A deep-learning-based surrogate model for Monte-Carlo simulations of the linear energy transfer in primary brain tumor patients treated with proton-beam radiotherapy.

Objective.This study explores the use of neural networks (NNs) as surrogate models for Monte-Carlo (MC) simulations in predicting the dose-averaged linear energy transfer (LETd) of protons in proton-beam therapy based on the planned dose distribution and patient anatomy in the form of computed tomography (CT) images. As LETdis associated with variability in the relative biological effectiveness (RBE) of protons, we also evaluate the implications of using NN predictions for normal tissue complication probability (NTCP) models within a variable-RBE context.Approach.The predictive performance of three-dimensional NN architectures was evaluated using five-fold cross-validation on a cohort of brain tumor patients (n= 151). The best-performing model was identified and externally validated on patients from a different center (n= 107). LETdpredictions were compared to MC-simulated results in clinically relevant regions of interest. We assessed the impact on NTCP models by leveraging LETdpredictions to derive RBE-weighted doses, using the Wedenberg RBE model.Main results.We found NNs based solely on the planned dose distribution, i.e. without additional usage of CT images, can approximate MC-based LETddistributions. Root mean squared errors (RMSE) for the median LETdwithin the brain, brainstem, CTV, chiasm, lacrimal glands (ipsilateral/contralateral) and optic nerves (ipsilateral/contralateral) were 0.36, 0.87, 0.31, 0.73, 0.68, 1.04, 0.69 and 1.24 keV µm-1, respectively. Although model predictions showed statistically significant differences from MC outputs, these did not result in substantial changes in NTCP predictions, with RMSEs of at most 3.2 percentage points.Significance.The ability of NNs to predict LETdbased solely on planned dose distributions suggests a viable alternative to compute-intensive MC simulations in a variable-RBE setting. This is particularly useful in scenarios where MC simulation data are unavailable, facilitating resource-constrained proton therapy treatment planning, retrospective patient data analysis and further investigations on the variability of proton RBE.

Modelling of RBE differences in selected points within similar spread-out Bragg-peaks (SOBP) placed at superficial and deep water phantom locations in passively scattered beams but not in scanned pencil beams: A hypothesis.

PURPOSE: To model relative biological effectiveness (RBE) differences found in two studies which used spread-out Bragg-peaks (SOBP) placed at (a) superficial depth and (b) at the maximum range depth. For pencil beam scanning (PBS), RBE at similar points within the SOBP did not change between the two extreme SOBP placement depths; in passively scattered beams (PSB), high RBE values (typically 1.2-1.3) were found within superficially- placed SOBP but reduced to lower values (1-1.07) at similar points within the extreme-depth positioned SOBP. The dose, LET (linear energy transfer) distributions along each SOBP were closely comparable regardless of placement depth, but significant changes in dose rate occurred with depth in the PSB beam. METHODS: The equations used allow α and ß changes with falling dose rate (the converse to FLASH studies) in PSB, resulting in reduced α/ß ratios, compatible with a reduction in micro-volumetric energy transfer (the product of Fluence and LET), with commensurate reductions in RBE. The experimental depth-distances, positions within SOBP, observed dose-rates and radiosensitivity ratios were used to estimate the changes in RBE. RESULTS: RBE values within a 5 % tolerance limit of the experimental results for PSB were found at the deepest SOBP placement. No RBE changes were predicted for PBS beams, as in the published results. CONCLUSIONS: Enhanced proton therapy toxicity might occur with PBS when compared with PSB for deeply positioned SOBP due to the maintenance of higher RBE. Scanned pencil beam users need to be vigilant about RBE and further research is indicated.

Particle tracking, recognition and LET evaluation of out-of-field proton therapy delivered to a phantom with implants.

Objective.This study aims to assess the composition of scattered particles generated in proton therapy for tumors situated proximal to some titanium (Ti) dental implants. The investigation involves decomposing the mixed field and recording Linear Energy Transfer (LET) spectra to quantify the influence of metallic dental inserts located behind the tumor.Approach.A therapeutic conformal proton beam was used to deliver the treatment plan to an anthropomorphic head phantom with two types of implants inserted in the target volume (made of Ti and plastic, respectively). The scattered radiation resulted during the irradiation was detected by a hybrid semiconductor pixel detector MiniPIX Timepix3 that was placed distal to the Spread-out Bragg peak. Visualization and field decomposition of stray radiation were generated using algorithms trained in particle recognition based on artificial intelligence neural networks (AI NN). Spectral sensitive aspects of the scattered radiation were collected using two angular positions of the detector relative to the beam direction: 0° and 60°.Results.Using AI NN, 3 classes of particles were identified: protons, electrons & photons, and ions & fast neutrons. Placing a Ti implant in the beam's path resulted in predominantly electrons and photons, contributing 52.2% of the total number of detected particles, whereas for plastic implants, the contribution was 65.4%. Scattered protons comprised 45.5% and 31.9% with and without metal inserts, respectively. The LET spectra were derived for each group of particles identified, with values ranging from 0.01 to 7.5 keVµm-1for Ti implants/plastic implants. The low-LET component was primarily composed of electrons and photons, while the high-LET component corresponded to protons and ions.Significance.This method, complemented by directional maps, holds the potential for evaluating and validating treatment plans involving stray radiation near organs at risk, offering precise discrimination of the mixed field, and enhancing in this way the LET calculation.

Cumulative Dose from Recurrent CT Scans: Exploring the DNA Damage Response in Human Non-Transformed Cells.

Recurrent computed tomography (CT) examination has become a common diagnostic procedure for several diseases and injuries. Though each singular CT scan exposes individuals at low doses of low linear energy transfer (LET) radiation, the cumulative dose received from recurrent CT scans poses an increasing concern for potential health risks. Here, we evaluated the biological effects of recurrent CT scans on the DNA damage response (DDR) in human fibroblasts and retinal pigment epithelial cells maintained in culture for five months and subjected to four CT scans, one every four weeks. DDR kinetics and eventual accumulation of persistent-radiation-induced foci (P-RIF) were assessed by combined immunofluorescence for γH2AX and 53BP1, i.e., γH2AX/53BP1 foci. We found that CT scan repetitions significantly increased both the number and size of γH2AX/53BP1 foci. In particular, after the third CT scan, we observed the appearance of giant foci that might result from the overlapping of individual small foci and that do not associate with irreversible growth arrest, as shown by DNA replication in the foci-carrying cells. Whether these giant foci represent coalescence of unrepaired DNA damage as reported following single exposition to high doses of high LET radiation is still unclear. However, morphologically, these giant foci resemble the recently described compartmentalization of damaged DNA that should facilitate the repair of DNA double-strand breaks but also increase the risk of chromosomal translocations. Overall, these results indicate that for a correct evaluation of the damage following recurrent CT examinations, it is necessary to consider the size and composition of the foci in addition to their number.

A dosiomics approach to treatment outcome modeling in carbon ion radiotherapy for skull base chordomas.

PURPOSE: To investigate the role of dosiomics features extracted from physical dose (DPHYS), RBE-weighted dose (DRBE) and dose-averaged Linear Energy Transfer (LETd), to predict the risk of local recurrence (LR) in skull base chordoma (SBC) treated with Carbon Ion Radiotherapy (CIRT). Thus, define and evaluate dosiomics-driven tumor control probability (TCP) models. MATERIALS AND METHODS: 54 SBC patients were retrospectively selected for this study. A regularized Cox proportional hazard model (r-Cox) and Survival Support Vector Machine (s-SVM) were tuned within a repeated Cross Validation (CV) and patients were stratified in low/high risk of LR. Models' performance was evaluated through Harrell's concordance statistic (C-index), and survival was represented through Kaplan-Meier (KM) curves. A multivariable logistic regression was fit to the selected feature sets to generate a dosiomics-driven TCP model for each map. These were compared to a reference model built with clinical parameters in terms of f-score and accuracy. RESULTS: The LETd maps reached a test C-index of 0.750 and 0.786 with r-Cox and s-SVM, and significantly separated KM curves. DPHYS maps and clinical parameters showed promising CV outcomes with C-index above 0.8, despite a poorer performance on the test set and patients stratification. The LETd-based TCP showed a significatively higher f-score (0.67[0.52-0.70], median[IQR]) compared to the clinical model (0.4[0.32-0.63], p < 0.025), while DPHYS achieved a significatively higher accuracy (DPHYS: 0.73[0.65-0.79], Clinical: 0.6 [0.52-0.72]). CONCLUSION: This analysis supports the role of LETd as relevant source of prognostic factors for LR in SBC treated with CIRT. This is reflected in the TCP modeling, where LETd and DPHYS showed an improved performance with respect to clinical models.

Comparison of linear energy transfer measurement for therapeutic carbon beam using CR-39 and TLD.

The measurement of linear energy transfer (LET) is crucial for the evaluation of the radiation effect in heavy ion therapy. As two detectors which are convenient to implant into the phantom, the performance of CR-39 and thermoluminescence detector (TLD) for LET measurement was compared by experiment and simulation in this study. The results confirmed the applicability of both detectors for LET measurements, but also revealed that the CR-39 detector would lead to potential overestimation of dose-averaged LET compared with the simulation by PHITS, while the TLD would have a large uncertainty measuring ions with LET larger than 20 keVµm-1. The results of this study were expected to improve the detection method of LET for therapeutic carbon beam and would finally be benefit to the quality assurance of heavy ion radiotherapy.

Comparison of mutation spectra induced by gamma-rays and carbon ion beams.

The ionizing radiation with high linear energy transfer (LET), such as a heavy ion beam, induces more serious biological effects than low LET ones, such as gamma- and X-rays. This indicates a difference in the DNA damage produced by low and high LET radiations and their biological effects. We have been studying the differences in DNA damage produced by gamma-rays and carbon ion beams. Therefore, we analyze mutations induced by both ionizing radiations to discuss the differences in their biological effects in this study. pUC19 plasmid DNA was irradiated by carbon ion beams in the solution containing 1M dimethyl sulfoxide to mimic a cellular condition. The irradiated DNA was cloned in competent cells of Escherichia coli. The clones harboring some mutations in the region of lacZα were selected, and the sequence alterations were analyzed. A one-deletion mutation is significant in the carbon-irradiated DNA, and the C:G↔T:A transition is minor. On the other hand, the gamma-irradiated DNA shows mainly G:C↔T:A transversion. These results suggest that carbon ion beams produce complex DNA damage, and gamma-rays are prone to single oxidative base damage, such as 8-oxoguanine. Carbon ion beams can also introduce oxidative base damage, and the damage species is 5-hydroxycytosine. This was consistent with our previous results of DNA damage caused by heavy ion beams. We confirmed the causal DNA damage by mass spectrometry for these mutations.

Potential Benefits of Combining Proton or Carbon Ion Therapy with DNA Damage Repair Inhibitors.

The use of charged particle radiotherapy is currently increasing, but combination therapy with DNA repair inhibitors remains to be exploited in the clinic. The high-linear energy transfer (LET) radiation delivered by charged particles causes clustered DNA damage, which is particularly effective in destroying cancer cells. Whether the DNA damage response to this type of damage is different from that elicited in response to low-LET radiation, and if and how it can be targeted to increase treatment efficacy, is not fully understood. Although several preclinical studies have reported radiosensitizing effects when proton or carbon ion irradiation is combined with inhibitors of, e.g., PARP, ATR, ATM, or DNA-PKcs, further exploration is required to determine the most effective treatments. Here, we examine what is known about repair pathway choice in response to high- versus low-LET irradiation, and we discuss the effects of inhibitors of these pathways when combined with protons and carbon ions. Additionally, we explore the potential effects of DNA repair inhibitors on antitumor immune signaling upon proton and carbon ion irradiation. Due to the reduced effect on healthy tissue and better immune preservation, particle therapy may be particularly well suited for combination with DNA repair inhibitors.

Radiation-induced DNA damage by proton, helium and carbon ions in human fibroblast cell: Geant4-DNA and MCDS-based study.

Background. Radiation-induced DNA damages such as Single Strand Break (SSB), Double Strand Break (DSB) and Complex DSB (cDSB) are critical aspects of radiobiology with implications in radiotherapy and radiation protection applications.Materials and Methods. This study presents a thorough investigation into the effects of protons (0.1-100 MeV/u), helium ions (0.13-100 MeV/u) and carbon ions (0.5-480 MeV/u) on DNA of human fibroblast cells using Geant4-DNA track structure code coupled with DBSCAN algorithm and Monte Carlo Damage Simulations (MCDS) code. Geant4-DNA-based simulations consider 1µm × 1µm × 0.5µm water box as the target to calculate energy deposition on event-by-event basis and the three-dimensional coordinates of the interaction location, and then DBSCAN algorithm is used to calculate yields of SSB, DSB and cDSB in human fibroblast cell. The study investigated the influence of Linear Energy Transfer (LET) of protons, helium ions and carbon ions on the yields of DNA damages. Influence of cellular oxygenation on DNA damage patterns is investigated using MCDS code.Results. The study shows that DSB and SSB yields are influenced by the LET of the particles, with distinct trends observed for different particles. The cellular oxygenation is a key factor, with anoxic cells exhibiting reduced SSB and DSB yields, underscoring the intricate relationship between cellular oxygen levels and DNA damage. The study introduced DSB/SSB ratio as an informative metric for evaluating the severity of radiation-induced DNA damage, particularly in higher LET regions.Conclusions. The study highlights the importance of considering particle type, LET, and cellular oxygenation in assessing the biological effects of ionizing radiation.

Novel radiation quality metrics accounting for proton energy spectra for RBE proton models.

BACKGROUND: For proton therapy, a relative biological effectiveness (RBE) of 1.1 is widely applied clinically. However, due to abundant evidence of variable RBE in vitro, and as suggested in studies of patient outcomes, RBE might increase by the end of the proton tracks, as described by several proposed variable RBE models. Typically, the dose averaged linear energy transfer ( LET d $\text{LET}_d$ ) has been used as a radiation quality metric (RQM) for these models. However, the optimal choice of RQM has not been fully explored. PURPOSE: This study aims to propose novel RQMs that effectively weight protons of different energies, and assess their predictive power for variable RBE in proton therapy. The overall objective is to identify an RQM that better describes the contribution of individual particles to the RBE of proton beams. METHODS: High-throughput experimental set-ups of in vitro cell survival studies for proton RBE determination are simulated utilizing the SHIELD-HIT12A Monte Carlo particle transport code. For every data point, the proton energy spectra are simulated, allowing the calculation of novel RQMs by applying different power levels to the spectra of LET or effective Q $Q$ ( Q eff $Q_\mathrm{eff}$ ) values. A phenomenological linear-quadratic-based RBE model is then applied to the in vitro data, using various RQMs as input variables, and the model performance is evaluated by root-mean-square-error (RMSE) for the logarithm of cell surviving fractions of each data point. RESULTS: Increasing the power level, that is, putting an even higher weight on higher LET particles when constructing the RQM is generally associated with an increased model performance, with dose averaged LET 3 $\text{LET}^3$ (i.e., dose averaged cubed LET, cLET d $\mathrm{cLET}_d$ ) resulting in a RMSE value 0.31, compared to 0.45 for a model based on (linearly weighted) LET d $\text{LET}_d$ , with similar trends also observed for track averaged and Q eff $Q_\mathrm{eff}$ -based RQMs. CONCLUSIONS: The results indicate that improved proton variable RBE models can be constructed assuming a non-linear RBE(LET) relationship for individual protons. If similar trends hold also for an in vitro-environment, variable RBE effects are likely better described by cLET d $\mathrm{cLET}_d$ or tracked averaged cubed LET ( cLET t $\mathrm{cLET}_t$ ), or corresponding Q eff $Q_\mathrm{eff}$ -based RQM, rather than linearly weighted LET d $\text{LET}_d$ or LET t $\text{LET}_t$ which is conventionally applied today.

Modeling Radiation-Induced Epithelial Cell Injury in Murine Three-Dimensional Esophageal Organoids.

Esophageal squamous cell carcinoma (ESCC) is a deadly consequence of radiation exposure to the esophagus. ESCC arises from esophageal epithelial cells that undergo malignant transformation and features a perturbed squamous cell differentiation program. Understanding the dose- and radiation quality-dependence of the esophageal epithelium response to radiation may provide insights into the ability of radiation to promote ESCC. We have explored factors that may play a role in esophageal epithelial radiosensitivity and their potential relationship to ESCC risk. We have utilized a murine three-dimensional (3D) organoid model that recapitulates the morphology and functions of the stratified squamous epithelium of the esophagus to study persistent dose- and radiation quality-dependent changes. Interestingly, although high-linear energy transfer (LET) Fe ion exposure induced a more intense and persistent alteration of squamous differentiation and 53BP1 DNA damage foci levels as compared to Cs, the MAPK/SAPK stress pathway signaling showed similar altered levels for most phospho-proteins with both radiation qualities. In addition, the lower dose of high-LET exposure also revealed nearly the same degree of morphological changes, even though only ~36% of the cells were predicted to be hit at the lower 0.1 Gy dose, suggesting that a bystander effect may be induced. Although p38 and ERK/MAPK revealed the highest levels following high-LET exposure, the findings reveal that even a low dose (0.1 Gy) of both radiation qualities can elicit a persistent stress signaling response that may critically impact the differentiation gradient of the esophageal epithelium, providing novel insights into the pathogenesis of radiation-induced esophageal injury and early stage esophageal carcinogenesis.

RadPhysBio: A Radiobiological Database for the Prediction of Cell Survival upon Exposure to Ionizing Radiation.

Based on the need for radiobiological databases, in this work, we mined experimental ionizing radiation data of human cells treated with X-rays, γ-rays, carbon ions, protons and α-particles, by manually searching the relevant literature in PubMed from 1980 until 2024. In order to calculate normal and tumor cell survival α and ß coefficients of the linear quadratic (LQ) established model, as well as the initial values of the double-strand breaks (DSBs) in DNA, we used WebPlotDigitizer and Python programming language. We also produced complex DNA damage results through the fast Monte Carlo code MCDS in order to complete any missing data. The calculated α/ß values are in good agreement with those valued reported in the literature, where α shows a relatively good association with linear energy transfer (LET), but not ß. In general, a positive correlation between DSBs and LET was observed as far as the experimental values are concerned. Furthermore, we developed a biophysical prediction model by using machine learning, which showed a good performance for α, while it underscored LET as the most important feature for its prediction. In this study, we designed and developed the novel radiobiological 'RadPhysBio' database for the prediction of irradiated cell survival (α and ß coefficients of the LQ model). The incorporation of machine learning and repair models increases the applicability of our results and the spectrum of potential users.

Dose and DNA damage modelling of diffusing alpha-emitters radiation therapy using Geant4.

PURPOSE: Diffusing alpha-emitters radiation therapy (DaRT) is a brachytherapy technique using α-particles to treat solid tumours. The high linear energy transfer (LET) and short range of α-particles make them good candidates for the targeted treatment of cancer. Treatment planning of DaRT requires a good understanding of the dose from α-particles and the other particles released in the 224Ra decay chain. METHODS: The Geant4 Monte Carlo toolkit has been used to simulate a DaRT seed to better understand the dose contribution from all particles and simulate the DNA damage due to this treatment. RESULTS: Close to the seed α-particles deliver the majority of dose, however at radial distances greater than 4 mm, the contribution of ß-particles is greater. The RBE has been estimated as a function of number of double strand breaks (DSBs) and complex DSBs. A maximum seed spacing of 5.5 mm and 6.5 mm was found to deliver at least 20 Gy RBE weighted dose between the seeds for RBEDSB and RBEcDSB respectively. CONCLUSIONS: The DNA damage changes with radial distance from the seed and has been found to become less complex with distance, which is potentially easier for the cell to repair. Close to the seed α-particles contribute the majority of dose, however the contribution from other particles cannot be neglected and may influence the choice of seed spacing.

Towards precise LET measurements based on energy deposition of therapeutic ions in Timepix3 detectors.

Objective.There is an increasing interest in calculating and measuring linear energy transfer (LET) spectra in particle therapy in order to assess their impact in biological terms. As such, the accuracy of the particle fluence energy spectra becomes paramount. This study focuses on quantifying energy depositions of distinct proton, helium, carbon, and oxygen ion beams using a silicon pixel detector developed at CERN to determine LET spectra in silicon.Approach.While detection systems have been investigated in this pursuit, the scarcity of detectors capable of providing per-ion data with high spatial and temporal resolution remains an issue. This gap is where silicon pixel detector technology steps in, enabling online tracking of single-ion energy deposition. The used detector consisted of a 300µm thick silicon sensor operated in partial depletion.Main results.During post-processing, artifacts in the acquired signals were identified and methods for their corrections were developed. Subsequently, a correlation between measured and Monte Carlo-based simulated energy deposition distributions was performed, relying on a two-step recalibration approach based on linear and saturating exponential models. Despite the observed saturation effects, deviations were confined below 7% across the entire investigated range of track-averaged LET values in silicon from 0.77 keVµm-1to 93.16 keVµm-1.Significance.Simulated and measured mean energy depositions were found to be aligned within 7%, after applying artifact corrections. This extends the range of accessible LET spectra in silicon to clinically relevant values and validates the accuracy and reliability of the measurements. These findings pave the way towards LET-based dosimetry through an approach to translate these measurements to LET spectra in water. This will be addressed in a future study, extending functionality of treatment planning systems into clinical routine, with the potential of providing ion-beam therapy of utmost precision to cancer patients.

A perspective on tumor radiation resistance following high-LET radiation treatment.

High-linear energy transfer (LET) radiation is a promising alternative to conventional low-LET radiation for therapeutic gain against cancer owing to its ability to induce complex and clustered DNA lesions. However, the development of radiation resistance poses a significant barrier. The potential molecular mechanisms that could confer resistance development are translesion synthesis (TLS), replication gap suppression (RGS) mechanisms, autophagy, epithelial-mesenchymal transition (EMT) activation, release of exosomes, and epigenetic changes. This article will discuss various types of complex clustered DNA damage, their repair mechanisms, mutagenic potential, and the development of radiation resistance strategies. Furthermore, it highlights the importance of careful consideration and patient selection when employing high-LET radiotherapy in clinical settings.

Assessment of fluence- and dose-averaged linear energy transfer with passive luminescence detectors in clinical proton beams.

Objective.This work investigates the use of passive luminescence detectors to determine different types of averaged linear energy transfer (LET-) for the energies relevant to proton therapy. The experimental results are compared to reference values obtained from Monte Carlo simulations.Approach.Optically stimulated luminescence detectors (OSLDs), fluorescent nuclear track detectors (FNTDs), and two different groups of thermoluminescence detectors (TLDs) were irradiated at four different radiation qualities. For each irradiation, the fluence- (LET-f) and dose-averaged LET (LET-d) were determined. For both quantities, two sub-types of averages were calculated, either considering the contributions from primary and secondary protons or from all protons and heavier, charged particles. Both simulated and experimental data were used in combination with a phenomenological model to estimate the relative biological effectiveness (RBE).Main results.All types ofLET-could be assessed with the luminescence detectors. The experimental determination ofLET-fis in agreement with reference data obtained from simulations across all measurement techniques and types of averaging. On the other hand,LET-dcan present challenges as a radiation quality metric to describe the detector response in mixed particle fields. However, excluding secondaries heavier than protons from theLET-dcalculation, as their contribution to the luminescence is suppressed by ionization quenching, leads to equal accuracy betweenLET-fandLET-d. Assessment of RBE through the experimentally determinedLET-dvalues agrees with independently acquired reference values, indicating that the investigated detectors can determineLET-with sufficient accuracy for proton therapy.Significance.OSLDs, TLDs, and FNTDs can be used to determineLET-and RBE in proton therapy. With the capability to determine dose through ionization quenching corrections derived fromLET-, OSLDs and TLDs can simultaneously ascertain dose,LET-, and RBE. This makes passive detectors appealing for measurements in phantoms to facilitate validation of clinical treatment plans or experiments related to proton therapy.

Dose averaged linear energy transfer optimization for large sacral chordomas in carbon ion therapy.

BACKGROUND: Carbon ion beams are well accepted as densely ionizing radiation with a high linear energy transfer (LET). However, the current clinical practice does not fully exploit the highest possible dose-averaged LET (LETd) and, consequently, the biological potential in the target. This aspect becomes worse in larger tumors for which inferior clinical outcomes and corresponding lower LETd was reported. PURPOSE: The vicinity to critical organs in general and the inferior overall survival reported for larger sacral chordomas treated with carbon ion radiotherapy (CIRT), makes the treatment of such tumors challenging. In this work it was aimed to increase the LETd in large volume tumors while maintaining the relative biological effectiveness (RBE)-weighted dose, utilizing the LETd optimization functions of a commercial treatment planning system (TPS). METHODS: Ten reference sequential boost carbon ion treatment plans, designed to mimic clinical plans for large sacral chordoma tumors, were generated. High dose clinical target volumes (CTV-HD) larger than 250 cm 3 $250 \,{\rm cm}^{3}$ were considered as large targets. The total RBE-weighted median dose prescription with the local effect model (LEM) was D RBE , 50 % = 73.6 Gy $\textrm {D}_{\rm RBE, 50\%}=73.6 \,{\rm Gy}$ in 16 fractions (nine to low dose and seven to high dose planning target volume). No LETd optimization was performed in the reference plans, while LETd optimized plans used the minimum LETd (Lmin) optimization function in RayStation 2023B. Three different Lmin values were investigated and specified for the seven boost fractions: L min = 60 keV / µ m $\textrm {L}_{\rm min}=60 \,{\rm keV}/{\umu }{\rm m}$ , L min = 80 keV / µ m $\textrm {L}_{\rm min}=80 \,{\rm keV}/{\umu }{\rm m}$ and L min = 100 keV / µ m $\textrm {L}_{\rm min}=100 \,{\rm keV}/{\umu }{\rm m}$ . To compare the LETd optimized against reference plans, LETd and RBE-weighted dose based goals similar to and less strict than clinical ones were specified for the target. The goals for the organs at risk (OAR) remained unchanged. Robustness evaluation was studied for eight scenarios ( ± 3.5 % $\pm 3.5\%$ range uncertainty and ± 3 mm $\pm 3 \,{\rm mm}$ setup uncertainty along the main three axes). RESULTS: The optimization method with L min = 60 keV / µ m $\textrm {L}_{\rm min}=60 \,{\rm keV}/{\umu }{\rm m}$ resulted in an optimal LETd distribution with an average increase of LET d , 98 % ${\rm {LET}}_{{\rm {d,}}98\%}$ (and LET d , 50 % ${\rm {LET}}_{{\rm {d,}}50\%}$ ) in the CTV-HD by 8.9 ± 1.5 keV / µ m $8.9\pm 1.5 \,{\rm keV}/{\umu }{\rm m}$ ( 27 % $27\%$ ) (and 6.9 ± 1.3 keV / µ m $6.9\pm 1.3 \,{\rm keV}/{\umu }{\rm m}$ ( 17 % $17\%$ )), without significant difference in the RBE-weighted dose. By allowing ± 5 % $\pm 5\%$ over- and under-dosage in the target, the LET d , 98 % ${\rm {LET}}_{{\rm {d,}}98\%}$ (and LET d , 50 % ${\rm {LET}}_{{\rm {d,}}50\%}$ ) can be increased by 11.3 ± 1.2 keV / µ m $11.3\pm 1.2 \,{\rm keV}/{\umu }{\rm m}$ ( 34 % $34\%$ ) (and 11.7 ± 3.4 keV / µ m $11.7\pm 3.4 \,{\rm keV}/{\umu }{\rm m}$ ( 29 % $29\%$ )), using the optimization parameters L min = 80 keV / µ m $\textrm {L}_{\rm min}=80 \,{\rm keV}/{\umu }{\rm m}$ . The pass rate for the OAR goals in the LETd optimized plans was in the same level as the reference plans. LETd optimization lead to less robust plans compared to reference plans. CONCLUSIONS: Compared to conventionally optimized treatment plans, the LETd in the target was increased while maintaining the RBE-weighted dose using TPS LETd optimization functionalities. Regularly assessing RBE-weighted dose robustness and acquiring more in-room images remain crucial and inevitable aspects during treatment.

Deep learning based linear energy transfer calculation for proton therapy.

Objective.This study aims to address the limitations of traditional methods for calculating linear energy transfer (LET), a critical component in assessing relative biological effectiveness (RBE). Currently, Monte Carlo (MC) simulation, the gold-standard for accuracy, is resource-intensive and slow for dose optimization, while the speedier analytical approximation has compromised accuracy. Our objective was to prototype a deep-learning-based model for calculating dose-averaged LET (LETd) using patient anatomy and dose-to-water (DW) data, facilitating real-time biological dose evaluation and LET optimization within proton treatment planning systems.Approach. 275 4-field prostate proton Stereotactic Body Radiotherapy plans were analyzed, rendering a total of 1100 fields. Those were randomly split into 880, 110, and 110 fields for training, validation, and testing. A 3D Cascaded UNet model, along with data processing and inference pipelines, was developed to generate patient-specific LETddistributions from CT images and DW. The accuracy of the LETdof the test dataset was evaluated against MC-generated ground truth through voxel-based mean absolute error (MAE) and gamma analysis.Main results.The proposed model accurately inferred LETddistributions for each proton field in the test dataset. A single-field LETdcalculation took around 100 ms with trained models running on a NVidia A100 GPU. The selected model yielded an average MAE of 0.94 ± 0.14 MeV cm-1and a gamma passing rate of 97.4% ± 1.3% when applied to the test dataset, with the largest discrepancy at the edge of fields where the dose gradient was the largest and counting statistics was the lowest.Significance.This study demonstrates that deep-learning-based models can efficiently calculate LETdwith high accuracy as a fast-forward approach. The model shows great potential to be utilized for optimizing the RBE of proton treatment plans. Future efforts will focus on enhancing the model's performance and evaluating its adaptability to different clinical scenarios.

Robustness of carbon-ion radiotherapy against DNA damage repair associated radiosensitivity variation based on a biophysical model.

BACKGROUND: Interpatient variation of tumor radiosensitivity is rarely considered during the treatment planning process despite its known significance for the therapeutic outcome. PURPOSE: To apply our mechanistic biophysical model to investigate the biological robustness of carbon ion radiotherapy (CIRT) against DNA damage repair interference (DDRi) associated patient-to-patient variability in radiosensitivity and its potential clinical advantages against conventional radiotherapy approaches. METHODS AND MATERIALS: The "UNIfied and VERSatile bio response Engine" (UNIVERSE) was extended by carbon ions and its predictions were compared to a panel of in vitro and in vivo data including various endpoints and DDRi settings within clinically relevant dose and linear energy transfer (LET) ranges. The implications of UNIVERSE predictions were then assessed in a clinical patient scenario considering DDRi variance. RESULTS: UNIVERSE tests well against the applied benchmarks. While in vitro survival curves were predicted with an R2 > 0.92, deviations from in vivo RBE data were less than 5.6% The conducted paradigmatic patient plan study implies a markedly reduced significance of DDRi based radiosensitivity variability in CIRT (13% change of D 50 ${{D}_{50}}$ in target) compared to conventional radiotherapy (62%) and that boosting the LET within the target further amplifies this robustness of CIRT (8%). In the case of heightened tumor radiosensitivity, a dose de-escalation strategy for photons allows a reduction of the maximum effective dose within the normal tissue (NT) from a D 2 ${{D}_2}$ of 2.65 to 1.64 Gy, which lies below the level found for CIRT ( D 2 ${{D}_2}$  = 2.41 Gy) for the analyzed plan and parameters. However, even after de-escalation, the integral effective dose in the NT is found to be substantially higher for conventional radiotherapy in comparison to CIRT ( D m e a n ${{D}_{mean}}$ of 0.75, 0.46, and 0.24 Gy for the conventional plan, its de-escalation and CIRT, respectively). CONCLUSIONS: The framework offers adequate predictions of in vitro and in vivo radiation effects of CIRT while allowing the consideration of DRRi based solely on parameters derived from photon data. The results of the patient planning study underline the potential of CIRT to minimize important sources of interpatient divergence in therapy outcome, especially when combined with techniques that allow to maximize the LET within the tumor. Despite the potential of de-escalation strategies for conventional radiotherapy to reduce the maximum effective dose in the NT, CIRT appears to remain a more favorable option due to its ability to reduce the integral effective dose within the NT.

Monte carlo simulation study on the dose and dose-averaged linear energy transfer distributions in carbon ion radiotherapy.

Dose-averaged linear energy transfer (LETd) is conventionally evaluated from the relative biological effectiveness (RBE)-LETd fitted function used in the treatment planning system. In this study, we calculated the physical doses and their linear energy transfer (LET) distributions for patterns of typical CIRT beams using Monte Carlo (MC) simulation. The LETd was then deduced from the MC simulation and compared with that obtained from the conventional method. The two types of LETd agreed well with each other, except around the distal end of the spread-out Bragg peak. Furthermore, an MC simulation was conducted with the material composition of water and realistic materials. The profiles of physical dose and LETd were in good agreement for both techniques. These results indicate that the previous studies to analyze the minimum LETd in CIRT cases are valid for practical situations, and the material composition conversion to water little affects the dose distribution in the irradiation field.