Manual or automatic delineation of the esophageal tumor in CT images is known to be very challenging. This is due to the low contrast between the tumor and adjacent tissues, the anatomical... Show moreManual or automatic delineation of the esophageal tumor in CT images is known to be very challenging. This is due to the low contrast between the tumor and adjacent tissues, the anatomical variation of the esophagus, as well as the occasional presence of foreign bodies (e.g. feeding tubes). Physicians therefore usually exploit additional knowledge such as endoscopic findings, clinical history, additional imaging modalities like PET scans. Achieving his additional information is time-consuming, while the results are error-prone and might lead to non-deterministic results. In this paper we aim to investigate if and to what extent a simplified clinical workflow based on CT alone, allows one to automatically segment the esophageal tumor with sufficient quality. For this purpose, we present a fully automatic end-to-end esophageal tumor segmentation method based on convolutional neural networks (CNNs). The proposed network, called Dilated Dense Attention Unet (DDAUnet), leverages spatial and channel attention gates in each dense block to selectively concentrate on determinant feature maps and regions. Dilated convolutional layers are used to manage GPU memory and increase the network receptive field. We collected a dataset of 792 scans from 288 distinct patients including varying anatomies with air pockets, feeding tubes and proximal tumors. Repeatability and reproducibility studies were conducted for three distinct splits of training and validation sets. The proposed network achieved a DSC value of 0.79 +/- 0.20, a mean surface distance of 5.4 +/- 20.2mm and 95% Hausdorff distance of 14.7 +/- 25.0mm for 287 test scans, demonstrating promising results with a simplified clinical workflow based on CT alone. Our code is publicly available via https://github.com/yousefis/DenseUnet_Esophagus_Segmentation. Show less
Elmahdy, M.S.; Beljaards, L.; Yousefi, S.; Sokooti, H.; Verbeek, F.; Heide, U.A. van der; Staring, M. 2021
Medical image registration and segmentation are two of the most frequent tasks in medical image analysis. As these tasks are complementary and correlated, it would be beneficial to apply them... Show moreMedical image registration and segmentation are two of the most frequent tasks in medical image analysis. As these tasks are complementary and correlated, it would be beneficial to apply them simultaneously in a joint manner. In this paper, we formulate registration and segmentation as a joint problem via a Multi-Task Learning (MTL) setting, allowing these tasks to leverage their strengths and mitigate their weaknesses through the sharing of beneficial information. We propose to merge these tasks not only on the loss level, but on the architectural level as well. We studied this approach in the context of adaptive image-guided radiotherapy for prostate cancer, where planning and follow-up CT images as well as their corresponding contours are available for training. At testing time the contours of the follow-up scans are not available, which is a common scenario in adaptive radiotherapy. The study involves two datasets from different manufacturers and institutes. The first dataset was divided into training (12 patients) and validation (6 patients), and was used to optimize and validate the methodology, while the second dataset (14 patients) was used as an independent test set. We carried out an extensive quantitative comparison between the quality of the automatically generated contours from different network architectures as well as loss weighting methods. Moreover, we evaluated the quality of the generated deformation vector field (DVF). We show that MTL algorithms outperform their Single-Task Learning (STL) counterparts and achieve better generalization on the independent test set. The best algorithm achieved a mean surface distance of 1.06 +/- 0.3 mm, 1.27 +/- 0.4 mm, 0.91 +/- 0.4 mm, and 1.76 +/- 0.8 mm on the validation set for the prostate, seminal vesicles, bladder, and rectum, respectively. The high accuracy of the proposed method combined with the fast inference speed, makes it a promising method for automatic re-contouring of follow-up scans for adaptive radiotherapy, potentially reducing treatment related complications and therefore improving patients quality-of-life after treatment. The source code is available at https://github.com/moelmahdy/JRS-MTL. Show less
Advanced endoscopic imaging is an emerging field in endoscopy practice, especially in optical diagnosis. Current technologies like virtual chromoendoscopy and small-field technologies allow... Show moreAdvanced endoscopic imaging is an emerging field in endoscopy practice, especially in optical diagnosis. Current technologies like virtual chromoendoscopy and small-field technologies allow visualization of subtle changes in mucosal and vascular patterns that are predictive of histology. The limiting factor in broadly utilizing these techniques is training and the need for reliable detection of these subtleties. This review provides the current evidence and limitations of training in advanced endoscopic imaging, and future directions of learning. A literature search was performed on PubMed and Medline through March 2020 with relevant keywords as advanced endoscopic imaging, training, and learning. References of relevant articles were screened for additional literature. Several didactic and web-based education programs are developed for training in virtual chromoendoscopy, autofluorescence imaging, confocal laser endomicroscopy, and volumetric laser endomicroscopy. Studies and post-hoc analysis on learning curves showed relatively steep learning curves after training, and web-based education seems to be as valuable as in-person didactic training for most techniques. However, consistent performance on expert level after training has not yet been demonstrated. Most advanced endoscopic imaging techniques are learned within a reasonable timeframe. Future efforts to enhance training and implementation of these techniques should focus on developing standardized and broadly incorporated training programs. The future role of artificial intelligence-assistance in advanced endoscopy and training has to be elucidated. Show less
Palagonia, E.; Mazzone, E.; Naeyer, G. de; D'Hondt, F.; Collins, J.; Wisz, P.; ... ; Dell'Oglio, P. 2020
Purpose To assess the available literature evidence that discusses the effect of surgical experience on patient outcomes in robotic setting. This information is used to help understand how we can... Show morePurpose To assess the available literature evidence that discusses the effect of surgical experience on patient outcomes in robotic setting. This information is used to help understand how we can develop a learning process that allows surgeons to maximally accommodate patient safety. Methods A literature search of the MEDLINE/PubMed and Scopus database was performed. Original and review articles published in the English language were included after an interactive peer-review process of the panel. Results Robotic surgical procedures require high level of experience to guarantee patient safety. This means that, for some procedures, the learning process might be longer than originally expected. In this context, structured training programs that assist surgeons to improve outcomes during their learning processes were extensively discussed. We identified few structured robotic curricula and demonstrated that for some procedures, curriculum trained surgeons can achieve outcomes rates during their initial learning phases that are at least comparable to those of experienced surgeons from high-volume centres. Finally, the importance of non-technical skills on patient safety and of their inclusion in robotic training programs was also assessed. Conclusion To guarantee safe robotic surgery and to optimize patient outcomes during the learning process, standardized and validated training programs are instrumental. To date, only few structured validated curricula exist for standardized training and further efforts are needed in this direction. Show less
Pezzotti, N.; Yousefi, S.; Elmahdy, M.S.; Gemert, J.H.F. van; Schuelke, C.; Doneva, M.; ... ; Staring, M. 2020
Adaptive intelligence aims at empowering machine learning techniques with the additional use of domain knowledge. In this work, we present the application of adaptive intelligence to accelerate MR... Show moreAdaptive intelligence aims at empowering machine learning techniques with the additional use of domain knowledge. In this work, we present the application of adaptive intelligence to accelerate MR acquisition. Starting from undersampled k-space data, an iterative learning-based reconstruction scheme inspired by compressed sensing theory is used to reconstruct the images. We developed a novel deep neural network to refine and correct prior reconstruction assumptions given the training data. The network was trained and tested on a knee MRI dataset from the 2019 fastMRI challenge organized by Facebook AI Research and NYU Langone Health. All submissions to the challenge were initially ranked based on similarity with a known groundtruth, after which the top 4 submissions were evaluated radiologically. Our method was evaluated by the fastMRI organizers on an independent challenge dataset. It ranked #1, shared #1, and #3 on respectively the 8x accelerated multi-coil, the 4x multi-coil, and the 4x single-coil tracks. This demonstrates the superior performance and wide applicability of the method. Show less
Objective: To develop a consensus-based set of generic competencies in antimicrobial prescribing and stewardship for European prescribers through a structured consensus procedure.Methods: The RAND... Show moreObjective: To develop a consensus-based set of generic competencies in antimicrobial prescribing and stewardship for European prescribers through a structured consensus procedure.Methods: The RAND-modified Delphi procedure comprised two online questionnaire rounds, a face-to-face meeting between rounds, and a final review. Our departure point was a set of competencies agreed previously by consensus among a UK multi-disciplinary panel, and which had been subsequently revised through consultation with ESCMID Study Group representatives. The 46 draft competency points were reviewed by an expert panel consisting of specialists in infectious diseases and clinical microbiology, and pharmacists. Each proposed competency was assessed using a nine-point Likert scale, for relevance as a minimum standard for all independent prescribers in all European countries.Results: A total of 65 expert panel members participated, from 24 European countries (one to six experts per country). There was very high satisfaction (98%) with the final competencies set, which included 35 competency points, in three sections: core concepts in microbiology, pathogenesis and diagnosing infections (11 points); antimicrobial prescribing (20 points); and antimicrobial stewardship (4 points).Conclusions: The consensus achieved enabled the production of generic antimicrobial prescribing and stewardship competencies for all European independent prescribers, and of possible global utility. These can be used for training and can be further adapted to the needs of specific professional groups. (C) 2018 The Authors. Published by Elsevier Ltd on behalf of European Society of Clinical Microbiology and Infectious Diseases. Show less
This thesis focused on the development of creative thinking across adolescence and into adulthood. To this end, a range of creativity tasks, both with and without an fMRI scanner, and before and... Show moreThis thesis focused on the development of creative thinking across adolescence and into adulthood. To this end, a range of creativity tasks, both with and without an fMRI scanner, and before and after training paradigms has been used to ex_amine both age- and experience-related effects on creative thinking performance during functional brain develop_ment. Chapter 1 provides a theoretical background for the research described in chapters 2 to 6. In Chapter 2, developmental trajectories of creative cognition across adolescence and early adulthood were examined using a set of tasks gauging both insight and divergent thinking in the verbal and visual domain. In Chapter 3, behavioral and neural differences for creative problem solving in middle-adolescents and adults were examined. Chapter 4 examined the neural correlates of divergent thinking in adults and adolescents. Chapter 5 focused on the effectiveness of creative ideation training in adolescents and adults. In Chapter 6, the benefits of training creativity in adolescents were examined using fMRI techniques. Finally, Chapter 7 summarizes the main results of the empirical studies presented in this thesis. Here, implications of the results are discussed and suggestions for future research are presented. Show less
Fetoscopic surgery is a surgical technique that is used to treat fetus(es) that are still inside the pregnant uterus. Coming years, more fetoscopic surgery will be performed. The most commonly... Show moreFetoscopic surgery is a surgical technique that is used to treat fetus(es) that are still inside the pregnant uterus. Coming years, more fetoscopic surgery will be performed. The most commonly performed procedure is laser surgery for twin-twin transfusion syndrome. This thesis shows learning curves for this procedure and current practice in relation to technical aspects and pregnancy outcomes. We show how to monitor performance and address specific subgroups in which laser surgery can be more complicated. Since teaching and training in fetoscopic surgery is challenging, we create and validate an evidence-based evaluation tool for the laser procedure. To conclude, we develop a standardized training curriculum with a high fidelity simulator model. Show less
In recent years a flow of media reports about unsafe situations in operating rooms have reached the general public. Awareness of the importance of patient safety also reached politicians. The... Show moreIn recent years a flow of media reports about unsafe situations in operating rooms have reached the general public. Awareness of the importance of patient safety also reached politicians. The report by the Dutch inspectorate of health care __Risico__s minimaal invasieve chirurgie onderschat__ (Risks minimally invasive surgery underestimated) stressed that patient safety is especially at risk in Minimally Invasive Surgery. Therefore patient safety became a focus of research and quality improvement, also in minimally invasive surgery. The current thesis aimes to give insight into patient safety risk factors in minimally invasive surgery. Of all examined risk factors minimally invasive surgical skills appeared to be directly related to patient safety. Therefore special focus for training of these skills is necessary. Previous research has shown that during simulation training objective assessment of economy of movements and time is possible. However, until recently there was no way to objectively assess one of the most important surgical skills: tissue handling. The development of a force sensor has made it possible to measure interaction forces with artificial tissue. In this thesis the clinical implications of a force sensor and the need of this new technology within training of minimally invasive surgical skills is examined. Show less