sparing) effect of FLASH in proton treatment using an in-vivo mouse ear model. The right ears of Balb/c mice were irradiated with 20 MeV protons at the ion microprobe SNAKE in Garching near Munich by using three dose rates (Conv = 0.06 Gy/s, Flash9 = 9.3 Gy/s and Flash930 = 930 Gy/s) at a total dose of 23 Gy or 33 Gy. The ear thickness, desquamation, and erythema combined in an inflammation score were measured for 180 days. The cytokines TGF-β1, TNF-α, IL1α, and IL1β were analyzed in the blood sampled in the first 4 weeks and at termination day. No differences in inflammation reactions were visible in the 23 Gy group for the different dose rates. In the 33 Gy group, the ear swelling and the inflammation score for Flash9 was reduced by (57 ± 12) % and (67 ± 17) % and for Flash930 by (40 ± 13) % and (50 ± 17) % compared to the Conv dose rate. No changes in the cytokines in the blood could be measured. However, an estimation of the irradiated blood volume demonstrates, that 100-times more blood is irradiated when using Conv compared to using Flash9 or Flash930. This indicates that blood might play a role in the underlying mechanisms in the protective effect of FLASH.

Deep learning-based synthetic dose-weighted LET map generation for intensity modulated proton therapy

The advantage of proton therapy as compared to photon therapy stems from the Bragg peak effect, which allows protons to deposit most of their energy directly at the tumor while sparing healthy tissue. However, even with such benefits, proton therapy does present certain challenges. The biological effectiveness differences between protons and photons are not fully incorporated into clinical treatment planning processes. In current clinical practice, the relative biological effectiveness (RBE) between protons and photons is set as constant 1.1. Numerous studies have suggested that the RBE of protons can exhibit significant variability. Given these findings, there is a substantial interest in refining proton therapy treatment planning to better account for the variable RBE. Dose-average linear energy transfer (LETd) is a key physical parameter for evaluating the RBE of proton therapy and aids in optimizing proton treatment plans. Calculating precise LETd distributions necessitates the use of intricate physical models and the execution of specialized Monte-Carlo simulation software, which is a computationally intensive and time-consuming progress. In response to these challenges, we propose a deep learning based framework designed to predict the LETd distribution map using the dose distribution map. This approach aims to simplify the process and increase the speed of LETd map generation in clinical settings. The proposed CycleGAN model has demonstrated superior performance over other GAN-based models. The mean absolute error (MAE), peak signal-to-noise ratio and normalized cross correlation of the LETd maps generated by the proposed method are 0.096 ± 0.019 keV μm−1, 24.203 ± 2.683 dB, and 0.997 ± 0.002, respectively. The MAE of the proposed method in the clinical target volume, bladder, and rectum are 0.193 ± 0.103, 0.277 ± 0.112, and 0.211 ± 0.086 keV μm−1, respectively. The proposed framework has demonstrated the feasibility of generating synthetic LETd maps from dose maps and has the potential to improve proton therapy planning by providing accurate LETd information.

Current Perspectives on Proton Therapy: Techniques Shaping the Future of Cancer Treatment

Traditional Radiation Therapy (RT) predominantly comprises a targeted therapeutic strategy focused on improving localised tumour control and achieving a cure while minimising the occurrence of adverse side effects. It could be feasible to take advantage of the better dose distribution by enabling larger RT dosages to the malignancy while preventing a rise in the toxicity of RT-induced healthy tissue, or by reducing adverse reactions to manageable levels. Poor local disease control and important dose-limiting normal tissue, which prevent safe dosage increase with conventional photon RT, have been the key justifications for RT. Proton treatment, on the other hand, delivers therapeutic protons or positive particles using proton beams. The potential advantage of protons’ physical properties allows for more localised RT delivery. By increasing the dosage to equitoxic levels, it is also possible to take advantage of the potential improvement in normal tissue sparing to support local tumour management and, ideally, longevity. Proton treatment preserves more important structures than photon therapy because of its unique physics. Thus, there is a need for wide usage of Proton Therapy (PT) for successive cancer treatment. The present review focuses on PT based on tumour site, clinical studies, biological barriers, instrumentation of PT, significance, and limitations.

Leveraging GAN-based CBCT-to-CT translation models for enhanced image quality and accurate photon and proton dose calculation in adaptive radiotherapy

PurposeTo develop a deep learning-based network for CBCT-to-CT translation. This approach aims to enhance CBCT image quality and enable more accurate photon and proton dose calculation in adaptive radiotherapy. MethodsThe proposed network is based on a simple generative adversarial network (GAN) and task-specific loss functions. The CBCT images were input into the network for synthetic CT (sCT) generation with regularization loss in the form of gradient loss (between CBCT and sCT) and style loss (between registered planning CT and sCT) to preserve the CBCT structural information and obtain accurate HU values through CT style transfer, respectively. CBCT and planning CT images collected from 58 patients with pelvic lesions were used for training. The model performance was quantitatively and qualitatively evaluated and compared to that of three other models (Res-G, CycleGAN and CUT). Quantitative analysis was performed to assess the noise reduction and dosimetric improvement achieved by the proposed network. The performance of the model was evaluated in an additional 6 pelvic patients. ResultsThe best visual effects on the test data were achieved using our model, which showed detailed information preservation and artifact removal ability. Quantitative values of 13.02 ± 4.40 HU, 56.59 ± 20.58 HU, 37.53 ± 3.06 dB and 93.43 ± 1.23% were obtained for the MAE, RSME, PSNR and SSIM, respectively. Our sCT also showed CT-like noise levels, with a mean global noise level (GNL) of 3.31 ± 0.32 HU. The dosimetric evaluation revealed a gamma pass rate higher than 99% under all criteria for photon treatment and high pass rates of 95.77 ± 1.21% (2%/2 mm) and 98.07 ± 0.45% (3%/3 mm)) for proton treatment. ConclusionWe introduce novel task-specific loss functions into a GAN-based network for generating high-quality sCT images. The model also has an excellent ability to remove artifacts and suppress noise. Dosimetric results show the potential for integration into the clinical ART workflow.

A quantitative framework for patient-specific collision detection in proton therapy.

Beam modifying accessories for proton therapy often need to be placed in close proximity of the patient for optimal dosimetry. However, proton treatment units are larger in size and as a result the planned treatment geometry may not be achievable due to collisions with the patient. A framework that can accurately simulate proton treatment geometry is desired. A quantitative framework was developed to model patient-specific proton treatment geometry, minimize air gap, and avoid collisions. The patient's external contour is converted into the International Electrotechnique Commission (IEC) gantry coordinates following the patient's orientation and each beam's gantry and table angles. All snout components are modeled by three-dimensional (3D) geometric shapes such as columns, cuboids, and frustums. Beam-specific parameters such as isocenter coordinates, snout type and extension are used to determine if any point on the external contour protrudes into the various snout components. A 3D graphical user interface is also provided to the planner to visualize the treatment geometry. In case of a collision, the framework's analytic algorithm quantifies the maximum protrusion of the external contour into the snout components. Without a collision, the framework quantifies the minimum distance of the external contour from the snout components and renders a warning if such distance is less than 5cm. Three different snout designs are modeled. Examples of potential collision and its aversion by snout retraction are demonstrated. Different patient orientations, including a sitting treatment position, as well as treatment plans with multiple isocenters, are successfully modeled in the framework. Finally, the dosimetric advantage of reduced air gap enabled by this framework is demonstrated by comparing plans with standard and reduced air gaps. Implementation of this framework reduces incidence of collisions in the treatment room. In addition, it enables the planners to minimize the air gap and achieve better plan dosimetry.

A Longitudinal Metagenomic Comparative Analysis of Oral Microbiome Shifts in Patients Receiving Proton Radiation versus Photon Radiation for Head and Neck Cancer.

Due to the radiation-sparing effects on salivary gland acini, changes in the composition of the oral microbiome may be a driver for improved outcomes in patients receiving proton radiation, with potentially worse outcomes in patients exposed to photon radiation therapy. To date, a head-to-head comparison of oral microbiome changes at a metagenomic level with longitudinal sampling has yet to be performed in these patient cohorts. To comparatively analyze oral microbiome shifts during head and neck radiation therapy, a prospective pilot cohort study was performed at the Maryland Proton Treatment Center and the University of Maryland Marlene and Stewart Greenebaum Comprehensive Cancer Center. A longitudinal metagenomic comparative analysis of oral microbiome shifts was performed at three time points (pre-radiation, during radiation, and immediately post-radiation). Head and neck cancer patients receiving proton radiation (n = 4) were compared to photon radiation (n = 4). Additional control groups included healthy age- and sex-matched controls (n = 5), head and neck cancer patients who never received radiation therapy (n = 8), and patients with oral inflammatory disease (n = 3). Photon therapy patients presented with lower microbial alpha diversity at all timepoints, and there was a trend towards reduced species richness as compared with proton therapy. Healthy controls and proton patients exhibited overall higher and similar diversity. A more dysbiotic state was observed in patients receiving photon therapy as compared to proton therapy, in which oral microbial homeostasis was maintained. Mucositis was observed in 3/4 photon patients and was not observed in any proton patients during radiation therapy. The bacterial de novo pyrimidine biosynthesis pathway and the nitrate reduction V pathway were comparatively higher following photon exposure. These functional changes in bacterial metabolism may suggest that photon exposure produces a more permissive environment for the proliferation of pathogenic bacteria. Oral microbiome dysbiosis in patients receiving photon radiation may be associated with increased mucositis occurrence. Proton radiation therapy for head and neck cancer demonstrates a safer side effect profile in terms of oral complications, oral microbiome dysbiosis, and functional metabolic status.

Proton Therapy Mediates Dose Reductions to Brain Structures Associated With Cognition in Children With Medulloblastoma

Background: Emerging evidence suggests proton radiotherapy may offer cognitive sparing advantages over photon radiotherapy, yet dosimetry has not been compared previously. The purpose of this study is to examine dosimetric correlates of cognitive outcomes in children with medulloblastoma treated with proton vs. photon radiotherapy.Methods: In this retrospective, bi-institutional study, dosimetric and cognitive data from 75 patients (39 photon and 36 proton) were analyzed. Doses to brain structures were compared between treatment modalities. Linear mixed effects models were used to create models of global IQ and cognitive domain scores.Results: The mean dose and dose to 40% of the brain (D40) were 2.7 and 4.1 Gy less among proton-treated patients as compared to photon-treated patients (p = 0.03 and 0.007, respectively). Mean doses to the left and right hippocampi were 11.2 Gy lower among proton-treated patients (p < 0.001 for both). Mean doses to the left and right temporal lobes were 6.9 and 7.1 Gy lower with proton treatment, respectively (p < 0.001 for both). Models of cognition found statistically significant associations between higher mean brain dose and reduced verbal comprehension, increased right temporal lobe D40 with reduced perceptual reasoning, and greater left temporal mean dose with reduced working memory. Higher brain D40 was associated with reduced processing speed and global IQ scores.Conclusions: Proton therapy reduces doses to normal brain structures as compared with photon treatment. This leads to reduced cognitive decline post-RT across multiple intellectual endpoints. Proton therapy should be offered to children receiving radiation for medulloblastoma.-Proton therapy for medulloblastoma delivers less dose to brain structures.-Less dose to brain structures leads to cognitive gain.-Children with medulloblastoma should be offered proton therapy.

Direct seawater splitting by photo-piezoelectric coupling based on surface protonated perovskite

Photocatalytic water splitting for hydrogen evolution has become one of the most important and effective ways to produce green and renewable hydrogen energy, the centre of which is the efficient, stable and inexpensive photocatalyst. The part of perovskite oxides have a wide range of properties such as ferroelectricity, piezoelectricity and pyroelectricity, as well as good photoresponsivity and chemical stability, which are promising for catalytic conversion applications. In this study, the Ruddlesden-Popper (RP) phase perovskite La2NiO4 is investigated from enhanced charge separation as well as accelerated surface redox reaction rates. In the case of the existence of piezoelectric response of La2NiO4, the photo-piezoelectric coupling strategy not only realize the catalytic pure water splitting, but also greatly improve the efficiency. The lattice structure is twisted when the catalyst is subjected to external stress and the electric dipole moment is no longer zero. The resulting piezoelectric potential not only enables the free charge in the bulk phase to migrate directionally, but also inhibits the recombination of photogenerated carriers, which ultimately achieves highly efficient photo-piezoelectric coupling for water splitting for hydrogen production. In addition, the protonation of the catalyst surface is used to further enhance the interfacial reaction of the catalysts, which improves the solution-catalysts contact efficiency and generates vacancy defects to strengthen the spontaneous polarization of the lattices, contributing to the enhancement of the redox reaction rate. This optimization strategy of surface protonation treatment combined with photo-piezoelectric coupling is not only applicable to nanocatalyst powders but also has a significant enhancement effect in perovskite ceramic membranes. When La2NiO4 powders are sintered into the porous ceramic membrane and the surface is protonated with 0.006 mol L−1 hydrochloric acid, the hydrogen production rate of seawater splitting under the photo-piezoelectric coupling condition can reach 7750.41 μmol m−2 h−1. This study contributes to further understanding of photo-piezoelectric coupling catalysis, leading to a new solutions for expanding the application of coupled multifield water splitting for hydrogen evolution.

Is noncoplanar plan more robust to inter-fractional variations than coplanar plan in treating bilateral HN tumors with pencil-beam scanning proton beams?

Noncoplanar plans (NCPs) are commonly used for proton treatment of bilateral head and neck (HN) malignancies. NCP requires additional verification setup imaging between beams to correct residual errors of robotic couch motion, which increases imaging dose and total treatment time. This study compared the quality and robustness of NCPs with those of coplanar plans (CPs). Under an IRB-approved study, CPs were created retrospectively for 10 bilateral HN patients previously treated with NCPs maintaining identical beam geometry of the original plan but excluding couch rotations. Plan robustness to the inter-fractional variation (IV) of both plans was evaluated through the Dose Volume Histograms (DVH) of weekly quality assurance CT (QACT) sets (39 total). In addition, delivery efficiency for both plans was compared using total treatment time (TTT) and beam-on time (BOT). No significant differences in plan quality were observed in terms of clinical target volume (CTV) coverage (D95) or organ-at-risk (OAR) doses (p>0.4 for all CTVs and OARs). No significant advantage of NCPs in the robustness to IV was found over CP, either. Changes in D95 of QA plans showed a linear correlation (slope=1.006, R2 >0.99) between NCP and CP for three CTV data points (CTV1, CTV2, and CTV3) in each QA plan (117 data points for 39 QA plans). NCPs showed significantly higher beam delivery time than CPs for TTT (539±50vs. 897±142s; p<0.001); however, no significant differences were observed for BOT. NCPs are not more robust to IV than CPs when treating bilateral HN tumors with pencil-beam scanning proton beams. CPs showed plan quality and robustness similar to NCPs while reduced treatment time (∼6min). This suggests that CPs may be a more efficient planning technique for bilateral HN cancer proton therapy.

Feasibility of Monte Carlo dropout-based uncertainty maps to evaluate deep learning-based synthetic CTs for adaptive proton therapy.

Deep learning has shown promising results to generate MRI-based synthetic CTs and to enable accurate proton dose calculations on MRIs. For clinical implementation of synthetic CTs, quality assurance tools that verify their quality and reliability are required but still lacking. This study aims to evaluate the predictive value of uncertainty maps generated with Monte Carlo dropout (MCD) for verifying proton dose calculations on deep-learning-based synthetic CTs (sCTs) derived from MRIs in online adaptive proton therapy. Two deep-learning models (DCNN and cycleGAN) were trained for CT image synthesis using 101 paired CT-MR images. sCT images were generated using MCD for each model by performing 10 inferences with activated dropout layers. The final sCT was obtained by averaging the inferred sCTs, while the uncertainty map was obtained from the HU variance corresponding to each voxel of 10 sCTs. The resulting uncertainty maps were compared to the observed HU-, range-, WET-, and dose-error maps between the sCT and planning CT. For range and WET errors, the generated uncertainty maps were projected along the 90-degree angle. To evaluate the dose distribution, a mask based on the 5%-isodose curve was applied to only include voxels along the beam paths. Pearson's correlation coefficients were calculated to determine the correlation between the uncertainty maps and HUs, range, WET, and dose errors. To evaluate the dosimetric accuracy of synthetic CTs, clinical proton treatment plans were recalculated and compared to the pCTs RESULTS: Evaluation of the correlation showed an average of r = 0.92 ± 0.03 and r = 0.92 ± 0.03 for errors between uncertainty-HU, r = 0.66 ± 0.09 and r = 0.62 ± 0.06 between uncertainty-range, r = 0.64 ± 0.06 and r = 0.58 ± 0.07 between uncertainty-WET, and r = 0.65 ± 0.09 and r = 0.67 ± 0.07 between uncertainty and dose difference for DCNN and cycleGAN model, respectively. Dosimetric comparison for target volumes showed an average 3%/3mm gamma pass rate of 99.76 ± 0.43 (DCNN) and 99.10 ± 1.27 (cycleGAN). The observed correlations between uncertainty maps and the various metrics (HU, range, WET, and dose errors) demonstrated the potential of MCD-based uncertainty maps as a reliable QA tool to evaluate the accuracy of deep learning-based sCTs.

A systematic approach for calibrating a Monte Carlo code to a treatment planning system for obtaining dose, LET, variable proton RBE and out-of-field dose

Objective. While integration of variable relative biological effectiveness (RBE) has not reached full clinical implementation, the importance of having the ability to recalculate proton treatment plans in a flexible, dedicated Monte Carlo (MC) code cannot be understated . Here we provide a step-wise method for calibrating dose from a MC code to a treatment planning system (TPS), to obtain required parameters for calculating linear energy transfer (LET), variable RBE and in general enabling clinical realistic research studies beyond the capabilities of a TPS. Approach. Initially, Pristine Bragg peaks (PBP) were calculated in both the Eclipse TPS and the FLUKA MC code. A rearranged Bortfeld energy-range relation was applied to the initial energy of the beam to fine-tune the range of the MC code at 80% dose level distal to the PBP. The energy spread was adapted by dividing the TPS range by the MC range for dose level 80%–20% distal to the PBP. Density and relative proton stopping power were adjusted by comparing the TPS and MC for different Hounsfield units. To find the relationship of dose per primary particle from the MC to dose per monitor unit in the TPS, integration was applied to the area of the Bragg curve. The calibration was validated for spread-out Bragg peaks (SOBP) in water and patient treatment plans. Following the validation, variable RBE were calculated using established models. Main results. The PBPs ranges were within ± mm threshold, and a maximum of 5.5% difference for the SOBPs was observed. The patient validation showed excellent dose agreement between the TPS and MC, with the greatest differences for the lung tumor patient. Significance. A procedure for calibrating a MC code to a TPS was developed and validated. The procedure enables MC-based calculation of dose, LET, variable RBE, advanced (secondary) particle tracking and more from treatment plans.

High-resolution three-dimensional dosimetry in clinically relevant volumes utilizing optically stimulated luminescence.

The continued development of new radiotherapy techniques requires dosimetry systems that satisfy increasingly rigorous requirements, such as high sensitivity, wide dose range, and high spatial resolution. An emerging requirement is the ability to read out doses in three dimensions (3D) with high precision and spatial resolution. A few dosimetry systems with 3D capabilities are available, but their application in a clinical workflow is limited for various reasons, primarily originating from their chemical nature. The search for a 3D dosimetry system with potential for clinical implementation is thusongoing. To demonstrate the capabilities of a novel optically-stimulated-luminescence (OSL)-based 3D dosimetry system capable of measuring radiation doses in clinically relevantvolumes. A laser-based readout system was used to measure dose distributions delivered by both photons and protons, utilizing the OSL from a mm YSO:Ce crystal. A homogeneous treatment plan consisting of two opposing photon fields was used to establish an inhomogeneity correction map of the crystal response and demonstrated the accuracy and precision of the system. The crystal was additionally irradiated with a photon treatment plan consisting of three overlapping mm fields delivered from different angles, and a proton treatment plan consisting of four pencil beams with energies 90 MeV ( ), 115 MeV, and 140 MeV. The system abilities were quantified by comparing the 3D-resolved measurements to Monte Carlosimulations. The dose map reproducibility of the system was found to be within 2% including both statistical and systematic errors. The measurements yielded integrated doses from a volume of mm with voxel volumes of just mm . An excellent agreement between the 3D-resolved measurements and the simulations was found for both photon- andproton-irradiation. The capabilities of the devised system for measuring clinically relevant fields of photons and proton pencil beams within a clinically relevant volume were demonstrated. The system poses as a promising candidate for clinical applications, and enables future research in the field of OSL-based tissue-equivalent 3Ddosimetry.

Estimating the Dimensions of the Radiological Field Suitable For the Treatment of Cancerous Tumors by Simulating a Proton Treatment System and Applying It to a Water Phantom Using the MCNP Code

Abstract: In this paper, the proton therapy system was simulated using the MCNPX code, and an analysis of the dimensions of the resulting radioactive field and dose distribution within the radiation field was conducted using the double beam scattering technique. The energy deposited in terms of depth was also studied within a water phantom with dimensions of 11×11×30〖 cm〗^3 divided into elemental volumes, each measuring 1x1x0.4〖 cm〗^3. This analysis was performed using a modified acceleration for range, where each cycle of the modified wheel was divided into 25 steps, with each step represented by an equivalent thickness of lead and Lexan. The study revealed the possibility of obtaining a radioactive field with dimensions of 7 x 7〖 cm〗^2 , concentrating about 95% to 100% of the maximum dose. Additionally, it was also found that the dose distribution can extend to a depth of 20 cm within the phantom. Keywords: Protons, MCNPX, Proton therapy system, Radiation dose.

A TOPAS model for lens-based proton radiography

Objective. Proton Radiography can be used in conjunction with proton therapy for patient positioning, real-time estimates of stopping power, and adaptive therapy in regions with motion. The modeling capability shown here can be used to evaluate lens-based radiography as an instantaneous proton-based radiographic technique. The utilization of user-friendly Monte Carlo program TOPAS enables collaborators and other users to easily conduct medical- and therapy- based simulations of the Los Alamos Neutron Science Center (LANSCE). The resulting transport model is an open-source Monte Carlo package for simulations of proton and heavy ion therapy treatments and concurrent particle imaging. Approach. The four-quadrupole, magnetic lens system of the 800-MeV proton beamline at LANSCE is modeled in TOPAS. Several imaging and contrast objects were modelled to assess transmission at energies from 230–930 MeV and different levels of particle collimation. At different proton energies, the strength of the magnetic field was scaled according to βγ, the inverse product of particle relativistic velocity and particle momentum. Main results. Materials with high atomic number, Z, (gold, gallium, bone-equivalent) generated more contrast than materials with low-Z (water, lung-equivalent, adipose-equivalent). A 5-mrad collimator was beneficial for tissue-to-contrast agent contrast, while a 10-mrad collimator was best to distinguish between different high-Z materials. Assessment with a step-wedge phantom showed water-equivalent path length did not scale directly according to predicted values but could be mapped more accurately with calibration. Poor image quality was observed at low energies (230 MeV), but improved as proton energy increased, with sub-mm resolution at 630 MeV. Significance. Proton radiography becomes viable for shallow bone structures at 330 MeV, and for deeper structures at 630 MeV. Visibility improves with use of high-Z contrast agents. This modality may be particularly viable at carbon therapy centers with accelerators capable of delivering high energy protons and could be performed with carbon therapy.

The dirty and clean dose concept: Towards creating proton therapy treatment plans with a photon-like dose response.

Applying tolerance doses for organs at risk (OAR) from photon therapy introduces uncertainties in proton therapy when assuming a constant relative biological effectiveness (RBE) of 1.1. This work introduces the novel dirty and clean dose concept, which allows for creating treatment plans with a more photon-like dose response for OAR and, thus, less uncertainties when applying photon-based tolerance doses. The concept divides the 1.1-weighted dose distribution into two parts: the clean and the dirty dose. The clean and dirty dose are deposited by protons with a linear energy transfer (LET) below and above a set LET threshold, respectively. For the former, a photon-like dose response is assumed, while for the latter, the RBE might exceed 1.1. To reduce the dirty dose in OAR, a MaxDirtyDose objective was added in treatment plan optimization. It requires setting two parameters: LET threshold and max dirty dose level. A simple geometry consisting of one target volume and one OAR in water was used to study the reduction in dirty dose in the OAR depending on the choice of the two MaxDirtyDose objective parameters during plan optimization. The best performing parameter combinations were used to create multiple dirty dose optimized (DDopt) treatment plans for two cranial patient cases. For each DDopt plan, 1.1-weighted dose, variable RBE-weighted dose using the Wedenberg RBE model and dose-average LETd distributions as well as resulting normal tissue complication probability (NTCP) values were calculated and compared to the reference plan (RefPlan) without MaxDirtyDose objectives. In the water phantom studies, LET thresholds between 1.5 and 2.5keV/µm yielded the best plans and were subsequently used. For the patient cases, nearly all DDopt plans led to a reduced Wedenberg dose in critical OAR. This reduction resulted from an LET reduction and translated into an NTCP reduction of up to 19 percentage points compared to the RefPlan. The 1.1-weighted dose in the OARs was slightly increased (patient 1: 0.45Gy(RBE), patient 2: 0.08Gy(RBE)), but never exceeded clinical tolerance doses. Additionally, slightly increased 1.1-weighted dose in healthy brain tissue was observed (patient 1: 0.81Gy(RBE), patient 2: 0.53Gy(RBE)). The variation of NTCP values due to variation of α/β from 2 to 3Gy was much smaller for DDopt (2 percentage points (pp)) than for RefPlans (5 pp). The novel dirty and clean dose concept allows for creating biologically more robust proton treatment plans with a more photon-like dose response. The reduced uncertainties in RBE can, therefore, mitigate uncertainties introduced by using photon-based tolerance doses for OAR.

AI Beam Angle Prediction in Proton Therapy for Brain Patients.

Patients receiving brain Intensity-modulated Proton Therapy (IMPT) are treated with 1-3 fields. Typically, beam angles must be selected manually, requiring substantial planner time and experience. The choice of beam angles has a major impact on the treatment plan quality. An AI model was developed to automate this process aiming to improve efficiency and potentially plan quality. AI performance was tested by comparing predicted beam angles to the human ground truth angles and evaluating plans resulting from each set of beam angles. Previously treated IMPT brain patients were divided into a training (n = 40) and validation set (n = 10) for the AI. Beam-angle selection in the AI was cast as a multi-label classification problem, training the convolution neural network with a Circular Earth Mover's Distance based regularization and multi-label circular-smooth label technique. An analytical post-processing algorithm is employed to minimize distance to the target and avoid organs-at-risk. Finalized gantry angle predictions are discretized in steps of 5°. For an independent test set of n = 10 patients beam angles were predicted by the AI and compared to those chosen by human planners. Both sets of angles were used to create treatment plans with an automated knowledge-based planning (KBP) tool for brain IMPT utilizing single-field optimization and robust optimization. Plan differences are therefore solely due to the choice of beam angles as automated KBP removes human optimization variability. Resulting plan quality was compared by standard clinical dosimetric parameters to the CTV, brain, brainstem, cochlea, cord, eye, lens, optic chiasm, optic nerves, pituitary, and temporal lobes. All Human and AI selected beam angles are shown in the table. For cases 1-6, AI and Human selected beam angles were within ±15° and resulting plans showed minimal dosimetric differences. In cases 7-9 AI beam angles reduced max dose while keeping organ-at-risk dose within ±2 Gy. In case 10 the AI chosen beam angles reduced cord Dmax by 9.1 Gy but increased left eye Dmax by 4.9 Gy, other organs-at-risk showed minimal differences. Gantry beam angle selection was automated by a newly developed AI model and tested on 10 brain IMPT patients. The comparison showed that the AI often chooses similar beam angles to the human planners. When there are differences, dosimetric analysis demonstrated that plans created from AI beam angles have at least the same quality as the human ones. Results motivate further research into this approach showing the AI being a promising tool to fill a current gap in the strive for automating proton treatment planning, increasing planning efficiency and potentially quality.

Clinical and Dosimetric Results of Proton or Photon Radiation Therapy for Large (>5 cm) Hepatocellular Carcinoma: A Retrospective Analysis

Purpose: To report the clinical and dosimetric attributes of patients with large unresectable hepatocellular carcinoma (HCC) undergoing proton or photon radiotherapy. Methods: We retrospectively analyzed the outcomes and dosimetric indices of 159 patients with >5 cm non-metastatic HCC who underwent definitive radiotherapy using either protons (N=105) or photons (N=54) between 2014 and 2018. Additional photon plans were performed in the 105 proton-treated patients using the same dose prescription criteria for intragroup dosimetric comparison. Results: After a median follow-up of 47 months, patients with biologically effective dose (BED10) ≥75 Gy exhibited significantly better local control (LC, 2-year: 85.6% vs. 20.5%, P<0.001), progression-free survival (PFS, median 7.4 vs. 3.2 months, P<0.001), and overall survival (OS, median 18.1 vs. 7.3 months, P<0.001) compared to those with BED10 <75 Gy. Notably, proton-treated patients had a significantly higher BED10 (96 vs. 67 Gy, P<0.001) and improved LC (2-year: 88.5% vs. 33.8%, P<0.001), PFS (median 7.4 vs. 3.3 months, P=0.001), and OS (median 18.9 vs. 8.3 months, P<0.001) than those undergoing photon radiotherapy. Furthermore, patients treated with protons had significantly lower V1 of the liver (P<0.001), mean upper gastrointestinal (UGI) tract dose (P<0.001), and mean splenic dose (P<0.001), with significantly decreased incidences of radiation-induced liver disease (P=0.007), grade ≥3 UGI bleeding (P=0.001), and grade ≥3 lymphopenia (P=0.003). On multivariate analysis, proton radiotherapy consistently correlated with superior LC (P<0.001), PFS (P<0.001), and OS (P<0.001). In intragroup dosimetric comparison, photon plans demonstrated significantly higher mean liver dose (MLD, P<0.001) compared with actually delivered proton treatments, and 72 (69%) of them had MLD exceeding 28 Gy which necessitated target dose de-escalation. Conclusions: In the context of large HCC radiotherapy, a higher target BED10 was associated with improved outcomes. Notably, proton therapy has demonstrated the capability to deliver ablative doses while also being accompanied by fewer instances of severe toxicity.

Selection of breathing phase number in 4D scanned proton treatment planning optimization for lung tumors

The standard four-dimensional (4D) treatment planning includes all breathing states in the optimization process, which is time-consuming. This work was aimed to optimize the number of intermediate phases needed for 4D proton treatment planning optimization to reduce the computational cost. Five 4D optimization strategies adopting different numbers of intermediate states and one three-dimensional (3D) optimization plan were studied for fifteen lung cancer patients treated with scanned protons, optimizing on all ten phases (4D_10), two extreme phases (4D_2), six phases during the exhalation stage (4D_6EX), six phases during the inhalation stage (4D_6IN), two extreme phases plus an intermediate state (4D_3) and average computed tomography image (3D), respectively. The 4D dose evaluation was conducted on all the ten phases, considering the interplay effect. The resulting doses accumulated on the reference phase were computed and compared. Compared to the 4D optimization plans, the 3D optimization plan performed inferiorly in target coverage, but superiorly in organ at risks (OARs) sparing. For the 4D optimization, all the five 4D plans showed similar performance in OARs protection. However, the 4D_6EX and 4D_6IN strategies out-performed the 4D_2 and 4D_3 plans in dose homogeneity. The computing times of the 4D_2, 4D_3, 4D_6EX and 4D_6IN approaches decreased to 32%, 41%, 66% and 67% of the 4D_10 method, respectively. Thus, our study suggested that the use of all phases during inhalation or exhalation stage might be a feasible approach substituting for the full phase strategy to reduce the calculation load while guaranteeing the plan quality for scanned proton therapy.

Inaugural Experience with Real-time Gated Liver Proton SBRT and Treatment Plan Quality Improvement

DIBH SBRT is routinely used for liver proton therapy. While intra-fraction target motion is limited with DIBH, acquisition of DIBH CT simulations in triplicate, as is done at our institution, reveals that variation does exist between each DIBH scan. The related target position can also vary correspondingly. The most common setup uncertainty for robust proton SBRT liver plan used at our institution is 5 mm sup-inf (SI) and 3 mm radially. Real-time gated proton therapy (RGPT) has the potential to provide instantaneous feedback for intra-fraction target motion to maximize patient safety and inform optimal treatment planning. Our first RGPT liver SBRT with intra-fraction motion under deep inspiration breath hold (DIBH). The potential treatment plan quality improvement brought by RGPT is investigated. The following metrics were used in establishing our RGPT proton DIBH SBRT liver program: the iso center is always set at the fiducial mark; the beam orientation is selected to achieve both good plan quality and tracking performance; daily CBCTs are acquired and verified using fiducial maker position with kV images; robust uncertainty is determined by the gating tolerance; SBRT plan has three beams with uniform dose. Target motion was monitored throughout treatment. To evaluate dose sparing for surrounding OARs, a plan with tighter gating tolerance (3 mm SI and 2 mm radially) is optimized for dosimetric comparison. Statistical analyses were conducted using a programming environment. Each of the three proton beams were delivered using DIBH over a total of 120-140 seconds. The average beam on time were 61.4, 66.9 and 62.8 seconds. The intra-fraction motion showed that targets could move up to 3 mm within the same DIBH. The motion increased with time. The table details the mean, maximum, standard deviation, and estimated upper 95% of directional shifts for three beams. Based on these results, plan delivery efficiency was maintained even with tighter gating tolerance. The comparison plan with tight gating tolerance showed significantly less dose (-25%) to the stomach in coronal view. RGPT successfully tracked fiducial marker motion for DIBH SBRT liver treatment. Despite target drift during DIBH, the uncertainty of our DIBH SBRT procedure was sufficient to cover target motion throughout treatment. Based on the target drift value, a maximum of 25 seconds for breath hold time should be employed. Utilizing a tighter gating tolerance of 3 mm SI and 2 mm radially has the potential to maintain target coverage while significantly reducing OAR dose. Aggregated RGPT-derived data may provide optimal treatment planning parameters such as variable uncertainty based on target location.