Hypofractionated stereotactic radiotherapy (HFSRT) versus single fraction stereotactic radiosurgery (SRS) to the resection cavity of brain metastases after surgical resection (SATURNUS): study protocol for a randomized phase III trial.

BACKGROUND: The brain is a common site for cancer metastases. In case of large and/or symptomatic brain metastases, neurosurgical resection is performed. Adjuvant radiotherapy is a standard procedure to minimize the risk of local recurrence and is increasingly performed as local stereotactic radiotherapy to the resection cavity. Both hypofractionated stereotactic radiotherapy (HFSRT) and single fraction stereotactic radiosurgery (SRS) can be applied in this case. Although adjuvant stereotactic radiotherapy to the resection cavity is widely used in clinical routine and recommended in international guidelines, the optimal fractionation scheme still remains unclear. The SATURNUS trial prospectively compares adjuvant HFSRT with SRS and seeks to detect the superiority of HFSRT over SRS in terms of local tumor control. METHODS: In this single center two-armed randomized phase III trial, adjuvant radiotherapy to the resection cavity of brain metastases with HFSRT (6 - 7 × 5 Gy prescribed to the surrounding isodose) is compared to SRS (1 × 12-20 Gy prescribed to the surrounding isodose). Patients are randomized 1:1 into the two different treatment arms. The primary endpoint of the trial is local control at the resected site at 12 months. The trial is based on the hypothesis that HFSRT is superior to SRS in terms of local tumor control. DISCUSSION: Although adjuvant stereotactic radiotherapy after resection of brain metastases is considered standard of care treatment, there is a need for further prospective research to determine the optimal fractionation scheme. To the best of our knowledge, the SATURNUS study is the only randomized phase III study comparing different regimes of postoperative stereotactic radiotherapy to the resection cavity adequately powered to detect the superiority of HFSRT regarding local control. TRIAL REGISTRATION: The study was retrospectively registered with ClinicalTrials.gov, number NCT05160818, on December 16, 2021. The trial registry record is available on  https://clinicaltrials.gov/study/NCT05160818 . The presented protocol refers to version V1.3 from March 21, 2021.

X-ray Dark-Field CT for Early Detection of Radiation-induced Lung Injury in a Murine Model.

Online supplemental material is available for this article.

Early detection of radiation-induced lung damage with X-ray dark-field radiography in mice.

OBJECTIVE: Assessing the advantage of x-ray dark-field contrast over x-ray transmission contrast in radiography for the detection of developing radiation-induced lung damage in mice. METHODS: Two groups of female C57BL/6 mice (irradiated and control) were imaged obtaining both contrasts monthly for 28 weeks post irradiation. Six mice received 20 Gy of irradiation to the entire right lung sparing the left lung. The control group of six mice was not irradiated. A total of 88 radiographs of both contrasts were evaluated for both groups based on average values for two regions of interest, covering (irradiated) right lung and healthy left lung. The ratio of these average values, R, was distinguished between healthy and damaged lungs for both contrasts. The time-point when deviations of R from healthy lung exceeded 3σ was determined and compared among contrasts. The Wilcoxon-Mann-Whitney test was used to test against the null hypothesis that there is no difference between both groups. A selection of 32 radiographs was assessed by radiologists. Sensitivity and specificity were determined in order to compare the diagnostic potential of both contrasts. Inter-reader and intra-reader accuracy were rated with Cohen's kappa. RESULTS: Radiation-induced morphological changes of lung tissue caused deviations from the control group that were measured on average 10 weeks earlier with x-ray dark-field contrast than with x-ray transmission contrast. Sensitivity, specificity, and accuracy doubled using dark-field radiography. CONCLUSION: X-ray dark-field radiography detects morphological changes of lung tissue associated with radiation-induced damage earlier than transmission radiography in a pre-clinical mouse model. KEY POINTS: • Significant deviations from healthy lung due to irradiation were measured after 16 weeks with x-ray dark-field radiography (p = 0.004). • Significant deviations occur on average 10 weeks earlier for x-ray dark-field radiography in comparison to x-ray transmission radiography. • Sensitivity and specificity doubled when using x-ray dark-field radiography instead of x-ray transmission radiography.

Craniospinal irradiation(CSI) in patients with leptomeningeal metastases: risk-benefit-profile and development of a prognostic score for decision making in the palliative setting.

BACKGROUND: The aim of our study was to assess the feasibility and oncologic outcomes in patients treated with spinal (SI) or craniospinal irradiation (CSI) in patients with leptomeningeal metastases (LM) and to suggest a prognostic score as to which patients are most likely to benefit from this treatment. METHODS: Nineteen patients treated with CSI at our institution were eligible for the study. Demographic data, primary tumor characteristics, outcome and toxicity were assessed retrospectively. The extent of extra-CNS disease was defined by staging CT-scans before the initiation of CSI. Based on outcome parameters a prognostic score was developed for stratification based on patient performance status and tumor staging. RESULTS: Median follow-up and overall survival (OS) for the whole group was 3.4 months (range 0.5-61.5 months). The median overall survival (OS) for patients with LM from breast cancer was 4.7 months and from NSCLC 3.3 months. The median OS was 7.3 months, 3.3 months and 1.5 months for patients with 0, 1 and 2 risk factors according to the proposed prognostic score (KPS < 70 and the presence of extra-CNS disease) respectively. Nonhematologic toxicities were mild. CONCLUSION: CSI demonstrated clinically meaningful survival that is comparable to the reported outcome of intrathecal chemotherapy. A simple scoring system could be used to better select patients for treatment with CSI in this palliative setting. In our opinion, the feasibility of performing CSI with modern radiotherapy techniques with better sparing of healthy tissue gives a further rationale for its use also in the palliative setting.

A proof of principle experiment for microbeam radiation therapy at the Munich compact light source.

Microbeam radiation therapy (MRT), a preclinical form of spatially fractionated radiotherapy, uses an array of microbeams of hard synchrotron X-ray radiation. Recently, compact synchrotron X-ray sources got more attention as they provide essential prerequisites for the translation of MRT into clinics while overcoming the limited access to synchrotron facilities. At the Munich compact light source (MuCLS), one of these novel compact X-ray facilities, a proof of principle experiment was conducted applying MRT to a xenograft tumor mouse model. First, subcutaneous tumors derived from the established squamous carcinoma cell line FaDu were irradiated at a conventional X-ray tube using broadbeam geometry to determine a suitable dose range for the tumor growth delay. For irradiations at the MuCLS, FaDu tumors were irradiated with broadbeam and microbeam irradiation at integral doses of either 3 Gy or 5 Gy and tumor growth delay was measured. Microbeams had a width of 50 µm and a center-to-center distance of 350 µm with peak doses of either 21 Gy or 35 Gy. A dose rate of up to 5 Gy/min was delivered to the tumor. Both doses and modalities delayed the tumor growth compared to a sham-irradiated tumor. The irradiated area and microbeam pattern were verified by staining of the DNA double-strand break marker γH2AX. This study demonstrates for the first time that MRT can be successfully performed in vivo at compact inverse Compton sources.

The dosimetric impact of stabilizing spinal implants in radiotherapy treatment planning with protons and photons: standard titanium alloy vs. radiolucent carbon-fiber-reinforced PEEK systems.

BACKGROUND: Throughout the last years, carbon-fibre-reinforced PEEK (CFP) pedicle screw systems were introduced to replace standard titanium alloy (Ti) implants for spinal instrumentation, promising improved radiotherapy (RT) treatment planning accuracy. We compared the dosimetric impact of both implants for intensity modulated proton (IMPT) and volumetric arc photon therapy (VMAT), with the focus on uncertainties in Hounsfield unit assignment of titanium alloy. METHODS: Retrospective planning was performed on CT data of five patients with Ti and five with CFP implants. Carbon-fibre-reinforced PEEK systems comprised radiolucent pedicle screws with thin titanium-coated regions and titanium tulips. For each patient, one IMPT and one VMAT plan were generated with a nominal relative stopping power (SP) (IMPT) and electron density (ρ) (VMAT) and recalculated onto the identical CT with increased and decreased SP or ρ by ±6% for the titanium components. RESULTS: Recalculated VMAT dose distributions hardly deviated from the nominal plans for both screw types. IMPT plans resulted in more heterogeneous target coverage, measured by the standard deviation σ inside the target, which increased on average by 7.6 ± 2.3% (Ti) vs 3.4 ± 1.2% (CFP). Larger SPs lead to lower target minimum doses, lower SPs to higher dose maxima, with a more pronounced effect for Ti screws. CONCLUSIONS: While VMAT plans showed no relevant difference in dosimetric quality between both screw types, IMPT plans demonstrated the benefit of CFP screws through a smaller dosimetric impact of CT-value uncertainties compared to Ti. Reducing metal components in implants will therefore improve dose calculation accuracy and lower the risk for tumor underdosage.

Early and late toxicity profiles of patients receiving immediate postoperative radiotherapy versus salvage radiotherapy for prostate cancer after prostatectomy.

PURPOSE: The present study aims to evaluate both early and late toxicity profiles of patients receiving immediate postoperative radiotherapy (RT; adjuvant RT or additive RT) compared to salvage RT. METHODS: We evaluated 253 patients with prostate cancer treated with either immediate postoperative (adjuvant RT, n = 42; additive RT, n = 39) or salvage RT (n = 137). Thirty-five patients received salvage treatment but did not achieve a postoperative prostate specific antigen (PSA) level <0.1 ng/ml and thus were excluded from analysis. RESULTS: A significantly higher rate of early grade 1/2 proctitis in the immediate postoperative RT group without additional pelvic RT was observed (p = 0.02). Patients in the immediate postoperative RT group without additional pelvic RT showed significantly more early urinary tract obstructions (p = 0.003). Toxicity rates of early (<3 months) and late (3-6 months) postoperative RT were similar (p > 0.05). Baseline recovery rate of erectile dysfunction was better in patients with immediate postoperative RT without additional pelvic RT (p = 0.02; hazard ratio (HR) = 2.22, 95%-confidence interval, 95%-CI: 1.12-4.37). Recovery rate of urinary incontinence showed no significant difference in all groups (p > 0.05). CONCLUSION: Patients receiving immediate postoperative RT (adjuvant or additive RT) without additional pelvic RT experience early gastrointestinal (GI) side effect proctitis and, as well as early genitourinary (GU) toxicity urinary tract obstruction more frequently than patients treated with salvage RT. Therefore, complete recovery after surgery is essential. However, we suggest basing the treatment decision on the patient's postoperative clinical condition and evaluation of any adverse risk factors, since many studies demonstrate a clear benefit for immediate postoperative RT (adjuvant or additive RT) in terms of oncological outcome.

Individualized radiotherapy by combining high-end irradiation and magnetic resonance imaging.

Image-guided radiotherapy (IGRT) has been integrated into daily clinical routine and can today be considered the standard especially with high-dose radiotherapy. Currently imaging is based on MV- or kV-CT, which has clear limitations especially in soft-tissue contrast. Thus, combination of magnetic resonance (MR) imaging and high-end radiotherapy opens a new horizon. The intricate technical properties of MR imagers pose a challenge to technology when combined with radiation technology. Several solutions that are almost ready for routine clinical application have been developed. The clinical questions include dose-escalation strategies, monitoring of changes during treatment as well as imaging without additional radiation exposure during treatment.

Pencil beam kernel-based dose calculations on CT data for a mixed neutron-gamma fission field applying tissue correction factors.

Objective.For fast neutron therapy with mixed neutron and gamma radiation at the fission neutron therapy facility MEDAPP at the research reactor FRM II in Garching, no clinical dose calculation software was available in the past. Here, we present a customized solution for research purposes to overcome this lack of three-dimensional dose calculation.Approach.The applied dose calculation method is based on two sets of decomposed pencil beam kernels for neutron and gamma radiation. The decomposition was performed using measured output factors and simulated depth dose curves and beam profiles in water as reference medium. While measurements were performed by applying the two-chamber dosimetry method, simulated data was generated using the Monte Carlo code MCNP. For the calculation of neutron dose deposition on CT data, tissue-specific correction factors were generated for soft tissue, bone, and lung tissue for the MEDAPP neutron spectrum. The pencil beam calculations were evaluated with reference to Monte Carlo calculations regarding accuracy and time efficiency.Main results.In water, dose distributions calculated using the pencil beam approach reproduced the input from Monte Carlo simulations. For heterogeneous media, an assessment of the tissue-specific correction factors with reference to Monte Carlo simulations for different tissue configurations showed promising results. Especially for scenarios where no lung tissue is present, the dose calculation could be highly improved by the applied correction method.Significance.With the presented approach, time-efficient dose calculations on CT data and treatment plan evaluations for research purposes are now available for MEDAPP.

Reduced side effects by proton microchannel radiotherapy: study in a human skin model.

The application of a microchannel proton irradiation was compared to homogeneous irradiation in a three-dimensional human skin model. The goal is to minimize the risk of normal tissue damage by microchannel irradiation, while preserving local tumor control through a homogeneous irradiation of the tumor that is achieved because of beam widening with increasing track length. 20 MeV protons were administered to the skin models in 10- or 50-µm-wide irradiation channels on a quadratic raster with distances of 500 µm between each channel (center to center) applying an average dose of 2 Gy. For comparison, other samples were irradiated homogeneously at the same average dose. Normal tissue viability was significantly enhanced after microchannel proton irradiation compared to homogeneous irradiation. Levels of inflammatory parameters, such as Interleukin-6, TGF-Beta, and Pro-MMP1, were significantly lower in the supernatant of the human skin tissue after microchannel irradiation than after homogeneous irradiation. The genetic damage as determined by the measurement of micronuclei in keratinocytes also differed significantly. This difference was quantified via dose modification factors (DMF) describing the effect of each irradiation mode relative to homogeneous X-ray irradiation, so that the DMF of 1.21 ± 0.20 after homogeneous proton irradiation was reduced to 0.23 ± 0.11 and 0.40 ± 0.12 after microchannel irradiation using 10- and 50-µm-wide channels, respectively. Our data indicate that proton microchannel irradiation maintains cell viability while significantly reducing inflammatory responses and genetic damage compared to homogeneous irradiation, and thus might improve protection of normal tissue after irradiation.

Development of a PTV margin for preclinical irradiation of orthotopic pancreatic tumors derived from a well-known recipe for humans.

In human radiotherapy a safety margin (PTV margin) is essential for successful irradiation and is usually part of clinical treatment planning. In preclinical radiotherapy research with small animals, most uncertainties and inaccuracies are present as well, but according to the literature a margin is used only scarcely. In addition, there is only little experience about the appropriate size of the margin, which should carefully be investigated and considered, since sparing of organs at risk or normal tissue is affected. Here we estimate the needed margin for preclinical irradiation by adapting a well-known human margin recipe from van Herck et al. to the dimensions and requirements of the specimen on a small animal radiation research platform (SARRP). We adjusted the factors of the described formula to the specific challenges in an orthotopic pancreatic tumor mouse model to establish an appropriate margin concept. The SARRP was used with its image-guidance irradiation possibility for arc irradiation with a field size of 10 × 10 mm2 for 5 fractions. Our goal was to irradiate the clinical target volume (CTV) of at least 90% of our mice with at least 95% of the prescribed dose. By carefully analyzing all relevant factors we gain a CTV to planning target volume (PTV) margin of 1.5 mm for our preclinical setup. The stated safety margin is strongly dependent on the exact setting of the experiment and has to be adjusted for other experimental settings. The few stated values in literature correspond well to our result. Even if using margins in the preclinical setting might be an additional challenge, we think it is crucial to use them to produce reliable results and improve the efficacy of radiotherapy.

Roadmap for precision preclinical x-ray radiation studies.

This Roadmap paper covers the field of precision preclinical x-ray radiation studies in animal models. It is mostly focused on models for cancer and normal tissue response to radiation, but also discusses other disease models. The recent technological evolution in imaging, irradiation, dosimetry and monitoring that have empowered these kinds of studies is discussed, and many developments in the near future are outlined. Finally, clinical translation and reverse translation are discussed.

Comparison of 3 Positioning Techniques for Fractionated High-precision Radiotherapy in an Orthotopic Mouse Model of Pancreatic Cancer.

Small-animal irradiators are widely used in oncologic research, and many experiments use mice to mimic radiation treatments in humans. To improve fractionated high-precision irradiation in mice with orthotopic pancreatic tumors, we evaluated 3 positioning methods: no positioning aid, skin marker, and immobilization devices (immobilization masks). We retrospectively evaluated the translation vector needed for optimal tumor alignment (by shifting the mouse in left-right, in cranio-caudal, and in anterior-posterior direction) on cone-beam CT from our small-animal radiotherapy system. Of the 3 methods, the skin marker method yielded the smallest mean translation vector (3.8 mm) and was the most precise method overall for most of the mice. In addition, the skin marker method required supplemental rotation (that is, roll, pitch, and yaw) for optimal tumor alignment only half as often as positioning without a positioning aid. Finally, the skin marker method had the highest scores for the quality of the fusion results. Overall, we preferred the skin marker method over the other 2 positioning methods with regard to optimal treatment planning and radiotherapy in an orthotopic mouse model of pancreatic cancer.

A comprehensive and efficient quality assurance program for an image-guided small animal irradiation system.

In the field of preclinical radiotherapy, many new developments were driven by technical innovations. To make research of different groups comparable in that context and reliable, high quality has to be maintained. Therefore, standardized protocols and programs should be used. Here we present a guideline for a comprehensive and efficient quality assurance program for an image-guided small animal irradiation system, which is meant to test all the involved subsystems (imaging, treatment planning, and the irradiation system in terms of geometric accuracy and dosimetric aspects) as well as the complete procedure (end-to-end test) in a time efficient way. The suggestions are developed on a Small Animal Radiation Research Platform (SARRP) from Xstrahl (Xstrahl Ltd., Camberley, UK) and are presented together with proposed frequencies (from monthly to yearly) and experiences on the duration of each test. All output and energy related measurements showed stable results within small variation. Also, the motorized parts (couch, gantry) and other geometrical alignments were very stable. For the checks of the imaging system, the results are highly dependent on the chosen protocol and differ according to the settings. We received nevertheless stable and comparably good results for our mainly used protocol. All investigated aspects of treatment planning were exactly fulfilled and also the end-to-end test showed satisfying values. The mean overall time we needed for our checks to have a well monitored machine is less than two hours per month.

Heat management of a compact x-ray source for microbeam radiotherapy and FLASH treatments.

BACKGROUND: Microbeam and x-ray FLASH radiation therapy are innovative concepts that promise reduced normal tissue toxicity in radiation oncology without compromising tumor control. However, currently only large third-generation synchrotrons deliver acceptable x-ray beam qualities and there is a need for compact, hospital-based radiation sources to facilitate clinical translation of these novel treatment strategies. PURPOSE: We are currently setting up the first prototype of a line-focus x-ray tube (LFxT), a promising technology that may deliver ultra-high dose rates (UHDRs) of more than 100 Gy/s from a table-top source. The operation of the source in the heat capacity limit allows very high dose rates with micrometer-sized focal spot widths. Here, we investigate concepts of effective heat management for the LFxT, a prerequisite for the performance of the source. METHODS: For different focal spot widths, we investigated the temperature increase numerically with Monte Carlo simulations and finite element analysis (FEA). We benchmarked the temperature and thermal stresses at the focal spot against a commercial x-ray tube with similar power characteristics. We assessed thermal loads at the vacuum chamber housing caused by scattering electrons in Monte Carlo simulations and FEA. Further, we discuss active cooling strategies and present a design of the rotating target. RESULTS: Conventional focal spot widths led to a temperature increase dominated by heat conduction, while very narrow focal spots led to a temperature increase dominated by the heat capacity of the target material. Due to operation in the heat capacity limit, the temperature increase at the focal spot was lower than for the investigated commercial x-ray tube. Hence, the thermal stress at the focal spot of the LFxT was considered uncritical. The target shaft and the vacuum chamber housing require active cooling to withstand the high heat loads. CONCLUSIONS: The heat capacity limit allows very high power densities at the focal spot of the LFxT and thus facilitates very high dose rates. Numerical simulations demonstrated that the heat load imparted by scattering electrons requires active cooling.

Scanning irradiation device for mice in vivo with pulsed and continuous proton beams.

A technical set-up for irradiation of subcutaneous tumours in mice with nanosecond-pulsed proton beams or continuous proton beams is described and was successfully used in a first experiment to explore future potential of laser-driven particle beams, which are pulsed due to the acceleration process, for radiation therapy. The chosen concept uses a microbeam approach. By focusing the beam to approximately 100 × 100 µm(2), the necessary fluence of 10(9) protons per cm(2) to deliver a dose of 20 Gy with one-nanosecond shot in the Bragg peak of 23 MeV protons is achieved. Electrical and mechanical beam scanning combines rapid dose delivery with large scan ranges. Aluminium sheets one millimetre in front of the target are used as beam energy degrader, necessary for adjusting the depth-dose profile. The required procedures for treatment planning and dose verification are presented. In a first experiment, 24 tumours in mice were successfully irradiated with 23 MeV protons and a single dose of 20 Gy in pulsed or continuous mode with dose differences between both modes of 10%. So far, no significant difference in tumour growth delay was observed.

Measurements to predict the time of target replacement of a helical tomotherapy.

Intensity-modulated radiation therapy (IMRT) requires more beam-on time than normal open field treatment. Consequently, the machines wear out and need more spare parts. A helical tomotherapy treatment unit needs a periodical tungsten target replacement, which is a time consuming event. To be able to predict the next replacement would be quite valuable. We observed unexpected variations towards the end of the target lifetime in the performed pretreatment measurements for patient plan verification. Thus, we retrospectively analyze the measurements of our quality assurance program. The time dependence of the quotient of two simultaneous dose measurements at different depths within a phantom for a fixed open field irradiation is evaluated. We also assess the time-dependent changes of an IMRT plan measurement and of a relative depth dose curve measurement. Additionally, we performed a Monte Carlo simulation with Geant4 to understand the physical reasons for the measured values. Our measurements show that the dose at a specified depth compared to the dose in shallower regions of the phantom declines towards the end of the target lifetime. This reproducible effect can be due to the lowering of the mean energy of the X-ray spectrum. These results are supported by the measurements of the IMRT plan, as well as the study of the relative depth dose curve. Furthermore, the simulation is consistent with these findings since it provides a possible explanation for the reduction of the mean energy for thinner targets. It could be due to the lowering of low energy photon self-absorption in a worn out and therefore thinner target. We state a threshold value for our measurement at which a target replacement should be initiated. Measurements to observe a change in the energy are good predictors of the need for a target replacement. However, since all results support the softening of the spectrum hypothesis, all depth-dependent setups are viable for analyzing the deterioration of the tungsten target. The suggested measurements and criteria to replace the target can be very helpful for every user of a TomoTherapy machine.

Prediction of multi-criteria optimization (MCO) parameter efficiency in volumetric modulated arc therapy (VMAT) treatment planning using machine learning (ML).

PURPOSE: To predict the impact of optimization parameter changes on dosimetric plan quality criteria in multi-criteria optimized volumetric-modulated-arc therapy (VMAT) planning prior to optimization using machine learning (ML). METHODS: A data base comprising a total of 21,266 VMAT treatment plans for 44 cranial and 18 spinal patient geometries was generated. The underlying optimization algorithm is governed by three highly composite parameters which model a combination of important aspects of the solution. Patient geometries were parametrized via volume- and shape properties of the voxel objects and overlap-volume histograms (OVH) of the planning-target-volume (PTV) and a relevant organ-at-risk (OAR). The impact of changes in one of the three optimization parameters on the maximally achievable value range of five dosimetric properties of the resulting dose distributions was studied. To predict the extent of this impact based on patient geometry, treatment site, and current parameter settings prior to optimization, three different ML-models were trained and tested. Precision-recall curves, as well as the area-under-curve (AUC) of the resulting receiver-operator-characteristic (ROC) curves were analyzed for model assessment. RESULTS: Successful identification of parameter regions resulting in a high variability of dosimetric plan properties depended on the choice of geometry features, the treatment indication and the plan property under investigation. AUC values between 0.82 and 0.99 could be achieved. The best average-precision (AP) values obtained from the corresponding precision/recall curves ranged from 0.71 to 0.99. CONCLUSIONS: Machine learning models trained on a database of pre-optimized treatment plans can help finding relevant optimization parameter ranges prior to optimization.

Deep Learning Based HPV Status Prediction for Oropharyngeal Cancer Patients.

Infection with the human papillomavirus (HPV) has been identified as a major risk factor for oropharyngeal cancer (OPC). HPV-related OPCs have been shown to be more radiosensitive and to have a reduced risk for cancer related death. Hence, the histological determination of HPV status of cancer patients depicts an essential diagnostic factor. We investigated the ability of deep learning models for imaging based HPV status detection. To overcome the problem of small medical datasets, we used a transfer learning approach. A 3D convolutional network pre-trained on sports video clips was fine-tuned, such that full 3D information in the CT images could be exploited. The video pre-trained model was able to differentiate HPV-positive from HPV-negative cases, with an area under the receiver operating characteristic curve (AUC) of 0.81 for an external test set. In comparison to a 3D convolutional neural network (CNN) trained from scratch and a 2D architecture pre-trained on ImageNet, the video pre-trained model performed best. Deep learning models are capable of CT image-based HPV status determination. Video based pre-training has the ability to improve training for 3D medical data, but further studies are needed for verification.

Establishment of Microbeam Radiation Therapy at a Small-Animal Irradiator.

PURPOSE: Microbeam radiation therapy is a preclinical concept in radiation oncology. It spares normal tissue more effectively than conventional radiation therapy at equal tumor control. The radiation field consists of peak regions with doses of several hundred gray, whereas doses between the peaks (valleys) are below the tissue tolerance level. Widths and distances of the beams are in the submillimeter range for microbeam radiation therapy. A similar alternative concept with beam widths and distances in the millimeter range is presented by minibeam radiation therapy. Although both methods were developed at large synchrotron facilities, compact alternative sources have been proposed recently. METHODS AND MATERIALS: A small-animal irradiator was fitted with a special 3-layered collimator that is used for preclinical research and produces microbeams of flexible width of up to 100 µm. Film dosimetry provided measurements of the dose distributions and was compared with Monte Carlo dose predictions. Moreover, the micronucleus assay in Chinese hamster CHO-K1 cells was used as a biological dosimeter. The focal spot size and beam emission angle of the x-ray tube were modified to optimize peak dose rate, peak-to-valley dose ratio (PVDR), beam shape, and field homogeneity. An equivalent collimator with slit widths of up to 500 µm produced minibeams and allowed for comparison of microbeam and minibeam field characteristics. RESULTS: The setup achieved peak entrance dose rates of 8 Gy/min and PVDRs >30 for microbeams. Agreement between Monte Carlo simulations and film dosimetry is generally better for larger beam widths; qualitative measurements validated Monte Carlo predicted results. A smaller focal spot enhances PVDRs and reduces beam penumbras but substantially reduces the dose rate. A reduction of the beam emission angle improves the PVDR, beam penumbras, and dose rate without impairing field homogeneity. Minibeams showed similar field characteristics compared with microbeams at the same ratio of beam width and distance but had better agreement with simulations. CONCLUSION: The developed setup is already in use for in vitro experiments and soon for in vivo irradiations. Deviations between Monte Carlo simulations and film dosimetry are attributed to scattering at the collimator surface and manufacturing inaccuracies and are a matter of ongoing research.