@article{ScheppachRauberStallhoferetal., author = {Scheppach, Markus W. and Rauber, David and Stallhofer, Johannes and Muzalyova, Anna and Otten, Vera and Manzeneder, Carolin and Schwamberger, Tanja and Wanzl, Julia and Schlottmann, Jakob and Tadic, Vidan and Probst, Andreas and Schnoy, Elisabeth and R{\"o}mmele, Christoph and Fleischmann, Carola and Meinikheim, Michael and Miller, Silvia and M{\"a}rkl, Bruno and Stallmach, Andreas and Palm, Christoph and Messmann, Helmut and Ebigbo, Alanna}, title = {Detection of duodenal villous atrophy on endoscopic images using a deep learning algorithm}, series = {Gastrointestinal Endoscopy}, journal = {Gastrointestinal Endoscopy}, publisher = {Elsevier}, doi = {10.1016/j.gie.2023.01.006}, abstract = {Background and aims Celiac disease with its endoscopic manifestation of villous atrophy is underdiagnosed worldwide. The application of artificial intelligence (AI) for the macroscopic detection of villous atrophy at routine esophagogastroduodenoscopy may improve diagnostic performance. Methods A dataset of 858 endoscopic images of 182 patients with villous atrophy and 846 images from 323 patients with normal duodenal mucosa was collected and used to train a ResNet 18 deep learning model to detect villous atrophy. An external data set was used to test the algorithm, in addition to six fellows and four board certified gastroenterologists. Fellows could consult the AI algorithm's result during the test. From their consultation distribution, a stratification of test images into "easy" and "difficult" was performed and used for classified performance measurement. Results External validation of the AI algorithm yielded values of 90 \%, 76 \%, and 84 \% for sensitivity, specificity, and accuracy, respectively. Fellows scored values of 63 \%, 72 \% and 67 \%, while the corresponding values in experts were 72 \%, 69 \% and 71 \%, respectively. AI consultation significantly improved all trainee performance statistics. While fellows and experts showed significantly lower performance for "difficult" images, the performance of the AI algorithm was stable. Conclusion In this study, an AI algorithm outperformed endoscopy fellows and experts in the detection of villous atrophy on endoscopic still images. AI decision support significantly improved the performance of non-expert endoscopists. The stable performance on "difficult" images suggests a further positive add-on effect in challenging cases.}, language = {en} } @misc{ZellmerRauberProbstetal., author = {Zellmer, Stephan and Rauber, David and Probst, Andreas and Weber, Tobias and Nagl, Sandra and R{\"o}mmele, Christoph and Schnoy, Elisabeth and Palm, Christoph and Messmann, Helmut and Ebigbo, Alanna}, title = {Verwendung k{\"u}nstlicher Intelligenz bei der Detektion der Papilla duodeni major}, series = {Zeitschrift f{\"u}r Gastroenterologie}, volume = {61}, journal = {Zeitschrift f{\"u}r Gastroenterologie}, number = {08}, publisher = {Thieme}, address = {Stuttgart}, doi = {10.1055/s-0043-1772000}, pages = {e593-e540}, abstract = {Einleitung Die Endoskopische Retrograde Cholangiopankreatikographie (ERCP) ist der Goldstandard in der Diagnostik und Therapie von Erkrankungen des pankreatobili{\"a}ren Trakts. Jedoch ist sie technisch sehr anspruchsvoll und weist eine vergleichsweise hohe Komplikationsrate auf. Ziele In der vorliegenden Machbarkeitsstudie soll gepr{\"u}ft werden, ob mithilfe eines Deep-learning-Algorithmus die Papille und das Ostium zuverl{\"a}ssig detektiert werden k{\"o}nnen und somit f{\"u}r Endoskopiker mit geringer Erfahrung ein geeignetes Hilfsmittel, insbesondere f{\"u}r die Ausbildungssituation, darstellen k{\"o}nnten. Methodik Wir betrachteten insgesamt 606 Bilddatens{\"a}tze von 65 Patienten. In diesen wurde sowohl die Papilla duodeni major als auch das Ostium segmentiert. Anschließend wurde eine neuronales Netz mittels eines Deep-learning-Algorithmus trainiert. Außerdem erfolgte eine 5-fache Kreuzvaldierung. Ergebnisse Bei einer 5-fachen Kreuzvaldierung auf den 606 gelabelten Daten konnte f{\"u}r die Klasse Papille eine F1-Wert von 0,7908, eine Sensitivit{\"a}t von 0,7943 und eine Spezifit{\"a}t von 0,9785 erreicht werden, f{\"u}r die Klasse Ostium eine F1-Wert von 0,5538, eine Sensitivit{\"a}t von 0,5094 und eine Spezifit{\"a}t von 0,9970 (vgl. [Tab. 1]). Unabh{\"a}ngig von der Klasse zeigte sich gemittelt (Klasse Papille und Klasse Ostium) ein F1-Wert von 0,6673, eine Sensitivit{\"a}t von 0,6519 und eine Spezifit{\"a}t von 0,9877 (vgl. [Tab. 2]). Schlussfolgerung In vorliegende Machbarkeitsstudie konnte das neuronale Netz die Papilla duodeni major mit einer hohen Sensitivit{\"a}t und sehr hohen Spezifit{\"a}t identifizieren. Bei der Detektion des Ostiums war die Sensitivit{\"a}t deutlich geringer. Zuk{\"u}nftig soll das das neuronale Netz mit mehr Daten trainiert werden. Außerdem ist geplant, den Algorithmus auch auf Videos anzuwenden. Somit k{\"o}nnte langfristig ein geeignetes Hilfsmittel f{\"u}r die ERCP etabliert werden.}, language = {de} } @misc{ScheppachRauberStallhoferetal., author = {Scheppach, Markus and Rauber, David and Stallhofer, Johannes and Muzalyova, Anna and Otten, Vera and Manzeneder, Carolin and Schwamberger, Tanja and Wanzl, Julia and Schlottmann, Jakob and Tadic, Vidan and Probst, Andreas and Schnoy, Elisabeth and R{\"o}mmele, Christoph and Fleischmann, Carola and Meinikheim, Michael and Miller, Silvia and M{\"a}rkl, Bruno and Palm, Christoph and Messmann, Helmut and Ebigbo, Alanna}, title = {Performance comparison of a deep learning algorithm with endoscopists in the detection of duodenal villous atrophy (VA)}, series = {Endoscopy}, volume = {55}, journal = {Endoscopy}, number = {S02}, publisher = {Thieme}, doi = {10.1055/s-0043-1765421}, pages = {S165}, abstract = {Aims VA is an endoscopic finding of celiac disease (CD), which can easily be missed if pretest probability is low. In this study, we aimed to develop an artificial intelligence (AI) algorithm for the detection of villous atrophy on endoscopic images. Methods 858 images from 182 patients with VA and 846 images from 323 patients with normal duodenal mucosa were used for training and internal validation of an AI algorithm (ResNet18). A separate dataset was used for external validation, as well as determination of detection performance of experts, trainees and trainees with AI support. According to the AI consultation distribution, images were stratified into "easy" and "difficult". Results Internal validation showed 82\%, 85\% and 84\% for sensitivity, specificity and accuracy. External validation showed 90\%, 76\% and 84\%. The algorithm was significantly more sensitive and accurate than trainees, trainees with AI support and experts in endoscopy. AI support in trainees was associated with significantly improved performance. While all endoscopists showed significantly lower detection for "difficult" images, AI performance remained stable. Conclusions The algorithm outperformed trainees and experts in sensitivity and accuracy for VA detection. The significant improvement with AI support suggests a potential clinical benefit. Stable performance of the algorithm in "easy" and "difficult" test images may indicate an advantage in macroscopically challenging cases.}, language = {en} } @misc{MeinikheimMendelProbstetal., author = {Meinikheim, Michael and Mendel, Robert and Probst, Andreas and Scheppach, Markus W. and Nagl, Sandra and Schnoy, Elisabeth and R{\"o}mmele, Christoph and Prinz, Friederike and Schlottmann, Jakob and Messmann, Helmut and Palm, Christoph and Ebigbo, Alanna}, title = {Einfluss von K{\"u}nstlicher Intelligenz auf die Performance von niedergelassenen Gastroenterolog:innen bei der Beurteilung von Barrett-{\"O}sophagus}, series = {Zeitschrift f{\"u}r Gastroenterologie}, volume = {61}, journal = {Zeitschrift f{\"u}r Gastroenterologie}, number = {8}, publisher = {Thieme}, address = {Stuttgart}, doi = {10.1055/s-0043-1771711}, abstract = {Einleitung Die Differenzierung zwischen nicht dysplastischem Barrett-{\"O}sophagus (NDBE) und mit Barrett-{\"O}sophagus assoziierten Neoplasien (BERN) w{\"a}hrend der endoskopischen Inspektion erfordert viel Expertise. Die fr{\"u}he Diagnosestellung ist wichtig f{\"u}r die weitere Prognose des Barrett-Karzinoms. In Deutschland werden Patient:innen mit einem Barrett-{\"O}sophagus (BE) in der Regel im niedergelassenen Sektor {\"u}berwacht. Ziele Ziel ist es, den Einfluss von einem auf K{\"u}nstlicher Intelligenz (KI) basierenden klinischen Entscheidungsunterst{\"u}tzungssystems (CDSS) auf die Performance von niedergelassenen Gastroenterolog:innen (NG) bei der Evaluation von Barrett-{\"O}sophagus (BE) zu untersuchen. Methodik Es erfolgte die prospektive Sammlung von 96 unver{\"a}nderten hochaufl{\"o}senden Videos mit F{\"a}llen von Patient:innen mit histologisch best{\"a}tigtem NDBE und BERN. Alle eingeschlossenen F{\"a}lle enthielten mindestens zwei der folgenden Darstellungsmethoden: HD-Weißlichtendoskopie, Narrow Band Imaging oder Texture and Color Enhancement Imaging. Sechs NG von sechs unterschiedlichen Praxen wurden als Proband:innen eingeschlossen. Es erfolgte eine permutierte Block-Randomisierung der Videof{\"a}lle in entweder Gruppe A oder Gruppe B. Gruppe A implizierte eine Evaluation des Falls durch Proband:innen zun{\"a}chst ohne KI und anschließend mit KI als CDSS. In Gruppe B erfolgte die Evaluation in umgekehrter Reihenfolge. Anschließend erfolgte eine zuf{\"a}llige Wiedergabe der so entstandenen Subgruppen im Rahmen des Tests. Ergebnis In diesem Test konnte ein von uns entwickeltes KI-System (Barrett-Ampel) eine Sensitivit{\"a}t von 92,2\%, eine Spezifit{\"a}t von 68,9\% und eine Accuracy von 81,3\% erreichen. Mit der Hilfe von KI verbesserte sich die Sensitivit{\"a}t der NG von 64,1\% auf 71,2\% (p<0,001) und die Accuracy von 66,3\% auf 70,8\% (p=0,006) signifikant. Eine signifikante Verbesserung dieser Parameter zeigte sich ebenfalls, wenn die Proband:innen die F{\"a}lle zun{\"a}chst ohne KI evaluierten (Gruppe A). Wurde der Fall jedoch als Erstes mit der Hilfe von KI evaluiert (Gruppe B), blieb die Performance nahezu konstant. Schlussfolgerung Es konnte ein performantes KI-System zur Evaluation von BE entwickelt werden. NG verbessern sich bei der Evaluation von BE durch den Einsatz von KI.}, language = {de} } @misc{MeinikheimMendelProbstetal., author = {Meinikheim, Michael and Mendel, Robert and Probst, Andreas and Scheppach, Markus W. and Schnoy, Elisabeth and Nagl, Sandra and R{\"o}mmele, Christoph and Prinz, Friederike and Schlottmann, Jakob and Golger, Daniela and Palm, Christoph and Messmann, Helmut and Ebigbo, Alanna}, title = {AI-assisted detection and characterization of early Barrett's neoplasia: Results of an Interim analysis}, series = {Endoscopy}, volume = {55}, journal = {Endoscopy}, number = {S02}, publisher = {Thieme}, doi = {10.1055/s-0043-1765437}, pages = {S169}, abstract = {Aims Evaluation of the add-on effect an artificial intelligence (AI) based clinical decision support system has on the performance of endoscopists with different degrees of expertise in the field of Barrett's esophagus (BE) and Barrett's esophagus-related neoplasia (BERN). Methods The support system is based on a multi-task deep learning model trained to solve a segmentation and several classification tasks. The training approach represents an extension of the ECMT semi-supervised learning algorithm. The complete system evaluates a decision tree between estimated motion, classification, segmentation, and temporal constraints, to decide when and how the prediction is highlighted to the observer. In our current study, ninety-six video cases of patients with BE and BERN were prospectively collected and assessed by Barrett's specialists and non-specialists. All video cases were evaluated twice - with and without AI assistance. The order of appearance, either with or without AI support, was assigned randomly. Participants were asked to detect and characterize regions of dysplasia or early neoplasia within the video sequences. Results Standalone sensitivity, specificity, and accuracy of the AI system were 92.16\%, 68.89\%, and 81.25\%, respectively. Mean sensitivity, specificity, and accuracy of expert endoscopists without AI support were 83,33\%, 58,20\%, and 71,48 \%, respectively. Gastroenterologists without Barrett's expertise but with AI support had a comparable performance with a mean sensitivity, specificity, and accuracy of 76,63\%, 65,35\%, and 71,36\%, respectively. Conclusions Non-Barrett's experts with AI support had a similar performance as experts in a video-based study.}, language = {en} } @article{MendelRauberdeSouzaJretal., author = {Mendel, Robert and Rauber, David and de Souza Jr., Luis Antonio and Papa, Jo{\~a}o Paulo and Palm, Christoph}, title = {Error-Correcting Mean-Teacher: Corrections instead of consistency-targets applied to semi-supervised medical image segmentation}, series = {Computers in Biology and Medicine}, volume = {154}, journal = {Computers in Biology and Medicine}, number = {March}, publisher = {Elsevier}, issn = {0010-4825}, doi = {10.1016/j.compbiomed.2023.106585}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:898-opus4-57790}, pages = {13}, abstract = {Semantic segmentation is an essential task in medical imaging research. Many powerful deep-learning-based approaches can be employed for this problem, but they are dependent on the availability of an expansive labeled dataset. In this work, we augment such supervised segmentation models to be suitable for learning from unlabeled data. Our semi-supervised approach, termed Error-Correcting Mean-Teacher, uses an exponential moving average model like the original Mean Teacher but introduces our new paradigm of error correction. The original segmentation network is augmented to handle this secondary correction task. Both tasks build upon the core feature extraction layers of the model. For the correction task, features detected in the input image are fused with features detected in the predicted segmentation and further processed with task-specific decoder layers. The combination of image and segmentation features allows the model to correct present mistakes in the given input pair. The correction task is trained jointly on the labeled data. On unlabeled data, the exponential moving average of the original network corrects the student's prediction. The combined outputs of the students' prediction with the teachers' correction form the basis for the semi-supervised update. We evaluate our method with the 2017 and 2018 Robotic Scene Segmentation data, the ISIC 2017 and the BraTS 2020 Challenges, a proprietary Endoscopic Submucosal Dissection dataset, Cityscapes, and Pascal VOC 2012. Additionally, we analyze the impact of the individual components and examine the behavior when the amount of labeled data varies, with experiments performed on two distinct segmentation architectures. Our method shows improvements in terms of the mean Intersection over Union over the supervised baseline and competing methods. Code is available at https://github.com/CloneRob/ECMT.}, language = {en} } @unpublished{RueckertRueckertPalm, author = {Rueckert, Tobias and Rueckert, Daniel and Palm, Christoph}, title = {Methods and datasets for segmentation of minimally invasive surgical instruments in endoscopic images and videos: A review of the state of the art}, doi = {10.48550/arXiv.2304.13014}, pages = {25}, abstract = {In the field of computer- and robot-assisted minimally invasive surgery, enormous progress has been made in recent years based on the recognition of surgical instruments in endoscopic images. Especially the determination of the position and type of the instruments is of great interest here. Current work involves both spatial and temporal information with the idea, that the prediction of movement of surgical tools over time may improve the quality of final segmentations. The provision of publicly available datasets has recently encouraged the development of new methods, mainly based on deep learning. In this review, we identify datasets used for method development and evaluation, as well as quantify their frequency of use in the literature. We further present an overview of the current state of research regarding the segmentation and tracking of minimally invasive surgical instruments in endoscopic images. The paper focuses on methods that work purely visually without attached markers of any kind on the instruments, taking into account both single-frame segmentation approaches as well as those involving temporal information. A discussion of the reviewed literature is provided, highlighting existing shortcomings and emphasizing available potential for future developments. The publications considered were identified through the platforms Google Scholar, Web of Science, and PubMed. The search terms used were "instrument segmentation", "instrument tracking", "surgical tool segmentation", and "surgical tool tracking" and result in 408 articles published between 2015 and 2022 from which 109 were included using systematic selection criteria.}, language = {en} } @inproceedings{MendelRauberPalm, author = {Mendel, Robert and Rauber, David and Palm, Christoph}, title = {Exploring the Effects of Contrastive Learning on Homogeneous Medical Image Data}, series = {Bildverarbeitung f{\"u}r die Medizin 2023: Proceedings, German Workshop on Medical Image Computing, July 2- 4, 2023, Braunschweig}, booktitle = {Bildverarbeitung f{\"u}r die Medizin 2023: Proceedings, German Workshop on Medical Image Computing, July 2- 4, 2023, Braunschweig}, publisher = {Springer Vieweg}, address = {Wiesbaden}, doi = {10.1007/978-3-658-41657-7}, pages = {128 -- 13}, abstract = {We investigate contrastive learning in a multi-task learning setting classifying and segmenting early Barrett's cancer. How can contrastive learning be applied in a domain with few classes and low inter-class and inter-sample variance, potentially enabling image retrieval or image attribution? We introduce a data sampling strategy that mines per-lesion data for positive samples and keeps a queue of the recent projections as negative samples. We propose a masking strategy for the NT-Xent loss that keeps the negative set pure and removes samples from the same lesion. We show cohesion and uniqueness improvements of the proposed method in feature space. The introduction of the auxiliary objective does not affect the performance but adds the ability to indicate similarity between lesions. Therefore, the approach could enable downstream auto-documentation tasks on homogeneous medical image data.}, language = {en} } @misc{RueckertRiederRauberetal., author = {R{\"u}ckert, Tobias and Rieder, Maximilian and Rauber, David and Xiao, Michel and Humolli, Eg and Feussner, Hubertus and Wilhelm, Dirk and Palm, Christoph}, title = {Augmenting instrument segmentation in video sequences of minimally invasive surgery by synthetic smoky frames}, series = {International Journal of Computer Assisted Radiology and Surgery}, volume = {18}, journal = {International Journal of Computer Assisted Radiology and Surgery}, number = {Suppl 1}, publisher = {Springer Nature}, doi = {10.1007/s11548-023-02878-2}, pages = {S54 -- S56}, language = {en} } @incollection{Palm, author = {Palm, Christoph}, title = {History, Core Concepts, and Role of AI in Clinical Medicine}, series = {AI in Clinical Medicine: A Practical Guide for Healthcare Professionals}, booktitle = {AI in Clinical Medicine: A Practical Guide for Healthcare Professionals}, editor = {Byrne, Michael F. and Parsa, Nasim and Greenhill, Alexandra T. and Chahal, Daljeet and Ahmad, Omer and Bargci, Ulas}, edition = {1. Aufl.}, publisher = {Wiley}, isbn = {978-1-119-79064-8}, doi = {10.1002/9781119790686.ch5}, pages = {49 -- 55}, abstract = {The field of AI is characterized by robust promises, astonishing successes, and remarkable breakthroughs. AI will play a major role in all domains of clinical medicine, but the role of AI in relation to the physician is not yet completely determined. The term artificial intelligence or AI is broad, and several different terms are used in this context that must be organized and demystified. This chapter will review the key concepts and methods of AI, and will introduce some of the different roles for AI in relation to the physician.}, language = {en} }