A joint physics and radiobiology DREAM team vision - Towards better response prediction models to advance radiotherapy.

Radiotherapy developed empirically through experience balancing tumour control and normal tissue toxicities. Early simple mathematical models formalized this practical knowledge and enabled effective cancer treatment to date. Remarkable advances in technology, computing, and experimental biology now create opportunities to incorporate this knowledge into enhanced computational models. The ESTRO DREAM (Dose Response, Experiment, Analysis, Modelling) workshop brought together experts across disciplines to pursue the vision of personalized radiotherapy for optimal outcomes through advanced modelling. The ultimate vision is leveraging quantitative models dynamically during therapy to ultimately achieve truly adaptive and biologically guided radiotherapy at the population as well as individual patient-based levels. This requires the generation of models that inform response-based adaptations, individually optimized delivery and enable biological monitoring to provide decision support to clinicians. The goal is expanding to models that can drive the realization of personalized therapy for optimal outcomes. This position paper provides their propositions that describe how innovations in biology, physics, mathematics, and data science including AI could inform models and improve predictions. It consolidates the DREAM team's consensus on scientific priorities and organizational requirements. Scientifically, it stresses the need for rigorous, multifaceted model development, comprehensive validation and clinical applicability and significance. Organizationally, it reinforces the prerequisites of interdisciplinary research and collaboration between physicians, medical physicists, radiobiologists, and computational scientists throughout model development. Solely by a shared understanding of clinical needs, biological mechanisms, and computational methods, more informed models can be created. Future research environment and support must facilitate this integrative method of operation across multiple disciplines.

Commissioning of pencil beam and Monte Carlo dose engines for non-isocentric treatments in scanned proton beam therapy.

This work describes the dosimetric commissioning of the treatment planning system (TPS) RayStation v6.1 from RaySearch Laboratories (Stockholm, Sweden) for a synchrotron-based scanned proton beam delivery with isocentric and non-isocentric setups at MedAustron. Focus was on the comparison of the pencil beam (PBv4.1) and Monte Carlo (MCv4.0) calculation algorithms. Commissioning of dose calculations was done first for 1D/2D dose delivery where the performance of the beam model in reproducing dosimetric properties for the delivery of single static pencil beams and mono-energetic layers with multiple spots was evaluated. The commissioning for 3D beam delivery employed test cases with increasing complexity: from box-shaped fields in homogeneous phantoms to the introduction of oblique incidences and inhomogeneities. Dose calculations were compared to the measured data for different air gaps and using beams with and without range shifter (RaShi). Depth-dose curves and spot shape comparisons showed good agreement of the results obtained with PBv4.1 and MCv4.0 algorithms at isocentric setup for open beam configurations (without RaShi). Comparison of transverse dose profiles for lateral heterogeneities at different depths showed better performance of the MCv4.0 algorithm in comparison to the PBv4.1 algorithm. In the case of 3D delivery comparisons of measured and TPS-calculated dose with MCv4.0 algorithm in box-shaped fields in water showed an average agreement within 2%. The results for dose calculations with the PBv4.1 algorithm showed larger deviations for beams with RaShi at all evaluated air gaps (from 64.8 cm to 14.8 cm). Our results suggest that the MCv4.0 algorithm shall be used in clinics for final dose calculation when beams with RaShi are used especially in the presence of large air gaps, inclined patient surface and lateral inhomogeneities. The detailed stepwise methodology implemented for the RayStation commissioning in this work could serve as further guidance for other facilities introducing a new TPS for proton beam therapy.

Dosimetric verification in water of a Monte Carlo treatment planning tool for proton, helium, carbon and oxygen ion beams at the Heidelberg Ion Beam Therapy Center.

The introduction of 'new' ion species in particle therapy needs to be supported by a thorough assessment of their dosimetric properties and by treatment planning comparisons with clinically used proton and carbon ion beams. In addition to the latter two ions, helium and oxygen ion beams are foreseen at the Heidelberg Ion Beam Therapy Center (HIT) as potential assets for improving clinical outcomes in the near future. We present in this study a dosimetric validation of a FLUKA-based Monte Carlo treatment planning tool (MCTP) for protons, helium, carbon and oxygen ions for spread-out Bragg peaks in water. The comparisons between the ions show the dosimetric advantages of helium and heavier ion beams in terms of their distal and lateral fall-offs with respect to protons, reducing the lateral size of the region receiving 50% of the planned dose up to 12 mm. However, carbon and oxygen ions showed significant doses beyond the target due to the higher fragmentation tail compared to lighter ions (p and He), up to 25%. The Monte Carlo predictions were found to be in excellent geometrical agreement with the measurements, with deviations below 1 mm for all parameters investigated such as target and lateral size as well as distal fall-offs. Measured and simulated absolute dose values agreed within about 2.5% on the overall dose distributions. The MCTP tool, which supports the usage of multiple state-of-the-art relative biological effectiveness models, will provide a solid engine for treatment planning comparisons at HIT.

Optimizing the modified microdosimetric kinetic model input parameters for proton and 4He ion beam therapy application.

Models able to predict relative biological effectiveness (RBE) values are necessary for an accurate determination of the biological effect with proton and 4He ion beams. This is particularly important when including RBE calculations in treatment planning studies comparing biologically optimized proton and 4He ion beam plans. In this work, we have tailored the predictions of the modified microdosimetric kinetic model (MKM), which is clinically applied for carbon ion beam therapy in Japan, to reproduce RBE with proton and 4He ion beams. We have tuned the input parameters of the MKM, i.e. the domain and nucleus radii, reproducing an experimental database of initial RBE data for proton and He ion beams. The modified MKM, with the best fit parameters obtained, has been used to reproduce in vitro cell survival data in clinically-relevant scenarios. A satisfactory agreement has been found for the studied cell lines, A549 and RENCA, with the mean absolute survival variation between the data and predictions within 2% and 5% for proton and 4He ion beams, respectively. Moreover, a sensitivity study has been performed varying the domain and nucleus radii and the quadratic parameter of the photon response curve. The promising agreement found in this work for the studied clinical-like scenarios supports the usage of the modified MKM for treatment planning studies in proton and 4He ion beam therapy.

A phenomenological relative biological effectiveness approach for proton therapy based on an improved description of the mixed radiation field.

Proton therapy treatment planning systems (TPSs) are based on the assumption of a constant relative biological effectiveness (RBE) of 1.1 without taking into account the found in vitro experimental variations of the RBE as a function of tissue type, linear energy transfer (LET) and dose. The phenomenological RBE models available in literature are based on the dose-averaged LET (LET D ) as an indicator of the physical properties of the proton radiation field. The LET D values are typically calculated taking into account primary and secondary protons, neglecting the biological effect of heavier secondaries. In this work, we have introduced a phenomenological RBE approach which considers the biological effect of primary protons, and of secondary protons, deuterons, tritons (Z = 1) and He fragments (3He and 4He, Z = 2). The calculation framework, coupled with a Monte Carlo (MC) code, has been successfully benchmarked against clonogenic in vitro data measured in this work for two cell lines and then applied to determine biological quantities for spread-out Bragg peaks and a prostate and a head case. The introduced RBE formalism, which depends on the mixed radiation field, the dose and the ratio of the linear-quadratic model parameters for the reference radiation [Formula: see text], predicts, when integrated in an MC code, higher RBE values in comparison to LET D -based parameterizations. This effect is particular enhanced in the entrance channel of the proton field and for low [Formula: see text] tissues. For the prostate and the head case, we found higher RBE-weighted dose values up to about 5% in the entrance channel when including or neglecting the Z = 2 secondaries in the RBE calculation. TPSs able to proper account for the mixed radiation field in proton therapy are thus recommended for an accurate determination of the RBE in the whole treatment field.

Biologically optimized helium ion plans: calculation approach and its in vitro validation.

Treatment planning studies on the biological effect of raster-scanned helium ion beams should be performed, together with their experimental verification, before their clinical application at the Heidelberg Ion Beam Therapy Center (HIT). For this purpose, we introduce a novel calculation approach based on integrating data-driven biological models in our Monte Carlo treatment planning (MCTP) tool. Dealing with a mixed radiation field, the biological effect of the primary (4)He ion beams, of the secondary (3)He and (4)He (Z = 2) fragments and of the produced protons, deuterons and tritons (Z = 1) has to be taken into account. A spread-out Bragg peak (SOBP) in water, representative of a clinically-relevant scenario, has been biologically optimized with the MCTP and then delivered at HIT. Predictions of cell survival and RBE for a tumor cell line, characterized by [Formula: see text] Gy, have been successfully compared against measured clonogenic survival data. The mean absolute survival variation ([Formula: see text]) between model predictions and experimental data was 5.3% ± 0.9%. A sensitivity study, i.e. quantifying the variation of the estimations for the studied plan as a function of the applied phenomenological modelling approach, has been performed. The feasibility of a simpler biological modelling based on dose-averaged LET (linear energy transfer) has been tested. Moreover, comparisons with biophysical models such as the local effect model (LEM) and the repair-misrepair-fixation (RMF) model were performed. [Formula: see text] values for the LEM and the RMF model were, respectively, 4.5% ± 0.8% and 5.8% ± 1.1%. The satisfactorily agreement found in this work for the studied SOBP, representative of clinically-relevant scenario, suggests that the introduced approach could be applied for an accurate estimation of the biological effect for helium ion radiotherapy.

Data-driven RBE parameterization for helium ion beams.

Helium ion beams are expected to be available again in the near future for clinical use. A suitable formalism to obtain relative biological effectiveness (RBE) values for treatment planning (TP) studies is needed. In this work we developed a data-driven RBE parameterization based on published in vitro experimental values. The RBE parameterization has been developed within the framework of the linear-quadratic (LQ) model as a function of the helium linear energy transfer (LET), dose and the tissue specific parameter (α/ß)ph of the LQ model for the reference radiation. Analytic expressions are provided, derived from the collected database, describing the RBEα = αHe/αph and Rß = ßHe/ßph ratios as a function of LET. Calculated RBE values at 2 Gy photon dose and at 10% survival (RBE10) are compared with the experimental ones. Pearson's correlation coefficients were, respectively, 0.85 and 0.84 confirming the soundness of the introduced approach. Moreover, due to the lack of experimental data at low LET, clonogenic experiments have been performed irradiating A549 cell line with (α/ß)ph = 5.4 Gy at the entrance of a 56.4 MeV u(-1)He beam at the Heidelberg Ion Beam Therapy Center. The proposed parameterization reproduces the measured cell survival within the experimental uncertainties. A RBE formula, which depends only on dose, LET and (α/ß)ph as input parameters is proposed, allowing a straightforward implementation in a TP system.

A Monte Carlo-based treatment-planning tool for ion beam therapy.

Ion beam therapy, as an emerging radiation therapy modality, requires continuous efforts to develop and improve tools for patient treatment planning (TP) and research applications. Dose and fluence computation algorithms using the Monte Carlo (MC) technique have served for decades as reference tools for accurate dose computations for radiotherapy. In this work, a novel MC-based treatment-planning (MCTP) tool for ion beam therapy using the pencil beam scanning technique is presented. It allows single-field and simultaneous multiple-fields optimization for realistic patient treatment conditions and for dosimetric quality assurance for irradiation conditions at state-of-the-art ion beam therapy facilities. It employs iterative procedures that allow for the optimization of absorbed dose and relative biological effectiveness (RBE)-weighted dose using radiobiological input tables generated by external RBE models. Using a re-implementation of the local effect model (LEM), the MCTP tool is able to perform TP studies using ions with atomic numbers Z ≤ 8. Example treatment plans created with the MCTP tool are presented for carbon ions in comparison with a certified analytical treatment-planning system. Furthermore, the usage of the tool to compute and optimize mixed-ion treatment plans, i.e. plans including pencil beams of ions with different atomic numbers, is demonstrated. The tool is aimed for future use in research applications and to support treatment planning at ion beam facilities.

Distributions of secondary particles in proton and carbon-ion therapy: a comparison between GATE/Geant4 and FLUKA Monte Carlo codes.

Monte Carlo simulations play a crucial role for in-vivo treatment monitoring based on PET and prompt gamma imaging in proton and carbon-ion therapies. The accuracy of the nuclear fragmentation models implemented in these codes might affect the quality of the treatment verification. In this paper, we investigate the nuclear models implemented in GATE/Geant4 and FLUKA by comparing the angular and energy distributions of secondary particles exiting a homogeneous target of PMMA. Comparison results were restricted to fragmentation of (16)O and (12)C. Despite the very simple target and set-up, substantial discrepancies were observed between the two codes. For instance, the number of high energy (>1 MeV) prompt gammas exiting the target was about twice as large with GATE/Geant4 than with FLUKA both for proton and carbon ion beams. Such differences were not observed for the predicted annihilation photon production yields, for which ratios of 1.09 and 1.20 were obtained between GATE and FLUKA for the proton beam and the carbon ion beam, respectively. For neutrons and protons, discrepancies from 14% (exiting protons-carbon ion beam) to 57% (exiting neutrons-proton beam) have been identified in production yields as well as in the energy spectra for neutrons.

A Monte Carlo-based treatment planning tool for proton therapy.

In the field of radiotherapy, Monte Carlo (MC) particle transport calculations are recognized for their superior accuracy in predicting dose and fluence distributions in patient geometries compared to analytical algorithms which are generally used for treatment planning due to their shorter execution times. In this work, a newly developed MC-based treatment planning (MCTP) tool for proton therapy is proposed to support treatment planning studies and research applications. It allows for single-field and simultaneous multiple-field optimization in realistic treatment scenarios and is based on the MC code FLUKA. Relative biological effectiveness (RBE)-weighted dose is optimized either with the common approach using a constant RBE of 1.1 or using a variable RBE according to radiobiological input tables. A validated reimplementation of the local effect model was used in this work to generate radiobiological input tables. Examples of treatment plans in water phantoms and in patient-CT geometries together with an experimental dosimetric validation of the plans are presented for clinical treatment parameters as used at the Italian National Center for Oncological Hadron Therapy. To conclude, a versatile MCTP tool for proton therapy was developed and validated for realistic patient treatment scenarios against dosimetric measurements and commercial analytical TP calculations. It is aimed to be used in future for research and to support treatment planning at state-of-the-art ion beam therapy facilities.

Investigating the robustness of ion beam therapy treatment plans to uncertainties in biological treatment parameters.

Uncertainties in determining clinically used relative biological effectiveness (RBE) values for ion beam therapy carry the risk of absolute and relative misestimations of RBE-weighted doses for clinical scenarios. This study assesses the consequences of hypothetical misestimations of input parameters to the RBE modelling for carbon ion treatment plans by a variational approach. The impact of the variations on resulting cell survival and RBE values is evaluated as a function of the remaining ion range. In addition, the sensitivity to misestimations in RBE modelling is compared for single fields and two opposed fields using differing optimization criteria. It is demonstrated for single treatment fields that moderate variations (up to ±50%) of representative nominal input parameters for four tumours result mainly in a misestimation of the RBE-weighted dose in the planning target volume (PTV) by a constant factor and only smaller RBE-weighted dose gradients. Ensuring a more uniform radiation quality in the PTV eases the clinical importance of uncertainties in the radiobiological treatment parameters, as for such a condition uncertainties tend to result only in a systematic misestimation of RBE-weighted dose in the PTV by a constant factor. Two opposed carbon ion fields with a constant RBE in the PTV are found to result in rather robust conditions. Treatments using two ion species may be used to achieve a constant RBE in the PTV irrespective of the size and depth of the spread-out Bragg peak.

FLUKA simulations of the response of tissue-equivalent proportional counters to ion beams for applications in hadron therapy and space.

For both cancer therapy with protons and ions (hadron therapy) and space radiation environments, the spatial energy deposition patterns of the radiation fields are of importance for quantifying the resulting radiation damage in biological structures. Tissue-equivalent proportional counters (TEPC) are the principal instruments for measuring imparted energy on a microscopic scale and for characterizing energy deposition patterns of radiation. Moreover, the distribution of imparted energy can serve as a complementary quantity to particle fluences of the primary beam and secondary fragments for characterizing a radiation field on a physical basis for radiobiological models. In this work, the Monte Carlo particle transport code FLUKA is used for simulating energy depositions in TEPC by ion beams. The capability of FLUKA in predicting imparted energy and derived quantities, such as lineal energy, for microscopic volumes is evaluated by comparing it with a large set of TEPC measurements for different ion beams with atomic numbers ranging from 1 to 26 and energies from 80 up to 1000 MeV/n. The influence of different physics configurations in the simulation is also discussed. It is demonstrated that FLUKA can simulate energy deposition patterns of ions in TEPC cavities accurately and that it provides an adequate description of the main features of the spectra.

Benchmarking nuclear models of FLUKA and GEANT4 for carbon ion therapy.

As carbon ions, at therapeutic energies, penetrate tissue, they undergo inelastic nuclear reactions and give rise to significant yields of secondary fragment fluences. Therefore, an accurate prediction of these fluences resulting from the primary carbon interactions is necessary in the patient's body in order to precisely simulate the spatial dose distribution and the resulting biological effect. In this paper, the performance of nuclear fragmentation models of the Monte Carlo transport codes, FLUKA and GEANT4, in tissue-like media and for an energy regime relevant for therapeutic carbon ions is investigated. The ability of these Monte Carlo codes to reproduce experimental data of charge-changing cross sections and integral and differential yields of secondary charged fragments is evaluated. For the fragment yields, the main focus is on the consideration of experimental approximations and uncertainties such as the energy measurement by time-of-flight. For GEANT4, the hadronic models G4BinaryLightIonReaction and G4QMD are benchmarked together with some recently enhanced de-excitation models. For non-differential quantities, discrepancies of some tens of percent are found for both codes. For differential quantities, even larger deviations are found. Implications of these findings for the therapeutic use of carbon ions are discussed.

Model-based control of mechanical ventilation: design and clinical validation.

BACKGROUND: We developed a model-based control system using end-tidal carbon dioxide fraction (FE'(CO(2))) to adjust a ventilator during clinical anaesthesia. METHODS: We studied 16 ASA I-II patients (mean age 38 (range 20-59) yr; weight 67 (54-87) kg) during i.v. anaesthesia for elective surgery. After periods of normal ventilation the patients were either hyper- or hypoventilated to assess precision and dynamic behaviour of the control system. These data were compared with a previous group where a fuzzy-logic controller had been used. Responses to different clinical events (invalid carbon dioxide measurement, limb tourniquet release, tube cuff leak, exhaustion of carbon dioxide absorbent, simulation of pulmonary embolism) were also noted. RESULTS: The model-based controller correctly maintained the setpoint. No significant difference was found for the static performance between the two controllers. The dynamic response of the model-based controller was more rapid (P<0.05). The mean rise time after a setpoint increase of 1 vol% was 313 (sd 90) s and 142 (17) s for fuzzy-logic and model-based control, respectively, and after a 1 vol% decrease was 355 (127) s and 177 (36) s, respectively. The new model-based controller had a consistent response to clinical artefacts. CONCLUSION: A model-based FE'(CO(2)) controller can be used in a clinical setting. It reacts appropriately to artefacts, and has a better dynamic response to setpoint changes than a previously described fuzzy-logic controller.