Assessment of the Sun Nuclear ArcCHECK to detect errors in 6MV FFF VMAT delivery of brain SABR using ROC analysis.

Institutions use a range of different detector systems for patient-specific quality assurance (QA) measurements conducted to assure that the dose delivered by a patient's radiotherapy treatment plan matches the calculated dose distribution. However, the ability of different detectors to detect errors from different sources is often unreported. This study contains a systematic evaluation of Sun Nuclear's ArcCHECK in terms of the detectability of potential machine-related treatment errors. The five investigated sources of error were multileaf collimator (MLC) leaf positions, gantry angle, collimator angle, jaw positions, and dose output. The study encompassed the clinical treatment plans of 29 brain cancer patients who received stereotactic ablative radiotherapy (SABR). Six error magnitudes were investigated per source of error. In addition, the Eclipse AAA beam model dosimetric leaf gap (DLG) parameter was varied with four error magnitudes. Error detectability was determined based on the area under the receiver operating characteristic (ROC) curve (AUC). Detectability of DLG errors was good or excellent (AUC >0.8) at an error magnitude of at least ±0.4 mm, while MLC leaf position and gantry angle errors reached good or excellent detectability at error magnitudes of at least 1.0 mm and 0.6°, respectively. Ideal thresholds, that is, gamma passing rates, to maximize sensitivity and specificity ranged from 79.1% to 98.7%. The detectability of collimator angle, jaw position, and dose output errors was poor for all investigated error magnitudes, with an AUC between 0.5 and 0.6. The ArcCHECK device's ability to detect errors from treatment machine-related sources was evaluated, and ideal gamma passing rate thresholds were determined for each source of error. The ArcCHECK was able to detect errors in DLG value, MLC leaf positions, and gantry angle. The ArcCHECK was unable to detect the studied errors in collimator angle, jaw positions, and dose output.

Radiation protection modelling for 2.5 Petawatt-laser production of ultrashort x-ray, proton and ion bunches: Monte Carlo model of the Munich CALA facility.

The 'Centre for Advanced Laser Applications' (CALA) is a new research institute for laser-based acceleration of electron beams for brilliant x-ray generation, laser-driven sub-nanosecond bunches of protons and heavy ions for biomedical applications like imaging and tumour therapy as well as for nuclear and high-field physics.The radiation sources emerging from experiments using the up to 2.5 petawatt laser pulses with 25 femtosecond duration will be mixed particle-species of high intensity, high energy and pulsed, thus posing new challenges compared to conventional radiation protection. Such worldwide pioneering laser experiments result in source characteristics that require careful a-priori radiation safety simulations.The FLUKA Monte-Carlo code was used to model the five CALA experimental caves, including the corridors, halls and air spaces surrounding the caves. Beams of electrons (<5 GeV), protons (<200 MeV),12C (<400MeV/u) and197Au (<10MeV/u) ions were simulated using spectra, divergences and bunch-charges based on expectations from recent scientific progress.Simulated dose rates locally can exceed 1.5 kSv h-1inside beam dumps. Vacuum pipes in the cave walls for laser transport and extraction channels for the generated x-rays result in small dose leakage to neighboring areas. Secondary neutrons contribute to most of the prompt dose rate outside caves into which the beam is delivered. This secondary radiation component causes non-negligible dose rates to occur behind walls to which large fluences of secondary particles are directed.By employing adequate beam dumps matched to beam-divergence, magnets, passive shielding and laser pulse repetition limits, average dose rates in- and outside the experimental building stay below design specifications (<0.5µSv h-1) for unclassified areas,<2.5µSv h-1for supervised areas,<7.5µSv h-1maximum local dose rate) and regulatory limits (<1mSv a-1for unclassified areas).

Evaluation of proton and photon dose distributions recalculated on 2D and 3D Unet-generated pseudoCTs from T1-weighted MR head scans.

Introduction: The recent developments of magnetic resonance (MR) based adaptive strategies for photon and, potentially for proton therapy, require a fast and reliable conversion of MR images to X-ray computed tomography (CT) values. CT values are needed for photon and proton dose calculation. The improvement of conversion results employing a 3D deep learning approach is evaluated. Material and methods: A database of 89 T1-weighted MR head scans with about 100 slices each, including rigidly registered CTs, was created. Twenty-eight validation patients were randomly sampled, and four patients were selected for application. The remaining patients were used to train a 2D and a 3D U-shaped convolutional neural network (Unet). A stack size of 32 slices was used for 3D training. For all application cases, volumetric modulated arc therapy photon and single-field uniform dose pencil-beam scanning proton plans at four different gantry angles were optimized for a generic target on the CT and recalculated on 2D and 3D Unet-based pseudoCTs. Mean (absolute) error (MAE/ME) and a gradient sharpness estimate were used to quantify the image quality. Three-dimensional gamma and dose difference analyses were performed for photon (gamma criteria: 1%, 1 mm) and proton dose distributions (gamma criteria: 2%, 2 mm). Range (80% fall off) differences for beam's eye view profiles were evaluated for protons. Results: Training 36 h for 1000 epochs in 3D (6 h for 200 epochs in 2D) yielded a maximum MAE of 147 HU (135 HU) for the application patients. Except for one patient gamma pass rates for photon and proton dose distributions were above 96% for both Unets. Slice discontinuities were reduced for 3D training at the cost of sharpness. Conclusions: Image analysis revealed a slight advantage of 2D Unets compared to 3D Unets. Similar dose calculation performance was reached for the 2D and 3D network.

Towards a novel small animal proton irradiation platform: the SIRMIO project.

Background: Precision small animal radiotherapy research is a young emerging field aiming to provide new experimental insights into tumor and normal tissue models in different microenvironments, to unravel complex mechanisms of radiation damage in target and non-target tissues and assess efficacy of novel therapeutic strategies. For photon therapy, modern small animal radiotherapy research platforms have been developed over the last years and are meanwhile commercially available. Conversely, for proton therapy, which holds potential for an even superior outcome than photon therapy, no commercial system exists yet. Material and methods: The project SIRMIO (Small Animal Proton Irradiator for Research in Molecular Image-guided Radiation-Oncology) aims at realizing and demonstrating an innovative portable prototype system for precision image-guided small animal proton irradiation, suitable for installation at existing clinical treatment facilities. The proposed design combines precise dose application with in situ multi-modal anatomical image guidance and in vivo verification of the actual treatment delivery. Results and conclusions: This manuscript describes the status of the different components under development, featuring a dedicated beamline for degradation and focusing of clinical proton beams, along with novel detector systems for in situimaging and range verification. The foreseen workflow includes pre-treatment proton transmission imaging, complemented by ultrasonic tumor localization, for treatment planning and position verification, followed by image-guided delivery with on site range verification by means of ionoacoustics (for pulsed beams) and positron-emission-tomography (PET, for continuous beams). The proposed compact and cost-effective system promises to open a new era in small animal proton therapy research, contributing to the basic understanding of in vivo radiation action to identify areas of potential breakthroughs for future translation into innovative clinical strategies.

Sensitivity of post treatment positron emission tomography/computed tomography to detect inter-fractional range variations in scanned ion beam therapy.

BACKGROUND: Ion therapy, especially with modern scanning beam delivery, offers very sharp dose gradients for highly conformal cancer treatment. However, it is very sensitive to uncertainties of tissue stopping properties as well as to anatomical changes and setup errors, making range verification highly desirable. To this end, positron emission tomography (PET) can be used to measure decay products of ß+-emitters created in interactions inside the patient. This work investigates the sensitivity of post treatment PET/CT (computed tomography) to detect inter-fractional range variations. MATERIAL AND METHODS: Fourteen patients of different indication underwent PET/CT monitoring after selected treatment fractions with scanned proton or carbon ion beams. In addition to PET/CT measurements, PET and dose distributions were simulated on different co-registered CT data. Pairs of PET data were then analyzed in terms of longitudinal shifts along the beam path, as surrogate of inter-fractional range deviations. These findings were compared to changes of dose-volume-histogram indexes and corresponding dose as well as CT shifts to disentangle the origin of possible PET shifts. RESULTS: Biological washout modeling (PET simulations) and low (<55 Bq/ml) activity concentrations (offline PET measurements, especially for 12C ions) were the main limitations for clinical treatment verification. For two selected cases, the benefit of improved washout modeling based on organ segmentation could be demonstrated. Overall, inter-fractional range shifts up to ±3 mm could be deduced from both PET measurements and simulations, and found well correlated (typically within 1.8 mm) to anatomical changes derived from CT scans, in agreement with dose data. CONCLUSIONS: Despite known limitations of post treatment PET/CT imaging, this work indicates its potential for assessing inter-fractional changes and points to future developments for improved PET-based treatment verification.

Practical implications for the quality assurance of modulated radiation therapy techniques using point detector arrays.

PURPOSE: Linac parameters potentially influencing the delivery quality of IMRT and VMAT plans are investigated with respect to threshold ranges, consequently to be considered in a linac based quality assurance procedure. Three commercially available 2D arrays are used to further investigate the influence of the measurement device. METHODS: Using three commercially available 2D arrays (Mx: MatriXXevolution , Oc: Octavius1500 , Mc: MapCHECK2), simple static measurements, measurements for MLC characterization and dynamic interplay of gantry movement, MLC movement and variable dose rate were performed. The results were evaluated with respect to each single array as well as among each other. RESULTS: Simple static measurements showed different array responses to dose, dose rate and profile homogeneity and revealed instabilities in dose delivery and profile shape during linac ramp up. Using the sweeping gap test, all arrays were able to detect small leaf misalignments down to ±0.1 mm, but this test also demonstrated up to 15% dose deviation due to profile instabilities and fast accelerating leaves during linac ramp up. Tests including gantry rotation showed different stability of gantry mounts for each array. Including gantry movement and dose rate variability, differences compared to static delivery were smaller compared to dose differences when simultaneously controling interplay of gantry movement, leaf movement and dose rate variability. CONCLUSION: Linac based QA is feasible with the tested commercially available 2D arrays. Limitations of each array and the linac ramp up characteristics should be carefully considered during individual plan generation and regularly checked in linac QA. Especially the dose and dose profile during linac ramp up should be checked regularly, as well as MLC positioning accuracy using a sweeping gap test. Additionally, dynamic interplay tests including various gantry rotation speeds and angles, various leaf speeds and various dose rates should be included.

Linearity of patient positioning detection : a phantom study of skin markers, cone beam computed tomography, and 3D ultrasound.

BACKGROUND: Three-dimensional ultrasound (3D-US) is a modality complementary to kilovoltage cone beam computed tomography (kV-CBCT) and skin markers for patient positioning detection. This study compares the linearity of evaluations based on measurements using a modern 3D-US system (Elekta Clarity®; Elekta, Stockholm, Sweden), a kV-CBCT system (Elekta iView®), and skin markers. MATERIALS AND METHODS: An investigator deliberately displaced a multimodal phantom by up to ± 30 mm along different axes. The following data points were acquired: 27 along the lateral axis, 29 along the longitudinal axis, 27 along the vertical axis, and 27 along the space diagonal. At each of these 110 positions, the displacements according to skin' markers were recorded and scans were performed using both 3D-US and kV-CBCT. Shifts were detected by matching bony anatomy or soft tissue density to a reference planning CT in the case of kV-CBCT and for 3D-US, by matching ultrasound volume data to a reference planning volume. A consensus value was calculated from the average of the four modalities. With respect to this consensus value, the linearity (offset and regression coefficient, i.e., slope), average offset, systematic error, and random error of all four modalities were calculated for each axis. RESULTS: Linearity was similar for all four modalities, with regression coefficients between 0.994 and 1.012, and all offsets below 1 mm. The systematic errors of skin markers and 3D-US were higher than for kV-CBCT, but random errors were similar. In particular, 3D-US demonstrated an average offset of 0.36 mm to the right, 0.08 mm inferiorly, and 0.15 mm anteriorly; the systematic error was 0.36 mm laterally, 0.35 mm longitudinally, and 0.22 mm vertically; the random error was 0.15 mm laterally, 0.30 mm longitudinally, and 0.12 mm vertically. A total of 109 out of 110 (99 %) 3D-US measurements were within 1 mm of the consensus value on either axis. CONCLUSION: The linearity of 3D-US was no worse than that of skin markers or kV-CBCT. Average offsets, systematic errors, and random errors were all below 1 mm. Optimal margins in the order of 1 mm could be achieved in the controlled laboratory setting of this phantom study.

Comparing cone-beam CT intensity correction methods for dose recalculation in adaptive intensity-modulated photon and proton therapy for head and neck cancer.

BACKGROUND: Adaptive intensity-modulated photon and proton radiotherapy (IMRT and IMPT) of head and neck (H&N) cancer requires frequent three-dimensional (3D) dose calculation. We compared two approaches for dose recalculation on the basis of intensity-corrected cone-beam (CB) x-ray computed tomography (CT) images. MATERIAL AND METHODS: For nine H&N tumor patients, virtual CTs (vCT) were generated by deformable image registration of the planning CT (pCT) to the CBCT. The second intensity correction approach used population-based lookup tables for scaling CBCT intensities to the pCT HU range (CBCTLUT). IMRT and IMPT plans were generated with a commercial treatment planning system. Dose recalculations on vCT and CBCTLUT were analyzed using a (3%, 3 mm) gamma-index analysis and comparison of normal tissue and tumor dose/volume parameters. A replanning CT (rpCT) acquired within three days of the CBCT served as reference. Single field uniform dose (SFUD) proton plans were created and recalculated on vCT and CBCTLUT for proton range comparison. RESULTS: Dose/volume parameters showed minor differences between rpCT, vCT and CBCTLUT in IMRT, but clinically relevant deviations between CBCTLUT and rpCT in the spinal cord for IMPT. Gamma-index pass-rates were found increased for vCT with respect to CBCTLUT in IMPT (by up to 21 percentage points) and IMRT (by up to 9 percentage points) for most cases. The SFUD-based proton range assessment showed improved agreement of vCT and rpCT, with 88-99% of the depth dose profiles in beam's eye view agreeing within 3 mm. For CBCTLUT, only 80-94% of the profiles fulfilled this criterion. CONCLUSION: vCT and CBCTLUT are suitable options for dose recalculation in adaptive IMRT. In the scope of IMPT, the vCT approach is preferable.

Evaluating an analytical prediction algorithm of positron emitter distributions in patient data for PET monitoring of carbon ion therapy: A simulation study.

In vivo treatment monitoring in ion therapy is one of the key issues for improving the treatment quality assurance procedures. Range verification is one of the most relevant and yet complex task used for in vivo treatment monitoring. In carbon ion therapy, positron emission tomography is the most widely used method. This technique exploits the ß+-activity of positron emitters created by nuclear interactions between the incoming beam and the irradiated tissue. Currently, high computational efforts and time-consuming Monte Carlo simulation platforms are typically used to predict positron emitter distributions. In order to avoid time-consuming simulations, an extended filtering approach was suggested to analytically predict positron emitter profiles from depth dose distributions in carbon ion therapy. The purpose of this work is to investigate such an analytical prediction model in patient anatomies of varying complexity, highlighting its potential and the need of further improvements, especially in highly heterogeneous anatomies where many air cavities are present in the beam path. The accuracy of range verification showed a mean relative error of ∼3% and a deviation between the simulation and the prediction below 2mm for the three patient cases analysed: a brain case and two head and neck cases. Additional investigations demonstrated the region of applicability of the method for cases of patient data. The analytical method enables range verification in carbon ion therapy by replacing computing-intensive Monte Carlo simulations and thus minimize the PET monitoring burden on the clinical workflow.

"Under the hood": artificial intelligence in personalized radiotherapy.

This review presents and discusses the ways in which artificial intelligence (AI) tools currently intervene, or could potentially intervene in the future, to enhance the diverse tasks involved in the radiotherapy workflow. The radiotherapy framework is presented on 2 different levels for the personalization of the treatment, distinct in tasks and methodologies. The first level is the clinically well-established anatomy-based workflow, known as adaptive radiation therapy. The second level is referred to as biology-driven workflow, explored in the research literature and recently appearing in some preliminary clinical trials for personalized radiation treatments. A 2-fold role for AI is defined according to these 2 different levels. In the anatomy-based workflow, the role of AI is to streamline and improve the tasks in terms of time and variability reductions compared to conventional methodologies. The biology-driven workflow instead fully relies on AI, which introduces decision-making tools opening uncharted frontiers that were in the past deemed challenging to explore. These methodologies are referred to as radiomics and dosiomics, handling imaging and dosimetric information, or multiomics, when complemented by clinical and biological parameters (ie, biomarkers). The review explicitly highlights the methodologies that are currently incorporated into clinical practice or still in research, with the aim of presenting the AI's growing role in personalized radiotherapy.

An experimental validation of a filtering approach for prompt gamma prediction in a research proton treatment planning system.

Objective.Prompt gamma (PG) radiation generated from nuclear reactions between protons and tissue nuclei can be employed for range verification in proton therapy. A typical clinical workflow for PG range verification compares the detected PG profile with a predicted one. Recently, a novel analytical PG prediction algorithm based on the so-called filtering formalism has been proposed and implemented in a research version of RayStation (RaySearch Laboratories AB), which is a widely adopted treatment planning system. This work validates the performance of the filtering PG prediction approach.Approach.The said algorithm is validated against experimental data and benchmarked with another well-established PG prediction algorithm implemented in a MATLAB-based software REGGUI. Furthermore, a new workflow based on several PG profile quality criteria and analytical methods is proposed for data selection. The workflow also calculates sensitivity and specificity information, which can help practitioners to decide on irradiation course interruption during treatment and monitor spot selection at the treatment planning stage. With the proposed workflow, the comparison can be performed on a limited number of selected high-quality irradiation spots without neighbouring-spot aggregation.Main results.The mean shifts between the experimental data and the predicted PG detection (PGD) profiles (ΔPGD) by the two algorithms are estimated to be1.5±2.1mm and-0.6±2.2mm for the filtering and REGGUI prediction methods, respectively. The ΔPGD difference between two algorithms is observed to be consistent with the beam model difference within uncertainty. However, the filtering approach requires a much shorter computation time compared to the REGGUI approach.Significance.The novel filtering approach is successfully validated against experimental data and another widely used PG prediction algorithm. The workflow designed in this work selects spots with high-quality PGD shift calculation results, and performs sensitivity and specificity analyses to assist clinical decisions.

Validation of dual-energy CT-based composition analysis using fresh animal tissues and composition-optimized tissue equivalent samples.

Objective.Proton therapy allows for highly conformal dose deposition, but is sensitive to range uncertainties. Several approaches currently under development measure composition-dependent secondary radiation to monitor the delivered proton rangein-vivo. To fully utilize these methods, an estimate of the elemental composition of the patient's tissue is often needed.Approach.A published dual-energy computed tomography (DECT)-based composition-extraction algorithm was validated against reference compositions obtained with two independent methods. For this purpose, a set of phantoms containing either fresh porcine tissue or tissue-mimicking samples with known, realistic compositions were imaged with a CT scanner at two different energies. Then, the prompt gamma-ray (PG) signal during proton irradiation was measured with a PG detector prototype. The PG workflow used pre-calculated Monte Carlo simulations to obtain an optimized estimate of the sample's carbon and oxygen contents. The compositions were also assessed with chemical combustion analysis (CCA), and the stopping-power ratio (SPR) was measured with a multi-layer ionization chamber. The DECT images were used to calculate SPR-, density- and elemental composition maps, and to assign voxel-wise compositions from a selection of human tissues. For a more comprehensive set of reference compositions, the original selection was extended by 135 additional tissues, corresponding to spongiosa, high-density bones and low-density tissues.Results.The root-mean-square error for the soft tissue carbon and oxygen content was 8.5 wt% and 9.5 wt% relative to the CCA result and 2.1 wt% and 10.3 wt% relative to the PG result. The phosphorous and calcium content were predicted within 0.4 wt% and 1.1 wt% of the CCA results, respectively. The largest discrepancies were encountered in samples whose composition deviated the most from tabulated compositions or that were more inhomogeneous.Significance.Overall, DECT-based composition estimations of relevant elements were in equal or better agreement with the ground truth than the established SECT-approach and could contribute toin-vivodose verification measurements.

A direct comparison of multi-energy x-ray and proton CT for imaging and relative stopping power estimation of plastic andex-vivophantoms.

Objective.Proton therapy administers a highly conformal dose to the tumour region, necessitating accurate prediction of the patient's 3D map of proton relative stopping power (RSP) compared to water. This remains challenging due to inaccuracies inherent in single-energy computed tomography (SECT) calibration. Recent advancements in spectral x-ray CT (xCT) and proton CT (pCT) have shown improved RSP estimation compared to traditional SECT methods. This study aims to provide the first comparison of the imaging and RSP estimation performance among dual-energy CT (DECT) and photon-counting CT (PCCT) scanners, and a pCT system prototype.Approach.Two phantoms were scanned with the three systems for their performance characterisation: a plastic phantom, filled with water and containing four plastic inserts and a wood insert, and a heterogeneous biological phantom, containing a formalin-stabilised bovine specimen. RSP maps were generated by converting CT numbers to RSP using a calibration based on low- and high-energy xCT images, while pCT utilised a distance-driven filtered back projection algorithm for RSP reconstruction. Spatial resolution, noise, and RSP accuracy were compared across the resulting images.Main results.All three systems exhibited similar spatial resolution of around 0.54 lp/mm for the plastic phantom. The PCCT images were less noisy than the DECT images at the same dose level. The lowest mean absolute percentage error (MAPE) of RSP,(0.28±0.07)%, was obtained with the pCT system, compared to MAPE values of(0.51±0.08)%and(0.80±0.08)%for the DECT- and PCCT-based methods, respectively. For the biological phantom, the xCT-based methods resulted in higher RSP values in most of the voxels compared to pCT.Significance.The pCT system yielded the most accurate estimation of RSP values for the plastic materials, and was thus used to benchmark the xCT calibration performance on the biological phantom. This study underlined the potential benefits and constraints of utilising such a novelex-vivophantom for inter-centre surveys in future.

A quantitative assessment of Geant4 for predicting the yield and distribution of positron-emitting fragments in ion beam therapy.

Objective.To compare the accuracy with which different hadronic inelastic physics models across ten Geant4 Monte Carlo simulation toolkit versions can predict positron-emitting fragments produced along the beam path during carbon and oxygen ion therapy.Approach.Phantoms of polyethylene, gelatin, or poly(methyl methacrylate) were irradiated with monoenergetic carbon and oxygen ion beams. Post-irradiation, 4D PET images were acquired and parent11C,10C and15O radionuclides contributions in each voxel were determined from the extracted time activity curves. Next, the experimental configurations were simulated in Geant4 Monte Carlo versions 10.0 to 11.1, with three different fragmentation models-binary ion cascade (BIC), quantum molecular dynamics (QMD) and the Liege intranuclear cascade (INCL++) - 30 model-version combinations. Total positron annihilation and parent isotope production yields predicted by each simulation were compared between simulations and experiments using normalised mean squared error and Pearson cross-correlation coefficient. Finally, we compared the depth of the maximum positron annihilation yield and the distal point at which the positron yield decreases to 50% of peak between each model and the experimental results.Main results.Performance varied considerably across versions and models, with no one version/model combination providing the best prediction of all positron-emitting fragments in all evaluated target materials and irradiation conditions. BIC in Geant4 10.2 provided the best overall agreement with experimental results in the largest number of test cases. QMD consistently provided the best estimates of both the depth of peak positron yield (10.4 and 10.6) and the distal 50%-of-peak point (10.2), while BIC also performed well and INCL generally performed the worst across most Geant4 versions.Significance.The best predictions of the spatial distribution of positron annihilations and positron-emitting fragment production along the beam path during carbon and oxygen ion therapy was obtained using Geant4 10.2.p03 with BIC or QMD. These version/model combinations are recommended for future heavy ion therapy research.

On the potential biological impact of radiation-induced acoustic emissions during ultra-high dose rate electron radiotherapy: a preliminary study.

Ionizing radiation pulses delivered at ultra-high dose rates in emerging FLASH radiotherapy can result in high-intensity low-frequency thermoacoustic emissions that may have a biological impact. This study aims at providing insights into the thermoacoustic emissions expected during FLASH radiotherapy and their likelihood of inducing acoustic cavitation. The characteristics of acoustic waves induced by the energy deposition of a pulsed electron beam similar to previous pre-clinical FLASH radiotherapy studies and their propagation in murine head-like phantoms are investigated in-silico. The results show that the generated pressures are sufficient to produce acoustic cavitation due to resonance in the irradiated object. It suggests that thermoacoustics may, in some irradiation scenarios, contribute to the widely misunderstood FLASH effect or cause adverse effects if not taken into account at the treatment planning stage.

Experience and new prospects of PET imaging for ion beam therapy monitoring.

Pioneering investigations on the usage of positron-emission-tomography (PET) for the monitoring of ion beam therapy with light (protons, helium) and heavier (stable and radioactive neon, carbon and oxygen) ions started shortly after the first realization of planar and tomographic imaging systems, which were able to visualize the annihilation of positrons resulting from irradiation induced or implanted positron emitting nuclei. And while the first clinical experience was challenged by the utilization of instrumentation directly adapted from nuclear medicine applications, new detectors optimized for this unconventional application of PET imaging are currently entering the phase of (pre)clinical testing for more reliable monitoring of treatment delivery during irradiation. Moreover, recent advances in detector technologies and beam production open several new exciting opportunities which will not only improve the performance of PET imaging under the challenging conditions of in-beam applications in ion beam therapy, but will also likely expand its field of application. In particular, the combination of PET and Compton imaging can enable the most efficient utilization of all possible radiative emissions for both stable and radioactive ion beams, while positronium lifetime imaging may enable probing new features of the underlying tumour and normal tissue environment. Thereby, PET imaging will not only provide means for volumetric reconstruction of the delivered treatment and in-vivo verification of the beam range, but can also shed new insights for biological optimization of the treatment or treatment response assessment.

Translation of AI into oncology clinical practice.

Artificial intelligence (AI) is a transformative technology that is capturing popular imagination and can revolutionize biomedicine. AI and machine learning (ML) algorithms have the potential to break through existing barriers in oncology research and practice such as automating workflow processes, personalizing care, and reducing healthcare disparities. Emerging applications of AI/ML in the literature include screening and early detection of cancer, disease diagnosis, response prediction, prognosis, and accelerated drug discovery. Despite this excitement, only few AI/ML models have been properly validated and fewer have become regulated products for routine clinical use. In this review, we highlight the main challenges impeding AI/ML clinical translation. We present different clinical use cases from the domains of radiology, radiation oncology, immunotherapy, and drug discovery in oncology. We dissect the unique challenges and opportunities associated with each of these cases. Finally, we summarize the general requirements for successful AI/ML implementation in the clinic, highlighting specific examples and points of emphasis including the importance of multidisciplinary collaboration of stakeholders, role of domain experts in AI augmentation, transparency of AI/ML models, and the establishment of a comprehensive quality assurance program to mitigate risks of training bias and data drifts, all culminating toward safer and more beneficial AI/ML applications in oncology labs and clinics.

ScatterNet for projection-based 4D cone-beam computed tomography intensity correction of lung cancer patients.

Background and purpose: In radiotherapy, dose calculations based on 4D cone beam CTs (4DCBCTs) require image intensity corrections. This retrospective study compared the dose calculation accuracy of a deep learning, projection-based scatter correction workflow (ScatterNet), to slower workflows: conventional 4D projection-based scatter correction (CBCTcor) and a deformable image registration (DIR)-based method (4DvCT). Materials and methods: For 26 lung cancer patients, planning CTs (pCTs), 4DCTs and CBCT projections were available. ScatterNet was trained with pairs of raw and corrected CBCT projections. Corrected projections from ScatterNet and the conventional workflow were reconstructed using MA-ROOSTER, yielding 4DCBCTSN and 4DCBCTcor. The 4DvCT was generated by 4DCT to 4DCBCT DIR, as part of the 4DCBCTcor workflow. Robust intensity modulated proton therapy treatment plans were created on free-breathing pCTs. 4DCBCTSN was compared to 4DCBCTcor and the 4DvCT in terms of image quality and dose calculation accuracy (dose-volume-histogram parameters and 3%/3mm gamma analysis). Results: 4DCBCTSN resulted in an average mean absolute error of 87HU and 102HU when compared to 4DCBCTcor and 4DvCT respectively. High agreement was observed in targets with median dose differences of 0.4Gy (4DCBCTSN-4DCBCTcor) and 0.3Gy (4DCBCTSN-4DvCT). The gamma analysis showed high average 3%/3mm pass rates of 96% for both 4DCBCTSN vs. 4DCBCTcor and 4DCBCTSN vs. 4DvCT. Conclusions: Accurate 4D dose calculations are feasible for lung cancer patients using ScatterNet for 4DCBCT correction. Average scatter correction times could be reduced from 10min (4DCBCTcor) to 3.9s, showing the clinical suitability of the proposed deep learning-based method.

Towards real-time EPID-based 3D in vivo dosimetry for IMRT with Deep Neural Networks: A feasibility study.

We investigate the potential of the Deep Dose Estimate (DDE) neural network to predict 3D dose distributions inside patients with Monte Carlo (MC) accuracy, based on transmitted EPID signals and patient CTs. The network was trained using as input patient CTs and first-order dose approximations (FOD). Accurate dose distributions (ADD) simulated with MC were given as training targets. 83 pelvic CTs were used to simulate ADDs and respective EPID signals for subfields of prostate IMRT plans (gantry at 0∘). FODs were produced as backprojections from the EPID signals. 581 ADD-FOD sets were produced and divided into training and test sets. An additional dataset simulated with gantry at 90∘ (lateral set) was used for evaluating the performance of the DDE at different beam directions. The quality of the FODs and DDE-predicted dose distributions (DDEP) with respect to ADDs, from the test and lateral sets, was evaluated with gamma analysis (3%,2 mm). The passing rates between FODs and ADDs were as low as 46%, while for DDEPs the passing rates were above 97% for the test set. Meaningful improvements were also observed for the lateral set. The high passing rates for DDEPs indicate that the DDE is able to convert FODs into ADDs. Moreover, the trained DDE predicts the dose inside a patient CT within 0.6 s/subfield (GPU), in contrast to 14 h needed for MC (CPU-cluster). 3D in vivo dose distributions due to clinical patient irradiation can be obtained within seconds, with MC-like accuracy, potentially paving the way towards real-time EPID-based in vivo dosimetry.

Design study of a novel geometrical arrangement for an in-beam small animal positron emission tomography scanner.

Objective.We designed a geometrical solution for a small animal in-beam positron emission tomography (PET) scanner to be used in the project SIRMIO (Small animal proton irradiator for research in molecular image-guided radiation-oncology). The system is based on 56 scintillator blocks of pixelated LYSO crystals. The crystals are arranged providing a pyramidal-step shape to optimize the geometrical coverage in a spherical configuration.Approach.Different arrangements have been simulated and compared in terms of spatial resolution and sensitivity. The chosen setup enables us to reach a good trade-off between a solid angle coverage and sufficient available space for the integration of additional components of the first design prototype of the SIRMIO platform. The possibility of moving the mouse holder inside the PET scanner furthermore allows for achieving the optimum placement of the irradiation area for all the possible tumor positions in the body of the mouse. The work also includes a study of the scintillator material where LYSO and GAGG are compared with a focus on the random coincidence noise due to the natural radioactivity of Lutetium in LYSO, justifying the choice of LYSO for the development of the final system.Main results.The best imaging performance can be achieved with a sub-millimeter spatial resolution and sensitivity of 10% in the center of the scanner, as verified in thorough simulations of point sources. The simulation of realistic irradiation scenarios of proton beams in PMMA targets with/without air gaps indicates the ability of the proposed PET system to detect range shifts down to 0.2 mm.Significance.The presented results support the choice of the identified optimal design for a novel spherical in-beam PET scanner which is currently under commissioning for application to small animal proton and light ion irradiation, and which might find also application, e.g. for biological image-guidance in x-ray irradiation.