In US-guided cardiac radioablation, a possible workflow includes simultaneous US and planning CT acquisitions, which can result in US transducer-induced metal artifacts on the planning CT scans. To reduce the impact of these artifacts, a metal artifact reduction (MAR) algorithm has been developed based on a deep learning Generative Adversarial Network called Cycle-MAR, and compared with iMAR (Siemens), O-MAR (Philips) and MDT (ReVision Radiology), and CCS-MAR (Combined Clustered Scan-based MAR). Cycle-MAR was trained with a supervised learning scheme using sets of paired clinical CT scans with and without simulated artifacts. It was then evaluated on CT scans with real artifacts of an anthropomorphic phantom, and on sets of clinical CT scans with simulated artifacts which were not used for Cycle-MAR training. Image quality metrics and HU value-based analysis were used to evaluate the performance of Cycle-MAR compared to the other algorithms. The proposed Cycle-MAR network effectively reduces the negative impact of the metal artifacts. For example, the calculated HU value improvement percentage for the cardiac structures in the clinical CT scans was 59.58%, 62.22%, and 72.84% after MDT, CCS-MAR, and Cycle-MAR application, respectively. The application of MAR algorithms reduces the impact of US transducer-induced metal artifacts on CT scans. In comparison to iMAR, O-MAR, MDT, and CCS-MAR, the application of developed Cycle-MAR network on CT scans performs better in reducing these metal artifacts.

Cardiac radioablation is a promising treatment for cardiac arrhythmias, but accurate dose delivery can be affected by heart motion. For this reason, real-time cardiac motion monitoring during radioablation is of paramount importance. Real-time ultrasound (US) guidance can be a solution. The US-guided cardiac radioablation workflow can be simplified by the simultaneous US and planning computed tomography (CT) acquisition, which can result in US transducer-induced metal artifacts on the planning CT scans. To reduce the impact of these artifacts, a new metal artifact reduction (MAR) algorithm (named: Combined Clustered Scan-based MAR [CCS-MAR]) has been developed and compared with iMAR (Siemens), O-MAR (Philips) and MDT (ReVision Radiology) algorithms. CCS-MAR is a fully automated sinogram inpainting-based MAR algorithm, which uses a two-stage correction process based on a normalized MAR method. The second stage aims to correct errors remaining from the first stage to create an artifact-free combined clustered scan for the process of metal artifact reduction. To evaluate the robustness of CCS-MAR, conventional CT scans and/or dual-energy CT scans from three anthropomorphic phantoms and transducers with different sizes were used. The performance of CCS-MAR for metal artifact reduction was compared with other algorithms through visual comparison, image quality metrics analysis, and HU value restoration evaluation. The results of this study show that CCS-MAR effectively reduced the US transducer-induced metal artifacts and that it improved HU value accuracy more or comparably to other MAR algorithms. These promising results justify future research into US transducer-induced metal artifact reduction for the US-guided cardiac radioablation purposes.

Background:The management of the cardio-respiratory motion of the target and the reduction of the uncertainties related to patient's positioning are two of the main challenges that stereotactic arrhythmia radio-ablation (STAR) has to overcome. A prototype of a system was developed that can automatically acquire and interpret echocardiographic images using an artificial intelligence (AI) algorithm to calculate cardiac displacement in real-time.Methods:We conducted a single center study enrolling consecutive patients with a history of ventricular arrhythmias (VA) in order to evaluate the feasibility of this automatic acquisition system. Echocardiographic images were automatically acquired from the parasternal and apical views with a dedicated probe. The system was designed to hold the probe fixed to the chest in the supine position during both free-breathing and short expiratory breath-hold sequences, to simulate STAR treatment. The primary endpoint was the percentage of patients reaching a score ≥2 in a multi-parametric assessment evaluating the quality of automatically acquired images. Moreover, we investigated the potential impact of clinical and demographic characteristics on achieving the primary endpoint.Results:We enrolled 24 patients (63 ± 14 years, 21% females). All of them had a history of VA and 21 (88%) had an ICD. Eight patients (33%) had coronary artery disease, 12 (50%) had non-ischemic cardiomyopathy, and 3 had idiopathic VA. Parasternal, as well as apical images were obtained from all patients except from one, in whom parasternal view could not be collected due to the patient's inability to maintain the supine position. The primary endpoint was achieved in 23 patients (96%) for the apical view, in 20 patients (87%) for the parasternal view, and in all patients in at least one of the two views. The images' quality was maximal (i.e., score = 4) in at least one of the two windows in 19 patients (79%). Atrial fibrillation arrhythmia was the only clinical characteristics associated with a poor score outcome in both imaging windows (apical p = 0.022, parasternal p = 0.014).Conclusions:These results provide the proof-of-concept for the feasibility of an automatic ultrasonographic image acquisition system associated with an AI algorithm for real-time monitoring of cardiac motion in patients with a history of VA.

This work presents an algorithm based on weak supervision to automatically localize an arthroscope on 3D ultrasound (US). The ultimate goal of this application is to combine 3D US with the 2D arthroscope view during knee arthroscopy, to provide the surgeon with a comprehensive view of the surgical site. The implemented algorithm consisted of a weakly supervised neural network, which was trained on 2D US images of different phantoms mimicking the imaging conditions during knee arthroscopy. Image-based classification was performed and the resulting class activation maps were used to localize the arthroscope. The localization performance was evaluated visually by three expert reviewers and by the calculation of objective metrics. Finally, the algorithm was also tested on a human cadaver knee. The algorithm achieved an average classification accuracy of 88.6% on phantom data and 83.3% on cadaver data. The localization of the arthroscope based on the class activation maps was correct in 92–100% of all true positive classifications for both phantom and cadaver data. These results are relevant because they show feasibility of automatic arthroscope localization in 3D US volumes, which is paramount to combining multiple image modalities that are available during knee arthroscopies.

This book provides a comprehensive overview of the role of 3D/4D ultrasound imaging in radiotherapy. It identifies barriers to the clinical implementation of ultrasound-guided radiotherapy, and provides a critical discussion of potential solutions as well as highlighting exciting new opportunities for research in this field.

This is a practical guide to the implementation of 3D/4D ultrasound imaging in radiography. Among its features are the coverage of the technology utilised for ultrasound-guided radiotherapy, clinical need and the advantages of using ultrasound. It is a useful tool for users that incorporates implementation, potential errors, uncertainties and training. This is a comprehensive review of the state-of-the-art technologies, which also looks at the future direction of this exciting field. Researchers, students, hospital physicists and radiographers will all find this book of use as it guides them through current clinical situation and examines the full potential of ultrasound in radiotherapy.

Metal artifact reduction (MAR) methods are used to reduce artifacts from metals or metal components in computed tomography (CT). In radiotherapy (RT), CT is the most used imaging modality for planning, whose quality is often affected by metal artifacts. The aim of this study is to systematically review the impact of MAR methods on CT Hounsfield Unit values, contouring of regions of interest, and dose calculation for RT applications. This systematic review is performed in accordance with the PRISMA guidelines; the PubMed and Web of Science databases were searched using the main keywords "metal artifact reduction", "computed tomography" and "radiotherapy". A total of 382 publications were identified, of which 40 (including one review article) met the inclusion criteria and were included in this review. The selected publications (except for the review article) were grouped into two main categories: commercial MAR methods and research-based MAR methods. Conclusion: The application of MAR methods on CT scans can improve treatment planning quality in RT. However, none of the investigated or proposed MAR methods was completely satisfactory for RT applications because of limitations such as the introduction of other errors (e.g., other artifacts) or image quality degradation (e.g., blurring), and further research is still necessary to overcome these challenges.

Chapter 10 explores the most recent and fascinating tools from the artificial intelligence (AI) world and their use in ultrasound-guided radiotherapy.

The first challenge encountered when introducing ultrasound into radiotherapy is described in chapter 3: localising the probe in the simulation suite and in the treatment room to enable comparison of the positions of anatomical structures at the two stages. The solutions available are introduced together with a detailed description and analysis of the sources of error and uncertainty. After the probe has been localised, the next steps are to evaluate if the image acquired is of sufficient quality and segment all the relevant structures of interest.

Knee arthroscopy is a complex minimally invasive surgery that can cause unintended injuries to femoral cartilage or postoperative complications, or both. Autonomous robotic systems using real-time volumetric ultrasound (US) imaging guidance hold potential for reducing significantly these issues and for improving patient outcomes. To enable the robotic system to navigate autonomously in the knee joint, the imaging system should provide the robot with a real-time comprehensive map of the surgical site. To this end, the first step is automatic image quality assessment, to ensure that the boundaries of the relevant knee structures are defined well enough to be detected, outlined, and then tracked. In this article, a recently developed one-class classifier deep learning algorithm was used to discriminate among the US images acquired in a simulated surgical scenario on which the femoral cartilage either could or could not be outlined. A total of 38 656 2-D US images were extracted from 151 3-D US volumes, collected from six volunteers, and were labeled as "1" or as "0" when an expert was or was not able to outline the cartilage on the image, respectively. The algorithm was evaluated using the expert labels as ground truth with a fivefold cross validation, where each fold was trained and tested on average with 15 640 and 6246 labeled images, respectively. The algorithm reached a mean accuracy of 78.4% ± 5.0, mean specificity of 72.5% ± 9.4, mean sensitivity of 82.8% ± 5.8, and mean area under the curve of 85% ± 4.4. In addition, interobserver and intraobserver tests involving two experts were performed on an image subset of 1536 2-D US images. Percent agreement values of 0.89 and 0.93 were achieved between two experts (i.e., interobserver) and by each expert (i.e., intraobserver), respectively. These results show the feasibility of the first essential step in the development of automatic US image acquisition and interpretation systems for autonomous robotic knee arthroscopy.

The tracking of the knee femoral condyle cartilage during ultrasound-guided minimally invasive procedures is important to avoid damaging this structure during such interventions. In this study, we propose a new deep learning method to track, accurately and efficiently, the femoral condyle cartilage in ultrasound sequences, which were acquired under several clinical conditions, mimicking realistic surgical setups. Our solution, that we name Siam-U-Net, requires minimal user initialization and combines a deep learning segmentation method with a siamese framework for tracking the cartilage in temporal and spatio-temporal sequences of 2D ultrasound images. Through extensive performance validation given by the Dice Similarity Coefficient, we demonstrate that our algorithm is able to track the femoral condyle cartilage with an accuracy which is comparable to experienced surgeons. It is additionally shown that the proposed method outperforms state-of-the-art segmentation models and trackers in the localization of the cartilage. We claim that the proposed solution has the potential for ultrasound guidance in minimally invasive knee procedures.

External beam radiotherapy (EBRT) is one of the curative treatment options for prostate cancer patients. The aim of this treatment option is to irradiate tumor tissue, while sparing normal tissue as much as possible. Frequent imaging during the course of the treatment (image guided radiotherapy) allows for determination of the location and shape of the prostate (target) and of the organs at risk. This information is used to increase accuracy in radiation dose delivery resulting in better tumor control and lower toxicity. Ultrasound imaging is harmless for the patient, it is cost-effective, and it allows for real-time volumetric organ tracking. For these reasons, it is an ideal technique for image guidance during EBRT workflows. Review papers have been published in which the use of ultrasound imaging in EBRT workflows for different cancer sites (prostate, breast, etc.) was extensively covered. This new review paper aims at providing the readers with an update on the current status for prostate cancer ultrasound guided EBRT treatments.

PurposeThe use of ultrasound imaging is not widespread in prostate cancer radiotherapy workflows, despite several advantages (eg, allowing real‐time volumetric organ tracking). This can be partially attributed to the need for a trained operator during acquisition and interpretation of the images. We introduce and evaluate an algorithm that can propose a patient‐specific transperineal ultrasound probe setup, based on a CT scan and anatomical structure delineations. The use of this setup during the simulation and treatment stage could improve usability of ultrasound imaging for relatively untrained operators (radiotherapists with less than 1 yr experience with ultrasound).MethodsThe internal perineum boundaries of three prostate cancer patients were identified based on bone masks extracted from their CT scans. After projection of these boundaries to the skin and exclusion of specific areas, this resulted in a skin area accessible for transperineal ultrasound probe placement in clinical practice. Several possible probe setups on this area were proposed by the algorithm and the optimal setup was automatically selected. In the end, this optimal setup was evaluated based on a comparison with a corresponding transperineal ultrasound volume acquired by a radiation oncologist.ResultsThe algorithm‐proposed setups allowed visualization of 100% of the clinically required anatomical structures, including the whole prostate and seminal vesicles, as well as the adjacent edges of the bladder and rectum. In addition, these setups allowed visualization of 94% of the anatomical structures, which were also visualized by the physician during the acquisition of an actual ultrasound volume.ConclusionProvided that the ultrasound probe setup proposed by the algorithm, is properly reproduced on the patient, it allows visualization of all clinically required structures for image guided radiotherapy purposes. Future work should validate these results on a patient population and optimize the workflow to enable a relatively untrained operator to perform the procedure.

The use of ultrasound guidance in prostate cancer radiotherapy workflows is not widespread. This can be partially attributed to the need for image interpretation by a trained operator during ultrasound image acquisition. In this work, a one-class regressor, based on DenseNet and Gaussian processes, was implemented to assess automatically the quality of transperineal ultrasound images of the male pelvic region. The implemented deep learning approach achieved a scoring accuracy of 94%, a specificity of 95% and a sensitivity of 93% with respect to the majority vote of three experts, which was comparable with the results of these experts. This is the first step towards a fully automatic workflow, which could potentially remove the need for image interpretation and thereby make the use of ultrasound imaging, which allows real-time volumetric organ tracking in the RT environment, more appealing for hospitals.

In many medical fields, spatial localization of devices is of paramount importance to e.g. deliver treatments correctly or to perform accurate diagnostic evaluations. Among the different technologies available, optical localization in the visible band of the spectrum offers many advantages: minimum interference with other instruments, flexibility in position of the devices used and extra features such as patient/operator surface tracking or vital signs monitoring. Its application in the domain of radiotherapy is novel, and potentially very beneficial. In this work we introduced a workflow to test the reliability of one of these systems, developed in house, to localize an ultrasound transducer for ultrasound guided radiation therapy (USgRT). Accuracy, precision, latency, sensitive volume and sensitivity to light intensity were evaluated. The workflow we proposed allowed us to establish that, while all the clinically relevant parameters of the system are acceptable, accuracy in locating the transducer could be as poor as 7-8 mm, therefore not acceptable for the intended purpose of the system.

Purpose: Interpretation of ultrasound (US) images is crucial during their acquisition. Typically, a trained operator needs to verify if the correct anatomical structures are visualized with sufficient quality. This is one of the key reasons why US imaging is not widely used in radiotherapy workflows. In this study, we explored the feasibility of providing the operator with patient-specific US probe setups for 4D transperineal (TPUS) imaging of prostate cancer patients. These setups were automatically calculated based on a CT scan of the patient. Methods: After bone mask extraction from a CT volume of a prostate cancer patient, internal perineum boundaries and the perineal area on the patients’ skin were automatically identified. Subsequently, the areas on which the US probe could not be positioned in clinical practice (e.g. around the anus) were removed resulting in a skin area accessible for TPUS imaging. Translation and rotation of a virtual probe on this skin area, while taking into account probe pressure effects, resulted in 7546 possible probe setups. Finally, we used anatomical structure delineations to assess if these setups allowed visualization of the prostate, the seminal vesicles, and the adjacent edges of bladder and rectum, as clinically required. Results: In total 392 of the suitable setups potentially allowed visualization of all clinically required structures without bone blockage. The best setup (out of 392) potentially allows visualization of 100% of the prostate and seminal vesicles, 78% of the bladder and 69% of the rectum. Conclusion: In this proof-of-concept study, we showed that it is feasible to provide the operator with patient-specific probe setups (based on a CT scan) that satisfy all clinical requirements. In future steps, we will investigate the clinical requirements for implementation in the radiotherapy workflow. In the end, the use of these setups can potentially remove the need for skilled operators.