Showing papers in "Medical Physics in 2022"

AAPM task group report 302: Surface‐guided radiotherapy

Abstract: Abstract The clinical use of surface imaging has increased dramatically, with demonstrated utility for initial patient positioning, real‐time motion monitoring, and beam gating in a variety of anatomical sites. The Therapy Physics Subcommittee and the Imaging for Treatment Verification Working Group of the American Association of Physicists in Medicine commissioned Task Group 302 to review the current clinical uses of surface imaging and emerging clinical applications. The specific charge of this task group was to provide technical guidelines for clinical indications of use for general positioning, breast deep‐inspiration breath hold treatment, and frameless stereotactic radiosurgery. Additionally, the task group was charged with providing commissioning and on‐going quality assurance (QA) requirements for surface‐guided radiation therapy (SGRT) as part of a comprehensive QA program including risk assessment. Workflow considerations for other anatomic sites and for computed tomography simulation, including motion management, are also discussed. Finally, developing clinical applications, such as stereotactic body radiotherapy (SBRT) or proton radiotherapy, are presented. The recommendations made in this report, which are summarized at the end of the report, are applicable to all video‐based SGRT systems available at the time of writing.

Ultra-high dose rate electron beams and the FLASH effect: from preclinical evidence to a new radiotherapy paradigm.

Abstract: In their seminal paper from 2014, Fauvadon et al. coined the term FLASH irradiation to describe ultra-high-dose-rate irradiation with dose rates greater than 40 Gy/s, which results in delivery times of fractions of a second. The experiments presented in that paper were performed with a high-dose-per-pulse 4.5-MeV electron beam, and the results served as the basis for the modern-day field of FLASH radiation therapy (RT). In this article, we review the studies that have been published after those early experiments, demonstrating the robust effects of FLASH RT on normal tissue sparing in preclinical models. We also outline the various irradiation parameters that have been used. Although the robustness of the biological response has been established, the mechanisms behind the FLASH effect are currently under investigation in a number of laboratories. However, differences in the magnitude of the FLASH effect between experiments in different labs have been reported. Reasons for these differences even within the same animal model are currently unknown, but likely has to do with the marked differences in irradiation parameter settings used. Here we show that these parameters are often not reported, which complicates large multi-study comparisons. For this reason, we propose a new standard for beam parameter reporting and discuss a systematic path to the clinical translation of FLASH radiation therapy. This article is protected by copyright. All rights reserved.

Ultra‐high dose rate dosimetry: Challenges and opportunities for FLASH radiation therapy

Abstract: Abstract The clinical translation of FLASH radiotherapy (RT) requires challenges related to dosimetry and beam monitoring of ultra‐high dose rate (UHDR) beams to be addressed. Detectors currently in use suffer from saturation effects under UHDR regimes, requiring the introduction of correction factors. There is significant interest from the scientific community to identify the most reliable solutions and suitable experimental approaches for UHDR dosimetry. This interest is manifested through the increasing number of national and international projects recently proposed concerning UHDR dosimetry. Attaining the desired solutions and approaches requires further optimization of already established technologies as well as the investigation of novel radiation detection and dosimetry methods. New knowledge will also emerge to fill the gap in terms of validated protocols, assessing new dosimetric procedures and standardized methods. In this paper, we discuss the main challenges coming from the peculiar beam parameters characterizing UHDR beams for FLASH RT. These challenges vary considerably depending on the accelerator type and technique used to produce the relevant UHDR radiation environment. We also introduce some general considerations on how the different time structure in the production of the radiation beams, as well as the dose and dose‐rate per pulse, can affect the detector response. Finally, we discuss the requirements that must characterize any proposed dosimeters for use in UDHR radiation environments. A detailed status of the current technology is provided, with the aim of discussing the detector features and their performance characteristics and/or limitations in UHDR regimes. We report on further developments for established detectors and novel approaches currently under investigation with a view to predict future directions in terms of dosimetry approaches, practical procedures, and protocols. Due to several on‐going detector and dosimetry developments associated with UHDR radiation environment for FLASH RT it is not possible to provide a simple list of recommendations for the most suitable detectors for FLASH RT dosimetry. However, this article does provide the reader with a detailed description of the most up‐to‐date dosimetric approaches, and describes the behavior of the detectors operated under UHDR irradiation conditions and offers expert discussion on the current challenges which we believe are important and still need to be addressed in the clinical translation of FLASH RT.

Abdomen CT Multi-organ Segmentation Using Token-based MLP-Mixer.

Abstract: BACKGROUND Manual contouring is very labor-intensive, time-consuming, and subject to intra- and inter-observer variability. An automated deep learning approach to fast and accurate contouring and segmentation is desirable during radiotherapy treatment planning. PURPOSE This work investigates an efficient deep-learning-based segmentation algorithm in abdomen computed tomography (CT) to facilitate radiation treatment planning. METHODS In this work, we propose a novel deep-learning model utilizing U-shaped Multi-Layer Perceptron Mixer (MLP-Mixer) and convolutional neural network (CNN) for multi-organ segmentation in abdomen CT images. The proposed model has a similar structure to V-net, while a proposed MLP-Convolutional block replaces each convolutional block. The MLP-Convolutional block consists of three components: an early convolutional block for local features extraction and feature resampling, a token-based MLP-Mixer layer for capturing global features with high efficiency, and a token projector for pixel-level detail recovery. We evaluate our proposed network using: 1) an institutional dataset with 60 patient cases, and 2) a public dataset (BCTV) with 30 patient cases. The network performance was quantitatively evaluated in three domains: 1) volume similarity between the ground truth contours and the network predictions using the Dice score coefficient (DSC), sensitivity, and precision; 2) surface similarity using Hausdorff distance (HD), mean surface distance (MSD) and residual mean square distance (RMS); 3) the computational complexity reported by the number of network parameters, training time, and inference time. The performance of the proposed network is compared with other state-of-the-art networks. RESULTS In the institutional dataset, the proposed network achieved the following volume similarity measures when averaged over all organs: DSC = 0.912, sensitivity = 0.917, precision = 0.917, average surface similarities were HD = 11.95mm, MSD = 1.90mm, RMS = 3.86mm. The proposed network achieved DSC = 0.786 and HD = 9.04mm on the public dataset. The network also shows statistically significant improvement, which is evaluated by a two-tailed Wilcoxon Mann-Whitney U test, on right lung (MSD where the maximum p-value is 0.001), spinal cord (sensitivity, precision, HD, RMSD where p-value ranges from 0.001 to 0.039), and stomach (DSC where the maximum p-value is 0.01) over all other competing networks. On the public dataset, the network report statistically significant improvement, which is shown by the Wilcoxon Mann-Whitney test, on pancreas (HD where the maximum p-value is 0.006), left (HD where the maximum p-value is 0.022) and right adrenal glands (DSC where the maximum p-value is 0.026). In both datasets, the proposed method can generate contours in less than five seconds. Overall, the proposed MLP-Vnet demonstrates comparable or better performance than competing methods with much lower memory complexity and higher speed. CONCLUSIONS The proposed MLP-Vnet demonstrates superior segmentation performance, in terms of accuracy and efficiency, relative to state-of-the-art methods. This reliable and efficient method demonstrates potential to streamline clinical workflows in abdominal radiotherapy, which may be especially important for online adaptive treatments. This article is protected by copyright. All rights reserved.

Design, realization, and characterization of a novel diamond detector prototype for FLASH radiotherapy dosimetry

Abstract: Abstract Purpose FLASH radiotherapy (RT) is an emerging technique in which beams with ultra‐high dose rates (UH‐DR) and dose per pulse (UH‐DPP) are used. Commercially available active real‐time dosimeters have been shown to be unsuitable in such conditions, due to severe response nonlinearities. In the present study, a novel diamond‐based Schottky diode detector was specifically designed and realized to match the stringent requirements of FLASH‐RT. Methods A systematic investigation of the main features affecting the diamond response in UH‐DPP conditions was carried out. Several diamond Schottky diode detector prototypes with different layouts were produced at Rome Tor Vergata University in cooperation with PTW‐Freiburg. Such devices were tested under electron UH‐DPP beams. The linearity of the prototypes was investigated up to DPPs of about 26 Gy/pulse and dose rates of approximately 1 kGy/s. In addition, percentage depth dose (PDD) measurements were performed in different irradiation conditions. Radiochromic films were used for reference dosimetry. Results The response linearity of the diamond prototypes was shown to be strongly affected by the size of their active volume as well as by their series resistance. By properly tuning the design layout, the detector response was found to be linear up to at least 20 Gy/pulse, well into the UH‐DPP range conditions. PDD measurements were performed by three different linac applicators, characterized by DPP values at the point of maximum dose of 3.5, 17.2, and 20.6 Gy/pulse, respectively. The very good superimposition of three curves confirmed the diamond response linearity. It is worth mentioning that UH‐DPP irradiation conditions may lead to instantaneous detector currents as high as several mA, thus possibly exceeding the electrometer specifications. This issue was properly addressed in the case of the PTW UNIDOS electrometers. Conclusions The results of the present study clearly demonstrate the feasibility of a diamond detector for FLASH‐RT applications.

A deep learning system for automated kidney stone detection and volumetric segmentation on non-contrast CT scans.

Abstract: PURPOSE Early detection and size quantification of renal calculi are important for optimizing treatment and preventing severe kidney stone disease. Prior work has shown that volumetric measurements of kidney stones are more informative and reproducible than linear measurements. Deep learning based systems that use abdominal non-contrast CT scans may assist in detection and reduce workload by removing the need for manual stone volume measurement. Prior to this work no such system had been developed for use on noisy low-dose CT or tested on a large-scale external dataset. METHODS We used a dataset of 91 CT colonography (CTC) scans with manually marked kidney stones combined with 89 CTC scans without kidney stones. To compare with a prior work half the data was used for training and half for testing. A set of CTC scans from 6,185 patients from a separate institution with patient-level labels were used as an external validation set. A 3D U-Net model was employed to segment the kidneys, followed by gradient-based anisotropic denoising, thresholding, and region growing. A 13 layer convolutional neural network classifier was then applied to distinguish kidney stones from false positive regions. RESULTS The system achieved a sensitivity of 0.86 at 0.5 false positives per scan on a challenging test set of low-dose CT with many small stones, an improvement over an earlier work which obtained a sensitivity of 0.52. The stone volume measurements correlated well with manual measurements (r2 = 0.95). For patient level classification the system achieved an area under the receiver operating characteristic (AU-ROC) of 0.95 on an external validation set (sensitivity = 0.88, specificity = 0.91 at the Youden point). A common cause of false positives were small atherosclerotic plaques in the renal sinus that simulated kidney stones. CONCLUSIONS Our deep learning based system showed improvements over a previously developed system that did not use deep learning, with even higher performance on an external validation set. This article is protected by copyright. All rights reserved.

Differentiation between immune checkpoint inhibitor‐related and radiation pneumonitis in lung cancer by CT radiomics and machine learning

Abstract: Abstract Purpose Consolidation immunotherapy after completion of chemoradiotherapy has become the standard of care for unresectable locally advanced non‐small cell lung cancer and can induce potentially severe and life‐threatening adverse events, including both immune checkpoint inhibitor‐related pneumonitis (CIP) and radiation pneumonitis (RP), which are very challenging for radiologists to diagnose. Differentiating between CIP and RP has significant implications for clinical management such as the treatments for pneumonitis and the decision to continue or restart immunotherapy. The purpose of this study is to differentiate between CIP and RP by a CT radiomics approach. Methods We retrospectively collected the CT images and clinical information of patients with pneumonitis who received immune checkpoint inhibitor (ICI) only (n = 28), radiotherapy (RT) only (n = 31), and ICI+RT (n = 14). Three kinds of radiomic features (intensity histogram, gray‐level co‐occurrence matrix [GLCM] based, and bag‐of‐words [BoW] features) were extracted from CT images, which characterize tissue texture at different scales. Classification models, including logistic regression, random forest, and linear SVM, were first developed and tested in patients who received ICI or RT only with 10‐fold cross‐validation and further tested in patients who received ICI+RT using clinicians’ diagnosis as a reference. Results Using 10‐fold cross‐validation, the classification models built on the intensity histogram features, GLCM‐based features, and BoW features achieved an area under curve (AUC) of 0.765, 0.848, and 0.937, respectively. The best model was then applied to the patients receiving combination treatment, achieving an AUC of 0.896. Conclusions This study demonstrates the promising potential of radiomic analysis of CT images for differentiating between CIP and RP in lung cancer, which could be a useful tool to attribute the cause of pneumonitis in patients who receive both ICI and RT.

A quantitative FLASH effectiveness model to reveal potentials and pitfalls of high dose rate proton therapy

Abstract: Abstract Purpose In ultrahigh dose rate radiotherapy, the FLASH effect can lead to substantially reduced healthy tissue damage without affecting tumor control. Although many studies show promising results, the underlying biological mechanisms and the relevant delivery parameters are still largely unknown. It is unclear, particularly for scanned proton therapy, how treatment plans could be optimized to maximally exploit this protective FLASH effect. Materials and Methods To investigate the potential of pencil beam scanned proton therapy for FLASH treatments, we present a phenomenological model, which is purely based on experimentally observed phenomena such as potential dose rate and dose thresholds, and which estimates the biologically effective dose during FLASH radiotherapy based on several parameters. We applied this model to a wide variety of patient geometries and proton treatment planning scenarios, including transmission and Bragg peak plans as well as single‐ and multifield plans. Moreover, we performed a sensitivity analysis to estimate the importance of each model parameter. Results Our results showed an increased plan‐specific FLASH effect for transmission compared with Bragg peak plans (19.7% vs. 4.0%) and for single‐field compared with multifield plans (14.7% vs. 3.7%), typically at the cost of increased integral dose compared to the clinical reference plan. Similar FLASH magnitudes were found across the different treatment sites, whereas the clinical benefits with respect to the clinical reference plan varied strongly. The sensitivity analysis revealed that the threshold dose as well as the dose per fraction strongly impacted the FLASH effect, whereas the persistence time only marginally affected FLASH. An intermediate dependence of the FLASH effect on the dose rate threshold was found. Conclusions Our model provided a quantitative measure of the FLASH effect for various delivery and patient scenarios, supporting previous assumptions about potentially promising planning approaches for FLASH proton therapy. Positive clinical benefits compared to clinical plans were achieved using hypofractionated, single‐field transmission plans. The dose threshold was found to be an important factor, which may require more investigation.

Development of an ultra‐thin parallel plate ionization chamber for dosimetry in FLASH radiotherapy

Abstract: Abstract Background Conventional air ionization chambers (ICs) exhibit ion recombination correction factors that deviate substantially from unity when irradiated with dose per pulse magnitudes higher than those used in conventional radiotherapy. This fact makes these devices unsuitable for the dosimetric characterization of beams in ultra‐high dose per pulse as used for FLASH radiotherapy. Purpose We present the design, development, and characterization of an ultra‐thin parallel plate IC that can be used in ultra‐high dose rate (UHDR) deliveries with minimal recombination. Methods The charge collection efficiency (CCE) of parallel plate ICs was modeled through a numerical solution of the coupled differential equations governing the transport of charged carriers produced by ionizing radiation. It was used to find out the optimal parameters for the purpose of designing an IC capable of exhibiting a linear response with dose (deviation less than 1%) up to 10 Gy per pulse at 4 μs pulse duration. As a proof of concept, two vented parallel plate IC prototypes have been built and tested in different ultra‐high pulse dose rate electron beams. Results It has been found that by reducing the distance between electrodes to a value of 0.25 mm it is possible to extend the dose rate operating range of parallel plate ICs to ultra‐high dose per pulse range, at standard voltage of clinical grade electrometers, well into several Gy per pulse. The two IC prototypes exhibit behavior as predicted by the numerical simulation. One of the so‐called ultra‐thin parallel plate ionization chamber (UTIC) prototypes was able to measure up to 10 Gy per pulse, 4 μs pulse duration, operated at 300 V with no significant deviation from linearity within the uncertainties (ElectronFlash Linac, SIT). The other prototype was tested up to 5.4 Gy per pulse, 2.5 μs pulse duration, operated at 250 V with CCE higher than 98.6% (Metrological Electron Accelerator Facility, MELAF at Physikalisch‐Technische Bundesanstalt, PTB). Conclusions This work demonstrates the ability to extend the dose rate operating range of ICs to ultra‐high dose per pulse range by reducing the spacing between electrodes. The results show that UTICs are suitable for measurement in UHDR electron beams.

A roadmap to clinical trials for FLASH

Abstract: Abstract While FLASH radiation therapy is inspiring enthusiasm to transform the field, it is neither new nor well understood with respect to the radiobiological mechanisms. As FLASH clinical trials are designed, it will be important to ensure we can deliver dose consistently and safely to every patient. Much like hyperthermia and proton therapy, FLASH is a promising new technology that will be complex to implement in the clinic and similarly will require customized credentialing for multi‐institutional clinical trials. There is no doubt that FLASH seems promising, but many technologies that we take for granted in conventional radiation oncology, such as rigorous dosimetry, 3D treatment planning, volumetric image guidance, or motion management, may play a major role in defining how to use, or whether to use, FLASH radiotherapy. Given the extended time frame for patients to experience late effects, we recommend moving deliberately but cautiously forward toward clinical trials. In this paper, we review the state of quality assurance and safety systems in FLASH, identify critical pre‐clinical data points that need to be defined, and suggest how lessons learned from previous technological advancements will help us close the gaps and build a successful path to evidence‐driven FLASH implementation.

Methods of monitoring thermal ablation of soft tissue tumors – A comprehensive review

Abstract: Thermal ablation is a form of hyperthermia in which oncologic control can be achieved by briefly inducing elevated temperatures, typically in the range 50-80°C, within a target tissue. Ablation modalities include high intensity focused ultrasound, radiofrequency ablation, microwave ablation, and laser interstitial thermal therapy which are all capable of generating confined zones of tissue destruction, resulting in fewer complications than conventional cancer therapies. Oncologic control is contingent upon achieving predefined coagulation zones; therefore, intraoperative assessment of treatment progress is highly desirable. Consequently, there is a growing interest in the development of ablation monitoring modalities. The first section of this review presents the mechanism of action and common applications of the primary ablation modalities. The following section outlines the state-of-the-art in thermal dosimetry which includes interstitial thermal probes and radiologic imaging. Both the physical mechanism of measurement and clinical or pre-clinical performance are discussed for each ablation modality. Thermal dosimetry must be coupled with a thermal damage model as outlined in Section 4. These models estimate cell death based on temperature-time history and are inherently tissue specific. In the absence of a reliable thermal model, the utility of thermal monitoring is greatly reduced. The final section of this review paper covers technologies that have been developed to directly assess tissue conditions. These approaches include visualization of non-perfused tissue with contrast-enhanced imaging, assessment of tissue mechanical properties using ultrasound and magnetic resonance elastography, and finally interrogation of tissue optical properties with interstitial probes. In summary, monitoring thermal ablation is critical for consistent clinical success and many promising technologies are under development but an optimal solution has yet to achieve widespread adoption.

AAPM Task Group Report 290: Respiratory motion management for particle therapy

Abstract: Abstract Dose uncertainty induced by respiratory motion remains a major concern for treating thoracic and abdominal lesions using particle beams. This Task Group report reviews the impact of tumor motion and dosimetric considerations in particle radiotherapy, current motion‐management techniques, and limitations for different particle‐beam delivery modes (i.e., passive scattering, uniform scanning, and pencil‐beam scanning). Furthermore, the report provides guidance and risk analysis for quality assurance of the motion‐management procedures to ensure consistency and accuracy, and discusses future development and emerging motion‐management strategies. This report supplements previously published AAPM report TG76, and considers aspects of motion management that are crucial to the accurate and safe delivery of particle‐beam therapy. To that end, this report produces general recommendations for commissioning and facility‐specific dosimetric characterization, motion assessment, treatment planning, active and passive motion‐management techniques, image guidance and related decision‐making, monitoring throughout therapy, and recommendations for vendors. Key among these recommendations are that: (1) facilities should perform thorough planning studies (using retrospective data) and develop standard operating procedures that address all aspects of therapy for any treatment site involving respiratory motion; (2) a risk‐based methodology should be adopted for quality management and ongoing process improvement.

Evaluating the clinical acceptability of deep learning contours of prostate and organs-at-risk in an automated prostate treatment planning process.

Abstract: BACKGROUND Radiation treatment is considered an effective and the most common treatment option for prostate cancer. The treatment planning process requires accurate and precise segmentation of the prostate and organs at risk (OARs), which is laborious and time-consuming when contoured manually. Artificial intelligence (AI)-based auto-segmentation has the potential to significantly accelerate the radiation therapy treatment planning process; however, the accuracy of auto-segmentation needs to be validated before its full clinical adoption. PURPOSE A commercial AI-based contouring model was trained to provide segmentation of the prostate and surrounding OARs. The segmented structures were input to a commercial auto-planning module for automated prostate treatment planning. This study comprehensively evaluates the performance of this contouring model in the automated prostate treatment planning process. METHODS AND MATERIALS A 3D U-Net-based model (INTContour, Carina AI) was trained and validated on 84 computed tomography (CT) scans and tested on an additional 23 CT scans from patients treated in our local institution. Prostate and OARs contours generated by the AI model (AI contour) were geometrically evaluated against Reference contours. The prostate contours were further evaluated against AI, Reference, and two additional observer contours for comparison using inter-observer variation (IOV) and 3D boundaries discrepancy analyses. A blinded evaluation was introduced to assess subjectively the clinical acceptability of the AI contours. Finally, treatment plans were created from an automated prostate planning workflow using the AI contours and were evaluated for their clinical acceptability following the RTOG-0815 protocol. RESULTS The AI contours demonstrated good geometric accuracy on OARs and prostate contours, with average Dice similarity coefficients (DSC) for bladder, rectum, femoral heads, seminal vesicles, and penile bulb of 0.93, 0.85, 0.96, 0.72, and 0.53, respectively. The DSC, 95% directed Hausdorff Distance (HD95), and Mean Surface Distance (MSD) for the prostate were 0.83±0.05, 6.07±1.87 mm, and 2.07±0.73 mm, respectively. No significant differences were found when comparing with IOV. In the double-blinded evaluation, 95.7% of the AI contours were scored as either "Perfect" (34.8%) or "Acceptable" (60.9%), while only one case (4.3%) was scored as "Unacceptable with minor changes required". In total, 69.6% of the AI contours were considered equal to or better than the Reference contours by an independent radiation oncologist. Automated treatment plans created from the AI contours produced similar and clinically-acceptable dosimetric distributions as those from plans created from Reference contours. CONCLUSIONS The investigated AI-based commercial model for prostate segmentation demonstrated good performance in clinical practice. Using this model, the implementation of an automated prostate treatment planning process is clinically feasible. This article is protected by copyright. All rights reserved.

The impact of phantom design and material-dependence on repeatability and reproducibility of CT-based radiomics features.

Abstract: PURPOSE To understand the design of radiomics phantom and material-dependence on repeatability and reproducibility of CT radiomics features METHODS: : A radiomics phantom consisting of various materials with uniformity, textural and biological components, was constructed. The phantom was scanned with different manufacturer CT scanners and the scans were repeated multiple times on the same scanner with different acquisition settings as kVp, mAs, orientation, field of view (FOV), slice thickness, pitch, reconstruction kernels and acquisition mode. A total of 72 phantom scans were included. For each scan, 18 different regions of interest (ROI) were contoured and 708 radiomics features were extracted from each ROI via an open source radiomics tool, IBEX. To relate the phantom data to patient data, the radiomics features from different phantom materials were compared with those extracted from 50 patients' images of five disease sites as brain, head-and-neck, breast, liver and lung cases using box-plots comparison and principal component analysis (PCA). The temporal stability of imaging features was then evaluated with respect to a controlled scenario (test-retest) via the intra-class correlation coefficient (ICC). The reproducibility of radiomics features with respect to different scanners or acquisition settings were further evaluated with concordance correlation coefficients (CCC). RESULTS Among all phantom materials, the biological component had feature values closest to human tissues, especially for tumors in brain and liver. The textural component showed similar ranges of variation to lung lesions, particularly for cartridges of rice, cereal, and the 3D-printed textural phantom with fine and rough-grid. It also showed that certain materials, such as polystyrene foam, plaster and peanuts, did not have comparable values to human tissue and could be excluded for future phantom design. High repeatability was observed in the test-retest study as indicated by an ICC value of 0.998 ± 0.020. All materials were used for feature stability analysis. For the inter-scanner study, shape-related features were the most-reliable category with 94% of features having CCC ≥0.9, while GOH were the least-reliable with only 14.6% meeting the criteria. For the intra-scanner study, the reproducibility of CT-based radiomics features showed material-dependence. In general, the instability of radiomics features introduced by kVp, mAs, pitch, acquisition mode and orientation were relatively mild. However, the homogeneous materials were more vulnerable to those changes compared to materials with textural patterns. Regardless of material compositions, resolution parameters like FOV and slice thickness, could have large impact on feature stability. Switching between standard and bone reconstruction kernels could also result significant changes to feature reproducibility. CONCLUSION We have built a radiomics phantom using materials that cover a wide span of tumor textures seen in oncological CT images. The designed phantom presents a preliminary opportunity for investigating reproducibility of radiomics features and the reproducibility can be material dependent. Thus, in the radiomics quality assurance design, it is important to choose appropriate materials that can provide a close range of radiomics features to patients with specific disease sites dependency taken into consideration. This article is protected by copyright. All rights reserved.

Proton beam dosimetry at ultra-high dose rates (FLASH): evaluation of GAFchromic™ (EBT3, EBT-XD) and OrthoChromic (OC-1) film performances.

Abstract: PURPOSE The ARRONAX cyclotron facility offers the possibility to deliver proton beams from low to ultra-high dose rates (UHDR). As a good control of the dosimetry is a prerequisite of UHDR experimentations, we evaluated in different conditions the usability and the dose rate dependency of several radiochromic films commonly used for dosimetry in radiotherapy. METHODS We compared the dose rate dependency of three types of radiochromic films: EBT3 and EBT-XD (GAFchromic™), and OC-1 (OrthoChrome Inc.), after proton irradiations at various mean dose rates (0.25, 40, 1500 and 7500 Gy/s) and for 10 doses (2-130 Gy). We also evaluated the dose rate dependency of each film considering beam structures, from single pulse to multiple pulses with various frequencies. RESULTS EBT3 and EBT-XD films showed differences of response between conventional (0.25 Gy/s) and UHDR (7500 Gy/s) conditions, above 10 Gy. On the contrary, OC-1 films did not present overall difference of response for doses except below 3 Gy. We observed an increase of the netOD with the mean dose rate for EBT3 and EBT-XD films. OC-1 films did not show any impact of the mean dose rate up to 7500 Gy/s, above 3 Gy. No difference was found based on the beam structure, for all three types of films. CONCLUSIONS EBT3 and EBT-XD radiochromic films should be used with caution for the dosimetry of UHDR proton beams over 10 Gy. Their overresponse, which increases with mean dose rate and dose, could lead to non-negligible overestimations of the absolute dose. OC-1 films are dose rate independent up to 7500Gy/s in proton beams. Films response is not impacted by the beam structure. A broader investigation of the usability of OC-1 films in UHDR conditions should be conducted at intermediate and higher mean dose rates and other beam energies. This article is protected by copyright. All rights reserved.

Impact of an artificial intelligence deep‐learning reconstruction algorithm for CT on image quality and potential dose reduction: A phantom study

Abstract: Abstract Background Recently, computed tomography (CT) manufacturers have developed deep‐learning‐based reconstruction algorithms to compensate for the limitations of iterative reconstruction (IR) algorithms, such as image smoothing and the spatial resolution's dependence on contrast and dose levels. Purpose To assess the impact of an artificial intelligence deep‐learning reconstruction (AI‐DLR) algorithm on image quality and dose reduction compared with a hybrid IR algorithm in chest CT for different clinical indications. Methods Acquisitions on the CT American College of Radiology (ACR) 464 and CT Torso CTU‐41 phantoms were performed at five dose levels (CTDIvol: 9.5/7.5/6/2.5/0.4 mGy) used for chest CT conditions. Raw data were reconstructed using filtered backprojection, two levels of IR (iDose4 levels 4 (i4) and 7 (i7)), and five levels of AI‐DLR (Precise Image; Smoother, Smooth, Standard, Sharp, Sharper). Noise power spectrum (NPS), task‐based transfer function, and detectability index (d′) were computed: d′‐modeled detection of a soft tissue mediastinal nodule (low‐contrast soft tissue chest nodule within the mediastinum [LCN]), ground‐glass opacity (GGO), or high‐contrast pulmonary (HCP) lesion. The subjective image quality of chest anthropomorphic phantom images was independently evaluated by two radiologists. They assessed image noise, image smoothing, contrast between vessels and fat in the mediastinum for mediastinal images, visual border detection between bronchus and lung parenchyma for parenchymal images, and overall image quality using a commonly used four‐ or five‐point scale. Results From Standard to Smoother levels, on average, the noise magnitude decreased (for all dose levels: −66.3% ± 0.5% for mediastinal images and −63.1% ± 0.1% for parenchymal images), the average NPS spatial frequency decreased (for all dose levels: −35.3% ± 2.2% for mediastinal images and −13.3% ± 2.2% for parenchymal images), and the detectability (d′) of the three lesions increased. The opposite pattern was found from Standard to Sharper levels. From Smoother to Sharper levels, the spatial resolution increased for the low‐contrast polyethylene insert and the opposite for the high‐contrast air insert. Compared to the i4 used in clinical practice, d′ values were higher using Smoother (mean for all dose levels: 338.7% ± 29.4%), Smooth (103.4% ± 11.2%), and Standard (34.1% ± 6.6%) levels for the LCN on mediastinal images and Smoother (169.5% ± 53.2% for GGO and 136.9% ± 1.6% for HCP) and Smooth (36.4% ± 22.1% and 24.1% ± 0.9%, respectively) levels for parenchymal images. Radiologists considered the images satisfactory for clinical use at these levels, but adaptation to the dose level of the protocol is required. Conclusion With AI‐DLR, the smoothest levels reduced the noise and improved the detectability of chest lesions but increased the image smoothing. The opposite was found with the sharpest levels. The choice of level depends on the dose level and type of image: mediastinal or parenchymal.

General and custom deep learning auto-segmentation models for organs in head and neck, abdomen, and male pelvis.

Abstract: PURPOSE To reduce workload and inconsistencies in organ segmentation for radiation treatment planning, we developed and evaluated general and custom auto-segmentation models on CT for three major tumor sites using a well-established deep convolutional neural network (DCNN). METHODS AND MATERIALS Five CT-based auto-segmentation models for 42 organs at risk (OARs) in head and neck (HN), abdomen (ABD) and male pelvis (MP) were developed using a full 3D DCNN architecture. Two types of DL models were separately trained using either general diversified multi-institutional datasets or custom well controlled single institution datasets. To improve segmentation accuracy, an adaptive spatial resolution approach for small and/or narrow OARs and a pseudo scan extension approach, when CT scan length is too short to cover entire organs, were implemented. The performance of the obtained models was evaluated based on accuracy and clinical applicability of the auto-segmented contours using qualitative visual inspection and quantitative calculation of dice similarity coefficient (DSC), mean distance to agreement (MDA), and time efficiency. RESULTS The five DL auto-segmentation models developed for the three anatomical sites were found to have high accuracy (DSC ranging from 0.8 - 0.98) for 74% OARs and marginally acceptable for 26% OARs. The custom models performed slightly better than the general models, even with smaller custom datasets used for the custom model training. The organ-based approaches improved auto-segmentation accuracy for small or complex organs (e.g., eye lens, optic nerves, inner ears, and bowels). Compared with traditional manual contouring times, the auto-segmentation times, including subsequent manual editing, if necessary, were substantially reduced by 88% for MP, 80% for HN, and 65% for ABD models. CONCLUSIONS The obtained auto-segmentation models, incorporating organ-based approaches were found to be effective and accurate for most OARs in the male pelvis, head and neck and abdomen. We have demonstrated that our multi-anatomical deep learning auto-segmentation models are clinically useful for radiation treatment planning. This article is protected by copyright. All rights reserved.

Time structure of pencil beam scanning proton FLASH beams measured with scintillator detectors and compared with log files.

Abstract: PURPOSE Key factors in FLASH treatments are the ultra-high dose rate (UHDR) and the time structure of the beam delivery. Measurements of the time structure in pencil beam scanning (PBS) proton FLASH treatments is challenging for many types of detectors since high temporal resolution is needed. In this study, a fast scintillator detector system was developed and used to measure the individual spot durations as well as the time when the beam moves between two positions (transition duration) during PBS proton FLASH and UHDR treatments. The spot durations were compared with machine log-file recordings. METHODS A detector system based on inorganic scintillating crystals was developed. The system consisted of four detector probes made of a sub-millimeter ZnSe:O crystal that was coupled via an optical fiber to an optical reader with 50kHz sampling rate. The detector system was used in two experiments, both performed with a PBS proton beam with 250MeV beam energy and 215nA requested nozzle beam current. The sampling rate enabled multiple measurements during each spot delivery and during the beam transition between spots. First, the detector was tested in a phantom experiment, where a total of 305 scan sequences were delivered to the four detectors. The number of spots delivered without beam interruption in a single scan sequence ranged from one to 35. The spot duration and transition duration were measured for each individual spot. Secondly, the detector system was used in vivo in pre-clinical experiments with FLASH irradiation of mouse legs placed in the entrance plateau of the beam. A single detector was placed 1cm downstream of the irradiated mouse leg. The mouse dose ranged from 30.5Gy to 44.2Gy and the field consisted of 35 spots. The spot durations as well as the mean dose rate (field dose divided by the measured field duration) for each mouse were determined using the detector and then compared with the corresponding log-files. RESULTS The phantom experiment showed that the logged total duration of an uninterrupted spot sequence was consistently shorter than the measured duration with a difference of -0.252ms (95%-confidence interval: [-0.255ms, -0.249ms]). This corresponded to 0.05-0.07% of the spot sequence duration in the mice experiments. For individual spots, the mean±1SD difference between logged and measured spot duration was -0.39±0.05ms for the first spot in a sequence, 0.13±0.04ms for the last spot in a sequence and -0.0017±0.09ms for the intermediate spots in a sequence. The measured spot transition durations were 0.20±0.04ms (5.1mm horizontal steps) and 0.50±0.04ms (5.0mm vertical steps). For the mouse experiments, the mean dose rate calculated from the measured field duration was 84.1-92.5Gy/s. It agreed with log files with a root-mean-square difference of 0.02Gy/s. CONCLUSIONS Fiber-coupled scintillator detectors were designed with sufficient temporal resolution to measure the spot and transition duration during PBS proton UHDR deliveries. Their small volume makes them feasible for in vivo use in pre-clinical FLASH studies. The logged spot durations were in excellent agreement with measurements but showed small systematic errors in the logged duration for the first and last spot in a sequence. This article is protected by copyright. All rights reserved.

Automatic segmentation of MR images for high-dose-rate cervical cancer brachytherapy using deep learning.

Abstract: PURPOSE Magnetic resonance (MR) imaging is the gold standard in image-guided brachytherapy (IGBT) due to its superior soft-tissue contrast for target and organ-at-risk (OARs) delineation. Accurate and fast segmentation of MR images are very important for high-quality IGBT treatment planning. The purpose of this work is to implement and evaluate deep learning models for the automatic segmentation of targets and OARs in MR image-based high-dose-rate (HDR) brachytherapy for cervical cancer. METHODS A 2D deep learning (DL) model using residual neural network architecture (ResNet50) was developed to contour the targets (GTV, HR CTV, and IR CTV) and OARs (bladder, rectum, sigmoid, and small intestine) automatically on axial MR slices of HDR brachytherapy patients. Furthermore, two additional 2D DL models using sagittal and coronal images were also developed. A 2.5D model was generated by combining the outputs from axial, sagittal, and coronal DL models. Similarly, a 2D and 2.5D DL models were also generated for the inception residual neural network (InceptionResNetv2 (InRN)) architecture. The geometric (dice similarity coefficient (DSCs), 95th percentile of Hausdorff distance (HD)) and dosimetric accuracy of 2D (axial only) and 2.5D (axial + sagittal + coronal) DL model generated contours were calculated and compared. RESULTS The mean (range) DSCs of ResNet50 across all contours were 0.674 (0.05-0.96) and 0.715 (0.26-0.96) for the 2D and 2.5D models, respectively. For InRN, these were 0.676 (0.11-0.96) and 0.723 (0.35-0.97) for the 2D and 2.5D models, respectively. The mean HD of ResNet50 across all contours was 15.6 mm (1.8-69 mm) and 12.1 mm (1.7-44 mm) for the 2D and 2.5D models, respectively. The similar results for InRN were 15.4 mm (2-68 mm) and 10.3 mm (2.7-39 mm) for the 2D and 2.5D models, respectively. The dosimetric parameters (D90) of GTV and HR CTV for manually contoured plans matched better with the 2.5D model (P > 0.6) and the results from the 2D model were slightly lower (P < 0.08). On the other hand, the IR CTV doses (D90) for all of the models were slightly lower (2D: -1.3 to -1.5 Gy and 2.5D: -0.5 to -0.6 Gy) and the differences were statistically significant for the 2D model (2D: P < 0.000002 and 2.5D: P > 0.06). In case of OARs, the 2.5D model segmentations resulted in closer dosimetry than 2D models (2D: P = 0.07-0.91 and 2.5D: P = 0.16-1.0). CONCLUSIONS The 2.5D deep learning models outperformed their respective 2D models for the automatic contouring of targets and OARs in MR image-based high-dose-rate (HDR) brachytherapy for cervical cancer. The InceptionResNetv2 model performed slightly better than ResNet50. This article is protected by copyright. All rights reserved.

In vivo assessment of hypertensive nephrosclerosis using ultrasound localization microscopy

Abstract: As a typical chronic kidney disease (CKD), hypertensive nephrosclerosis (HN) is a common syndrome of hypertension, characterized by chronic kidney microvascular damage. Early diagnosis of microvascular damage using conventional ultrasound imaging encounters challenges in sensitivity and specificity owing to the inherent diffraction limit. Ultrasound localization microscopy (ULM) has been developed to obtain microvasculature and microvascular hemodynamics within the kidney, and would be a promising tool for the early diagnosis of CKD.In this study, the advantage of quantitative indexes obtained by using ULM (mean arterial blood flow speeds of different segments of interlobular arteries) over indexes obtained using conventional clinical serum (β2-microglobulin, serum urea nitrogen, and creatinine) and urine (24-h urine volume and urine protein) tests and ultrasound Doppler imaging (peak systolic velocity [PSV], end-diastolic velocity [EDV], and resistance index [RI]) and contrast-enhanced ultrasound imaging (CEUS; rise time [RT], peak intensity [IMAX], mean transit time [mTT], and area under the time-intensity curve [AUC]) for early diagnosis of HN were investigated. Examinations were carried out on six spontaneously hypertensive rats (SHR) and five normal Wistar-Kyoto (WKY) rats at the age of 10 weeks.The experimental results show that the indicators derived from conventional clinical inspections (serum and urine tests) and ultrasound imaging (PSV, EDV, RI, RT, IMAX, mTT, and AUC) do not show significant difference between hypertensive and healthy rats (p > 0.05), while the TTP of the SHR group (28.52 ± 5.52 s) derived from CEUS is significantly higher than that of the WKY group (18.68 ± 7.32 s; p < 0.05). The mean blood flow speed in interlobular artery of SHR (12.47 ± 1.06 mm/s) derived from ULM is significantly higher than that of WKY rats (10.13 ± 1.17 mm/s; p < 0.01).The advantages of ULM over conventional clinical inspections and ultrasound imaging methods for early diagnosis of HN were validated. The quantitative results show that ULM can effectively diagnose HN at the early stage by detecting the blood flow speed changes of interlobular arteries. ULM may promise a reliable technique for early diagnosis of HN in the future.

Treatment planning for Flash radiotherapy: general aspects and applications to proton beams.

Abstract: The increased radioresistence of healthy tissues when irradiated at very high dose rates (known as the Flash effect) is a radiobiological mechanism that is currently investigated in order to increase the therapeutic ratio of radiotherapy treatments. To maximize the benefits of the clinical application of Flash, a patient-specific balance between different properties of the dose distribution should be found, i.e. Flash needs to be one of the variables considered in treatment planning. We investigated the Flash potential of three proton therapy planning and beam delivery techniques, each on a different anatomical region. Based on a set of beam delivery parameters, on hypotheses on the dose and dose rate thresholds needed for the Flash effect to occur, and on two definitions of Flash dose rate, we generated exemplary illustrations of the capabilities of current proton therapy equipment to generate Flash dose distributions. All techniques investigated could both produce dose distributions comparable with a conventional proton plan and reach the Flash regime, to an extent that was strongly dependent on the dose per fraction and the Flash dose threshold. The beam current, Flash dose rate threshold and dose rate definition typically had a more moderate effect on the amount of Flash dose in normal tissue. A systematic estimation of the impact of Flash on different patient anatomies and treatment protocols is possible only if Flash-specific treatment planning features become readily available. Planning evaluation tools such as a voxel-based dose delivery time structure, and the inclusion in the optimization cost function of parameters directly associated with Flash (e.g. beam current, spot delivery sequence and scanning speed), are needed to generate treatment plans that are taking full advantage of the potential benefits of the Flash effect. This article is protected by copyright. All rights reserved.

GPU-accelerated Monte Carlo-based online adaptive proton therapy - a feasibility study.

Abstract: PURPOSE To develop an online Graphic-Processing-Unit (GPU)-accelerated Monte-Carlo-based adaptive radiation therapy (ART) workflow for pencil beam scanning (PBS) proton therapy to address inter-fraction anatomical changes in patients treated with PBS. METHODS AND MATERIALS A four-step workflow was developed using our in-house developed GPU-accelerated Monte-Carlo-based treatment planning system to implement online Monte-Carlo-based ART for PBS. The first step conducts diffeomorphic demon-based deformable image registration (DIR) to propagate contours on the initial planning CT (pCT) to the verification CT (vCT) to form a new structure set. The second step performs forward dose calculation of the initial plan on the vCT with the propagated contours after manual approval (possible modifications involved). The third step triggers a re-optimization of the plan depending on whether the verification dose meets the clinical requirements or not. A robust evaluation will be done for both the verification plan in the second step and the re-opotimized plan in the third step. The fourth step involves a two-stage (before and after delivery) patient specific quality assurance (PSQA) of the re-optimized plan. The before-delivery PSQA is to compare the plan dose to the dose calculated using an independent fast open-source Monte Carlo code, MCsquare. The after-delivery PSQA is to compare the plan dose to the dose re-calculated using the log file (spot MU, spot position, and spot energy) collected during the delivery. Jaccard index (JI), Dice similarity coefficients (DSCs), and Hausdorff distance (HD) were used to assess the quality of the propagated contours in the first step. A commercial plan evaluation software, ClearCheck™, was integrated into the workflow to carry out efficient plan evaluation. 3D Gamma analysis was used during the fourth step to ensure the accuracy of the plan dose from re-optimization. Three patients with three different disease sites were chosen to evaluate the feasibility of the online ART workflow for PBS. RESULTS For all three patients, the propagated contours were found to have good volume conformance [JI (lowest-highest: 0.833-0.983) and DSC (0.909-0.992)] but sub-optimal boundary coincidence [HD (2.37-20.76 mm)] for organs at risk (OARs). The verification dose evaluated by ClearCheck™ showed significant degradation of the target coverage due to the inter-fractional anatomical changes. Re-optimization on the vCT resulted in great improvement of the plan quality to a clinically acceptable level. 3D Gamma analyses of PSQA confirmed the accuracy of the plan dose before delivery (mean Gamma index = 98.74% with a threshold of 2%/2 mm/10%), and after delivery based on the log files (mean Gamma index = 99.05% with a threshold of 2%/2 mm/10%). The average time cost for the complete execution of the workflow was around 858 seconds, excluding the time for manual intervention. CONCLUSION The proposed online ART workflow for PBS was demonstrated to be efficient and effective by generating a re-optimized plan that significantly improved the plan quality. This article is protected by copyright. All rights reserved.

Phantom task-based image quality assessment of three generations of rapid kV-switching dual-energy CT systems on virtual monoenergetic images.

Abstract: PURPOSE To compare the spectral performance of three rapid kV switching Dual-Energy CT (DECT) systems on virtual monoenergetic images (VMIs) at low-energy levels on abdominal imaging. METHODS A multi-energy phantom was scanned on three DECT systems equipped with three different Gemstone Spectral Imaging™ (GSI) platforms: GSI (1st generation, GSI-1st ), GSI-Pro (2nd generation, GSI-2nd ) and GSI-Xtream (3rd generation, GSI-3rd ). Acquisitions on the phantom were performed with a CTDIvol close to 11mGy. For all platforms, raw data were reconstructed using filtered-back projection (FBP) and a hybrid iterative reconstruction algorithm (ASIR-V at 50%; AV50). A deep-learning image reconstruction (DLR) algorithm (TrueFidelity™) was used only for the GSI-3rd . Noise power spectrum (NPS) and task-based transfer function (TTF) were evaluated from 40 to 80keV of VMIs. A detectability index (d') was computed to assess the detection of two contrast-enhanced lesions according to the keV level used. RESULTS For all GSI platforms, the noise magnitude decreased from 40 to 70keV, and using AV50 compared to FBP. The average NPS spatial frequency (fav ) and spatial resolution (TTF50% ) were similar from 40 to 70 keV and decreased with AV50 compared to FBP. Compared to AV50, using DLR reduced the noise magnitude (-27%±3%) and improved fav values (10%±0%) and altering spatial resolution (2%±5%). For the two lesions, d' values peaked at 70keV for GSI-1st and GSI-2nd platforms and at 40/50keV for GSI-3rd , for all reconstruction algorithms. The highest d' values were found for the GSI-3rd with DLR. CONCLUSION Differences in image quality were found between the GSI platforms for VMIs at low keV. New DLR algorithm on the GSI-3rd platform reduced noise and improved spatial resolution and detectability without changing the noise texture for VMIs at low keV. The choice of the best energy level in VMIs depends on the platform and the reconstruction algorithm. This article is protected by copyright. All rights reserved.

Increase of the transmission and emittance acceptance through a cyclotron‐based proton therapy gantry

Abstract: Abstract Purpose In proton therapy, the gantry, as the final part of the beamline, has a major effect on beam intensity and beam size at the isocenter. Most of the conventional beam optics of cyclotron‐based proton gantries have been designed with an imaging factor between 1 and 2 from the coupling point (CP) at the gantry entrance to the isocenter (patient location) meaning that to achieve a clinically desirable (small) beam size at isocenter, a small beam size is also required at the CP. Here we will show that such imaging factors are limiting the emittance which can be transported through the gantry. We, therefore, propose the use of large beam size and low divergence beam at the CP along with an imaging factor of 0.5 (2:1) in a new design of gantry beam optics to achieve substantial improvements in transmission and thus increase beam intensity at the isocenter. Methods The beam optics of our gantry have been re‐designed to transport higher emittance without the need of any mechanical modifications to the gantry beamline. The beam optics has been designed using TRANSPORT, with the resulting transmissions being calculated using Monte Carlo simulations (BDSIM code). Finally, the new beam optics have been tested with measurements performed on our Gantry 2 at PSI. Results With the new beam optics, we could maximize transmission through the gantry for a fixed emittance value. Additionally, we could transport almost four times higher emittance through the gantry compared to conventional optics, whilst achieving good transmissions through the gantry (>50%) with no increased losses in the gantry. As such, the overall transmission (cyclotron to isocenter) can be increased by almost a factor of 6 for low energies. Additionally, the point‐to‐point imaging inherent to the optics allows adjustment of the beam size at the isocenter by simply changing the beam size at the CP. Conclusion We have developed a new gantry beam optics which, by selecting a large beam size and low divergence at the gantry entrance and using an imaging factor of 0.5 (2:1), increases the emittance acceptance of the gantry, leading to a substantial increase in beam intensity at low energies. We expect that this approach could easily be adapted for most types of existing gantries.

A radiomics‐boosted deep‐learning model for COVID‐19 and non‐COVID‐19 pneumonia classification using chest x‐ray images

Abstract: To develop a deep learning model design that integrates radiomics analysis for enhanced performance of COVID-19 and non-COVID-19 pneumonia detection using chest x-ray images.As a novel radiomics approach, a 2D sliding kernel was implemented to map the impulse response of radiomic features throughout the entire chest x-ray image; thus, each feature is rendered as a 2D map in the same dimension as the x-ray image. Based on each of the three investigated deep neural network architectures, including VGG-16, VGG-19, and DenseNet-121, a pilot model was trained using x-ray images only. Subsequently, two radiomic feature maps (RFMs) were selected based on cross-correlation analysis in reference to the pilot model saliency map results. The radiomics-boosted model was then trained based on the same deep neural network architecture using x-ray images plus the selected RFMs as input. The proposed radiomics-boosted design was developed using 812 chest x-ray images with 262/288/262 COVID-19/non-COVID-19 pneumonia/healthy cases, and 649/163 cases were assigned as training-validation/independent test sets. For each model, 50 runs were trained with random assignments of training/validation cases following the 7:1 ratio in the training-validation set. Sensitivity, specificity, accuracy, and ROC curves together with area-under-the-curve (AUC) from all three deep neural network architectures were evaluated.After radiomics-boosted implementation, all three investigated deep neural network architectures demonstrated improved sensitivity, specificity, accuracy, and ROC AUC results in COVID-19 and healthy individual classifications. VGG-16 showed the largest improvement in COVID-19 classification ROC (AUC from 0.963 to 0.993), and DenseNet-121 showed the largest improvement in healthy individual classification ROC (AUC from 0.962 to 0.989). The reduced variations suggested improved robustness of the model to data partition. For the challenging non-COVID-19 pneumonia classification task, radiomics-boosted implementation of VGG-16 (AUC from 0.918 to 0.969) and VGG-19 (AUC from 0.964 to 0.970) improved ROC results, while DenseNet-121 showed a slight yet insignificant ROC performance reduction (AUC from 0.963 to 0.949). The achieved highest accuracy of COVID-19/non-COVID-19 pneumonia/healthy individual classifications were 0.973 (VGG-19)/0.936 (VGG-19)/ 0.933 (VGG-16), respectively.The inclusion of radiomic analysis in deep learning model design improved the performance and robustness of COVID-19/non-COVID-19 pneumonia/healthy individual classification, which holds great potential for clinical applications in the COVID-19 pandemic.

Predicting the outcome of radiotherapy in brain metastasis by integrating the clinical and MRI‐based deep learning features

Abstract: Abstract Background A considerable proportion of metastatic brain tumors progress locally despite stereotactic radiation treatment, and it can take months before such local progression is evident on follow‐up imaging. Prediction of radiotherapy outcome in terms of tumor local failure is crucial for these patients and can facilitate treatment adjustments or allow for early salvage therapies. Purpose In this work, a novel deep learning architecture is introduced to predict the outcome of local control/failure in brain metastasis treated with stereotactic radiation therapy using treatment‐planning magnetic resonance imaging (MRI) and standard clinical attributes. Methods At the core of the proposed architecture is an InceptionResentV2 network to extract distinct features from each MRI slice for local outcome prediction. A recurrent or transformer network is integrated into the architecture to incorporate spatial dependencies between MRI slices into the predictive modeling. A visualization method based on prediction difference analysis is coupled with the deep learning model to illustrate how different regions of each lesion on MRI contribute to the model's prediction. The model was trained and optimized using the data acquired from 99 patients (116 lesions) and evaluated on an independent test set of 25 patients (40 lesions). Results The results demonstrate the promising potential of the MRI deep learning features for outcome prediction, outperforming standard clinical variables. The prediction model with only clinical variables demonstrated an area under the receiver operating characteristic curve (AUC) of 0.68. The MRI deep learning models resulted in AUCs in the range of 0.72 to 0.83 depending on the mechanism to integrate information from MRI slices of each lesion. The best prediction performance (AUC = 0.86) was associated with the model that combined the MRI deep learning features with clinical variables and incorporated the inter‐slice dependencies using a long short‐term memory recurrent network. The visualization results highlighted the importance of tumor/lesion margins in local outcome prediction for brain metastasis. Conclusions The promising results of this study show the possibility of early prediction of radiotherapy outcome for brain metastasis via deep learning of MRI and clinical attributes at pre‐treatment and encourage future studies on larger groups of patients treated with other radiotherapy modalities.

Technical note: Validation of an ultrahigh dose rate pulsed electron beam monitoring system using a current transformer for FLASH preclinical studies

Abstract: Abstract Purpose The Oriatron eRT6 is a linear accelerator (linac) used in FLASH preclinical studies able to reach dose rates ranging from conventional (CONV) up to ultrahigh (UHDR). This work describes the implementation of commercially available beam current transformers (BCTs) as online monitoring tools compatible with CONV and UHDR irradiations for preclinical FLASH studies. Methods Two BCTs were used to measure the output of the Oriatron eRT6 linac. First, the correspondence between the set nominal beam parameters and those measured by the BCTs was checked. Then, we established the relationship between the total exit charge (measured by BCTs) and the absorbed dose to water. The influence of the pulse width (PW) and the pulse repetition frequency (PRF) at UHDR was characterized, as well as the short‐ and long‐term stabilities of the relationship between the exit charge and the dose at CONV and UHDR. Results The BCTs were able to determine consistently the number of pulses, PW, and PRF. For fixed PW and pulse height, the exit charge measured from BCTs was correlated with the dose, and linear relationships were found with uncertainties of 0.5 % and 3 % in CONV and UHDR mode, respectively. Short‐ and long‐term stabilities of the dose‐to‐charge ratio were below 1.6 %. Conclusions We implemented commercially available BCTs and demonstrated their ability to act as online beam monitoring systems to support FLASH preclinical studies with CONV and UHDR irradiations. The implemented BCTs support dosimetric measurements, highlight variations among multiple measurements in a row, enable monitoring of the physics parameters used for irradiation, and are an important step for the safety of the clinical translation of FLASH radiation therapy.

Real-time estimation of patient-specific dose distributions for medical CT using the deep dose estimation.

Abstract: PURPOSE With the rising number of CT examinations and the trend towards personalized medicine, patient-specific dose estimates are becoming more and more important in CT imaging. However, current approaches are often too slow or too inaccurate to be applied routinely. Therefore, we propose the so-called deep dose estimation (DDE) to provide highly accurate patient dose distributions in real-time. METHODS To combine accuracy and computational performance, the DDE algorithm uses a deep convolutional neural network to predict patient dose distributions. To do so, a U-net like architecture is trained to reproduce Monte Carlo simulations from a two-channel input consisting of a CT reconstruction and a first-order dose estimate. Here, the corresponding training data were generated using CT simulations based on 45 whole-body patient scans. For each patient, simulations were performed for different anatomies (pelvis, abdomen, thorax, head), different tube voltages (80 kV, 100 kV, 120 kV), different scan trajectories (circle, spiral), and with and without bowtie filtration and tube current modulation. Similar simulations were performed using a second set of 8 whole-body CT scans from the Visceral project to generate testing data. Finally, the DDE algorithm was evaluated with respect to the generalization to different scan parameters and the accuracy of organ dose and effective dose estimates based on an external organ segmentation. RESULTS DDE dose distributions were quantified in terms of the the mean absolute percentage error (MAPE) and a gamma analysis with respect to the ground truth Monte Carlo simulation. Both measures indicate that DDE generalizes well to different scan parameters and different anatomical regions with a maximum MAPE of 6.3 % and a minimum gamma passing rate of 91 %. Evaluating the organ dose values for all organs listed in the ICRP recommendation, shows an average error of 3.1 % and maximum error of 7.2 % (bone surface). CONCLUSIONS The DDE algorithm provides an efficient approach to determine highly accurate dose distributions. Being able to process a whole-body CT scan in about 1.5 s, it provides a valuable alternative to Monte Carlo simulations on a GPU. Here, the main advantage of DDE is that it can be used on top of any existing Monte Carlo code such that real-time performance can be achieved without major adjustments. Thus, DDE opens up new options not only for dosimetry but also for scan and protocol optimization. This article is protected by copyright. All rights reserved.

Design of static and dynamic ridge filters for FLASH-IMPT: a simulation study.

Abstract: PURPOSE This paper focused on the design and optimization of ridge filter-based intensity-modulated proton therapy (IMPT), and its potential applications for FLASH. Differing from the standard pencil beam scanning (PBS) mode, no energy/layer switching is required and total treatment time can be shortened. METHODS Unique dose influence matrices were generated as a proton beam traverse through slabs of different thicknesses (i.e. modulation by different layers). To establish the references for comparison, conventional IMPT plans (single field) were created using a large-scale non-linear solver. The spot weights from the reference IMPT plans were used as inputs for optimizing the design of ridge filters. Two designs were evaluated: model A (static) and model B (dynamic). The ridge filters designs were first verified (by GEANT4 simulation) in a water phantom and then in a H&N case. Direct comparison was made between the GEANT4 simulation results of two models and their respective references, with regard to plan quality, dose-averaged dose rate (DADR), and total treatment time. RESULTS In both the water phantom and the H&N case, two models are able to modulate dose distributions with high conformity, showing no significant difference relative to the reference plans. Dose rate volume histograms (DRVHs) suggest that in order to achieve a dose rate of 40 Gy/s over 90% PTV, the beam intensity needs to be 2.5×1011 protons/s for both models. For a fraction dose of 10 Gy, the total treatment time (including both irradiation time and dead time) can be shortened by a factor of 4.9 (model A) and 6.5 (model B), relative to the reference plans. CONCLUSION Two proposed designs (both static and dynamic) can be used for PBS-IMPT requiring no layer switching. They are promising candidates for FLASH-IMPT capable of reducing treatment time and achieving high dose rates, while maintaining dose conformity simultaneously. This article is protected by copyright. All rights reserved.

Patient‐derived heterogeneous breast phantoms for advanced dosimetry in mammography and tomosynthesis

Abstract: Abstract Background Understanding the magnitude and variability of the radiation dose absorbed by the breast fibroglandular tissue during mammography and digital breast tomosynthesis (DBT) is of paramount importance to assess risks versus benefits. Although homogeneous breast models have been proposed and used for decades for this purpose, they do not accurately reflect the actual heterogeneous distribution of the fibroglandular tissue in the breast, leading to biases in the estimation of dose from these modalities. Purpose To develop and validate a method to generate patient‐derived, heterogeneous digital breast phantoms for breast dosimetry in mammography and DBT. Methods The proposed phantoms were developed starting from patient‐based models of compressed breasts, generated for multiple thicknesses and representing the two standard views acquired in mammography and DBT, that is, cranio‐caudal (CC) and medio‐lateral‐oblique (MLO). Internally, the breast phantoms were defined as consisting of an adipose/fibroglandular tissue mixture, with a nonspatially uniform relative concentration. The parenchyma distributions were obtained from a previously described model based on patient breast computed tomography data that underwent simulated compression. Following these distributions, phantoms with any glandular fraction (1%–100%) and breast thickness (12–125 mm) can be generated, for both views. The phantoms were validated, in terms of their accuracy for average normalized glandular dose (DgN) estimation across samples of patient breasts, using 88 patient‐specific phantoms involving actual patient distribution of the fibroglandular tissue in the breast, and compared to that obtained using a homogeneous model similar to those currently used for breast dosimetry. Results The average DgN estimated for the proposed phantoms was concordant with that absorbed by the patient‐specific phantoms to within 5% (CC) and 4% (MLO). These DgN estimates were over 30% lower than those estimated with the homogeneous models, which overestimated the average DgN by 43% (CC), and 32% (MLO) compared to the patient‐specific phantoms. Conclusions The developed phantoms can be used for dosimetry simulations to improve the accuracy of dose estimates in mammography and DBT.