TY - GEN A1 - Roser, David A1 - Meinikheim, Michael A1 - Mendel, Robert A1 - Palm, Christoph A1 - Probst, Andreas A1 - Muzalyova, Anna A1 - Scheppach, Markus W. A1 - Nagl, Sandra A1 - Schnoy, Elisabeth A1 - Römmele, Christoph A1 - Schulz, Dominik Andreas Helmut Otto A1 - Schlottmann, Jakob A1 - Prinz, Friederike A1 - Rauber, David A1 - Rückert, Tobias A1 - Matsumura, Tomoaki A1 - Fernandez-Esparrach, G. A1 - Parsa, Nasim A1 - Byrne, Michael F. A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Human-Computer Interaction: Impact of Artificial Intelligence on the diagnostic confidence of endoscopists assessing videos of Barrett’s esophagus T2 - Endoscopy N2 - Aims Human-computer interactions (HCI) may have a relevant impact on the performance of Artificial Intelligence (AI). Studies show that although endoscopists assessing Barrett’s esophagus (BE) with AI improve their performance significantly, they do not achieve the level of the stand-alone performance of AI. One aspect of HCI is the impact of AI on the degree of certainty and confidence displayed by the endoscopist. Indirectly, diagnostic confidence when using AI may be linked to trust and acceptance of AI. In a BE video study, we aimed to understand the impact of AI on the diagnostic confidence of endoscopists and the possible correlation with diagnostic performance. Methods 22 endoscopists from 12 centers with varying levels of BE experience reviewed ninety-six standardized endoscopy videos. Endoscopists were categorized into experts and non-experts and randomly assigned to assess the videos with and without AI. Participants were randomized in two arms: Arm A assessed videos first without AI and then with AI, while Arm B assessed videos in the opposite order. Evaluators were tasked with identifying BE-related neoplasia and rating their confidence with and without AI on a scale from 0 to 9. Results The utilization of AI in Arm A (without AI first, with AI second) significantly elevated confidence levels for experts and non-experts (7.1 to 8.0 and 6.1 to 6.6, respectively). Only non-experts benefitted from AI with a significant increase in accuracy (68.6% to 75.5%). Interestingly, while the confidence levels of experts without AI were higher than those of non-experts with AI, there was no significant difference in accuracy between these two groups (71.3% vs. 75.5%). In Arm B (with AI first, without AI second), experts and non-experts experienced a significant reduction in confidence (7.6 to 7.1 and 6.4 to 6.2, respectively), while maintaining consistent accuracy levels (71.8% to 71.8% and 67.5% to 67.1%, respectively). Conclusions AI significantly enhanced confidence levels for both expert and non-expert endoscopists. Endoscopists felt significantly more uncertain in their assessments without AI. Furthermore, experts with or without AI consistently displayed higher confidence levels than non-experts with AI, irrespective of comparable outcomes. These findings underscore the possible role of AI in improving diagnostic confidence during endoscopic assessment. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1782859 SN - 1438-8812 VL - 56 IS - S 02 SP - 79 PB - Georg Thieme Verlag ER - TY - GEN A1 - Roser, David A1 - Meinikheim, Michael A1 - Mendel, Robert A1 - Palm, Christoph A1 - Muzalyova, Anna A1 - Rauber, David A1 - Rückert, Tobias A1 - Parsa, Nasim A1 - Byrne, Michael F. A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Mensch-Maschine-Interaktion: Einfluss künstlicher Intelligenz auf das diagnostische Vertrauen von Endoskopikern bei der Beurteilung des Barrett-Ösophagus T2 - Zeitschrift für Gastroenterologie N2 - Ziele:  Das Ziel der Studie war es, den Einfluss von KI auf die diagnostische Sicherheit (Konfidenzniveau) von Endoskopikern anhand von BÖ-Videos zu untersuchen und mögliche Korrelationen mit der Untersuchungsqualität zu erforschen. Methodik:  22 Endoskopiker aus zwölf Zentren mit unterschiedlicher Barrett-Erfahrung untersuchten 96 standardisierte Endoskopievideos. Die Untersucher wurden in Experten und Nicht-Experten eingeteilt und nach dem Zufallsprinzip für die Bewertung der Videos mit oder ohne KI eingeteilt. Die Teilnehmer wurden in zwei Gruppen aufgeteilt: Arm A bewertete zunächst Videos ohne KI und dann mit KI, während Arm B die umgekehrte Reihenfolge einhielt. Die Untersucher hatten die Aufgabe, BÖ-assoziierte Neoplasien zu erkennen und ihr Konfidenzniveau sowohl mit als auch ohne KI auf einer Skala von 0 bis 9 anzugeben. Ergebnis:  In Arm A erhöhte der Einsatz von KI das Konfidenzniveau bei beiden signifikant (p<0.001). Bemerkenswert ist, dass jedoch nur Nicht-Experten durch die KI eine signifikante Verbesserung der Sensitivität und Spezifität (p<0.001 bzw. p<0.05) erfuhren. Während Experten ohne KI im Vergleich zu Nicht-Experten mit KI ein höheres Konfidenzniveau aufwiesen, gab es keinen signifikanten Unterschied in der Genauigkeit. In Arm B zeigten beide Gruppen eine signifikante Abnahme des Konfidenzniveaus (p<0.001) bei gleichbleibender Genauigkeit. Darüber hinaus wurden in 9% der Entscheidungen trotz korrekter KI eine falsche Wahl getroffen. Schlussfolgerung:  Der Einsatz künstlicher Intelligenz steigerte das Konfidenzniveau sowohl bei Experten als auch bei Nicht-Experten signifikant – ein Effekt, der im Studienmodell reversibel war. Darüber hinaus wiesen Experten mit oder ohne KI durchweg höhere Konfidenzniveaus auf als Nicht-Experten mit KI, trotz vergleichbarer Ergebnisse. Zudem konnte beobachtet werden, dass die Untersucher in 9% der Fälle die KI zuungunsten des Patienten ignorierten. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1789656 VL - 62 IS - 09 SP - e575 EP - e576 PB - Georg Thieme Verlag KG ER - TY - JOUR A1 - Wallner, M. A1 - Gutbrod, Max A1 - Rauber, David A1 - Ebigbo, Alanna A1 - Probst, Andreas A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Roser, David T1 - KI-gestützte Detektion und Segmentierung von Magenkarzinomen in westlichen endoskopischen Bilddaten anhand eines fine-tuned Deep-Learning Ansatzes JF - Zeitschrift für Gastroenterologie N2 - Diese vorläufige monozentrische Studie zeigt, dass ein aus einem Barrett-Ösophagus-KI-System feinjustiertes Deep-Learning-Modell Magenkarzinome in westlichen multimodalen endoskopischen Bilddaten zuverlässig detektieren und präzise segmentieren kann. Die hohe Segmentierungsgenauigkeit und Detektionssensitivität über verschiedene Bildmodalitäten hinweg unterstreichen die Machbarkeit eines pathologiegestützten KI-Ansatzes auch in einer westlichen Niedriginzidenzpopulation. Aufgrund der ausschließlichen Verwendung von Bildern mit sichtbaren Tumoren lassen sich keine Aussagen zur Spezifität treffen; eine Übertragbarkeit auf Screening- oder Mischkollektive ist daher limitiert. Weitere Studien mit a) größerem Datensatz inklusive Videodaten, b) externer Validierung an einer multizentrischen westlichen Kohorte, sowie c) Anwendung und Prüfung an nicht-neoplastischen Vergleichsbildern oder anderen Pathologien sind erforderlich. Nach unserem Kenntnisstand zählt dieses System zu den ersten in einer westlichen Population entwickelten endoskopischen KI-Ansätzen zur Magenkarzinomdetektion, und zu wenigen, die vollständige ESD-präparatbasierte Referenzdaten für Training und Validierung nutzen. Y1 - 2026 U6 - https://doi.org/10.1055/s-0046-1817751 VL - 64 IS - 03 SP - e64 EP - e65 PB - Thieme ER - TY - JOUR A1 - Ebigbo, Alanna A1 - Mendel, Robert A1 - Scheppach, Markus W. A1 - Probst, Andreas A1 - Shahidi, Neal A1 - Prinz, Friederike A1 - Fleischmann, Carola A1 - Römmele, Christoph A1 - Gölder, Stefan Karl A1 - Braun, Georg A1 - Rauber, David A1 - Rückert, Tobias A1 - Souza Jr., Luis Antonio de A1 - Papa, João Paulo A1 - Byrne, Michael F. A1 - Palm, Christoph A1 - Messmann, Helmut T1 - Vessel and tissue recognition during third-space endoscopy using a deep learning algorithm JF - Gut N2 - In this study, we aimed to develop an artificial intelligence clinical decision support solution to mitigate operator-dependent limitations during complex endoscopic procedures such as endoscopic submucosal dissection and peroral endoscopic myotomy, for example, bleeding and perforation. A DeepLabv3-based model was trained to delineate vessels, tissue structures and instruments on endoscopic still images from such procedures. The mean cross-validated Intersection over Union and Dice Score were 63% and 76%, respectively. Applied to standardised video clips from third-space endoscopic procedures, the algorithm showed a mean vessel detection rate of 85% with a false-positive rate of 0.75/min. These performance statistics suggest a potential clinical benefit for procedure safety, time and also training. KW - Artificial Intelligence KW - Endoscopy KW - Medical Image Computing Y1 - 2022 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:898-opus4-54293 VL - 71 IS - 12 SP - 2388 EP - 2390 PB - BMJ CY - London ER - TY - JOUR A1 - Meinikheim, Michael A1 - Mendel, Robert A1 - Palm, Christoph A1 - Probst, Andreas A1 - Muzalyova, Anna A1 - Scheppach, Markus W. A1 - Nagl, Sandra A1 - Schnoy, Elisabeth A1 - Römmele, Christoph A1 - Schulz, Dominik Andreas Helmut Otto A1 - Schlottmann, Jakob A1 - Prinz, Friederike A1 - Rauber, David A1 - Rückert, Tobias A1 - Matsumura, Tomoaki A1 - Fernández-Esparrach, Glòria A1 - Parsa, Nasim A1 - Byrne, Michael F. A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Influence of artificial intelligence on the diagnostic performance of endoscopists in the assessment of Barrett’s esophagus: a tandem randomized and video trial JF - Endoscopy N2 - Background This study evaluated the effect of an artificial intelligence (AI)-based clinical decision support system on the performance and diagnostic confidence of endoscopists in their assessment of Barrett’s esophagus (BE). Methods 96 standardized endoscopy videos were assessed by 22 endoscopists with varying degrees of BE experience from 12 centers. Assessment was randomized into two video sets: group A (review first without AI and second with AI) and group B (review first with AI and second without AI). Endoscopists were required to evaluate each video for the presence of Barrett’s esophagus-related neoplasia (BERN) and then decide on a spot for a targeted biopsy. After the second assessment, they were allowed to change their clinical decision and confidence level. Results AI had a stand-alone sensitivity, specificity, and accuracy of 92.2%, 68.9%, and 81.3%, respectively. Without AI, BE experts had an overall sensitivity, specificity, and accuracy of 83.3%, 58.1%, and 71.5%, respectively. With AI, BE nonexperts showed a significant improvement in sensitivity and specificity when videos were assessed a second time with AI (sensitivity 69.8% [95%CI 65.2%–74.2%] to 78.0% [95%CI 74.0%–82.0%]; specificity 67.3% [95%CI 62.5%–72.2%] to 72.7% [95%CI 68.2%–77.3%]). In addition, the diagnostic confidence of BE nonexperts improved significantly with AI. Conclusion BE nonexperts benefitted significantly from additional AI. BE experts and nonexperts remained significantly below the stand-alone performance of AI, suggesting that there may be other factors influencing endoscopists’ decisions to follow or discard AI advice. KW - Artificial Intelligence KW - Endoscopy KW - Medical Image Computing Y1 - 2024 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:898-opus4-72818 VL - 56 SP - 641 EP - 649 PB - Georg Thieme Verlag CY - Stuttgart ER - TY - GEN A1 - Ebigbo, Alanna A1 - Rauber, David A1 - Ayoub, Mousa A1 - Birzle, Lisa A1 - Matsumura, Tomoaki A1 - Probst, Andreas A1 - Steinbrück, Ingo A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Meinikheim, Michael A1 - Scheppach, Markus W. A1 - Palm, Christoph A1 - Messmann, Helmut T1 - Early Esophageal Cancer and the Generalizability of Artificial Intelligence T2 - Endoscopy N2 - Aims Artificial Intelligence (AI) systems in gastrointestinal endoscopy are narrow because they are trained to solve only one specific task. Unlike Narrow-AI, general AI systems may be able to solve multiple and unrelated tasks. We aimed to understand whether an AI system trained to detect, characterize, and segment early Barrett’s neoplasia (Barrett’s AI) is only capable of detecting this pathology or can also detect and segment other diseases like early squamous cell cancer (SCC). Methods 120 white light (WL) and narrow-band endoscopic images (NBI) from 60 patients (1 WL and 1 NBI image per patient) were extracted from the endoscopic database of the University Hospital Augsburg. Images were annotated by three expert endoscopists with extensive experience in the diagnosis and endoscopic resection of early esophageal neoplasias. An AI system based on DeepLabV3+architecture dedicated to early Barrett’s neoplasia was tested on these images. The AI system was neither trained with SCC images nor had it seen the test images prior to evaluation. The overlap between the three expert annotations („expert-agreement“) was the ground truth for evaluating AI performance. Results Barrett’s AI detected early SCC with a mean intersection over reference (IoR) of 92% when at least 1 pixel of the AI prediction overlapped with the expert-agreement. When the threshold was increased to 5%, 10%, and 20% overlap with the expert-agreement, the IoR was 88%, 85% and 82%, respectively. The mean Intersection Over Union (IoU) – a metric according to segmentation quality between the AI prediction and the expert-agreement – was 0.45. The mean expert IoU as a measure of agreement between the three experts was 0.60. Conclusions In the context of this pilot study, the predictions of SCC by a Barrett’s dedicated AI showed some overlap to the expert-agreement. Therefore, features learned from Barrett’s cancer-related training might be helpful also for SCC prediction. Our results allow different possible explanations. On the one hand, some Barrett’s cancer features generalize toward the related task of assessing early SCC. On the other hand, the Barrett’s AI is less specific to Barrett’s cancer than a general predictor of pathological tissue. However, we expect to enhance the detection quality significantly by extending the training to SCC-specific data. The insight of this study opens the way towards a transfer learning approach for more efficient training of AI to solve tasks in other domains. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1783775 VL - 56 IS - S 02 SP - S428 PB - Thieme CY - Stuttgart ER - TY - GEN A1 - Scheppach, Markus W. A1 - Mendel, Robert A1 - Rauber, David A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Meinikheim, Michael A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Artificial Intelligence (AI) improves endoscopists’ vessel detection during endoscopic submucosal dissection (ESD) T2 - Endoscopy N2 - Aims While AI has been successfully implemented in detecting and characterizing colonic polyps, its role in therapeutic endoscopy remains to be elucidated. Especially third space endoscopy procedures like ESD and peroral endoscopic myotomy (POEM) pose a technical challenge and the risk of operator-dependent complications like intraprocedural bleeding and perforation. Therefore, we aimed at developing an AI-algorithm for intraprocedural real time vessel detection during ESD and POEM. Methods A training dataset consisting of 5470 annotated still images from 59 full-length videos (47 ESD, 12 POEM) and 179681 unlabeled images was used to train a DeepLabV3+neural network with the ECMT semi-supervised learning method. Evaluation for vessel detection rate (VDR) and time (VDT) of 19 endoscopists with and without AI-support was performed using a testing dataset of 101 standardized video clips with 200 predefined blood vessels. Endoscopists were stratified into trainees and experts in third space endoscopy. Results The AI algorithm had a mean VDR of 93.5% and a median VDT of 0.32 seconds. AI support was associated with a statistically significant increase in VDR from 54.9% to 73.0% and from 59.0% to 74.1% for trainees and experts, respectively. VDT significantly decreased from 7.21 sec to 5.09 sec for trainees and from 6.10 sec to 5.38 sec for experts in the AI-support group. False positive (FP) readings occurred in 4.5% of frames. FP structures were detected significantly shorter than true positives (0.71 sec vs. 5.99 sec). Conclusions AI improved VDR and VDT of trainees and experts in third space endoscopy and may reduce performance variability during training. Further research is needed to evaluate the clinical impact of this new technology. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1782891 VL - 56 IS - S 02 SP - S93 PB - Thieme CY - Stuttgart ER - TY - GEN A1 - Zellmer, Stephan A1 - Rauber, David A1 - Probst, Andreas A1 - Weber, Tobias A1 - Braun, Georg A1 - Römmele, Christoph A1 - Nagl, Sandra A1 - Schnoy, Elisabeth A1 - Messmann, Helmut A1 - Ebigbo, Alanna A1 - Palm, Christoph T1 - Artificial intelligence as a tool in the detection of the papillary ostium during ERCP T2 - Endoscopy N2 - Aims Endoscopic retrograde cholangiopancreaticography (ERCP) is the gold standard in the diagnosis as well as treatment of diseases of the pancreatobiliary tract. However, it is technically complex and has a relatively high complication rate. In particular, cannulation of the papillary ostium remains challenging. The aim of this study is to examine whether a deep-learning algorithm can be used to detect the major duodenal papilla and in particular the papillary ostium reliably and could therefore be a valuable tool for inexperienced endoscopists, particularly in training situation. Methods We analyzed a total of 654 retrospectively collected images of 85 patients. Both the major duodenal papilla and the ostium were then segmented. Afterwards, a neural network was trained using a deep-learning algorithm. A 5-fold cross-validation was performed. Subsequently, we ran the algorithm on 5 prospectively collected videos of ERCPs. Results 5-fold cross-validation on the 654 labeled data resulted in an F1 value of 0.8007, a sensitivity of 0.8409 and a specificity of 0.9757 for the class papilla, and an F1 value of 0.5724, a sensitivity of 0.5456 and a specificity of 0.9966 for the class ostium. Regardless of the class, the average F1 value (class papilla and class ostium) was 0.6866, the sensitivity 0.6933 and the specificity 0.9861. In 100% of cases the AI-detected localization of the papillary ostium in the prospectively collected videos corresponded to the localization of the cannulation performed by the endoscopist. Conclusions In the present study, the neural network was able to identify the major duodenal papilla with a high sensitivity and high specificity. In detecting the papillary ostium, the sensitivity was notably lower. However, when used on videos, the AI was able to identify the location of the subsequent cannulation with 100% accuracy. In the future, the neural network will be trained with more data. Thus, a suitable tool for ERCP could be established, especially in the training situation. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1783138 VL - 56 IS - S 02 SP - S198 PB - Thieme CY - Stuttgart ER - TY - GEN A1 - Scheppach, Markus W. A1 - Nunes, Danilo Weber A1 - Arizi, X. A1 - Rauber, David A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Meinikheim, Michael A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Procedural phase recognition in endoscopic submucosal dissection (ESD) using artificial intelligence (AI) T2 - Endoscopy N2 - Aims Recent evidence suggests the possibility of intraprocedural phase recognition in surgical operations as well as endoscopic interventions such as peroral endoscopic myotomy and endoscopic submucosal dissection (ESD) by AI-algorithms. The intricate measurement of intraprocedural phase distribution may deepen the understanding of the procedure. Furthermore, real-time quality assessment as well as automation of reporting may become possible. Therefore, we aimed to develop an AI-algorithm for intraprocedural phase recognition during ESD. Methods A training dataset of 364385 single images from 9 full-length ESD videos was compiled. Each frame was classified into one procedural phase. Phases included scope manipulation, marking, injection, application of electrical current and bleeding. Allocation of each frame was only possible to one category. This training dataset was used to train a Video Swin transformer to recognize the phases. Temporal information was included via logarithmic frame sampling. Validation was performed using two separate ESD videos with 29801 single frames. Results The validation yielded sensitivities of 97.81%, 97.83%, 95.53%, 85.01% and 87.55% for scope manipulation, marking, injection, electric application and bleeding, respectively. Specificities of 77.78%, 90.91%, 95.91%, 93.65% and 84.76% were measured for the same parameters. Conclusions The developed algorithm was able to classify full-length ESD videos on a frame-by-frame basis into the predefined classes with high sensitivities and specificities. Future research will aim at the development of quality metrics based on single-operator phase distribution. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1783804 VL - 56 IS - S 02 SP - S439 PB - Thieme CY - Stuttgart ER - TY - GEN A1 - Scheppach, Markus W. A1 - Rauber, David A1 - Stallhofer, Johannes A1 - Muzalyova, Anna A1 - Otten, Vera A1 - Manzeneder, Carolin A1 - Schwamberger, Tanja A1 - Wanzl, Julia A1 - Schlottmann, Jakob A1 - Tadic, Vidan A1 - Probst, Andreas A1 - Schnoy, Elisabeth A1 - Römmele, Christoph A1 - Fleischmann, Carola A1 - Meinikheim, Michael A1 - Miller, Silvia A1 - Märkl, Bruno A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Performance comparison of a deep learning algorithm with endoscopists in the detection of duodenal villous atrophy (VA) T2 - Endoscopy N2 - 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. Y1 - 2023 U6 - https://doi.org/10.1055/s-0043-1765421 VL - 55 IS - S02 PB - Thieme ER - TY - GEN A1 - Scheppach, Markus W. A1 - Weber Nunes, Danilo A1 - Arizi, X. A1 - Rauber, David A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Single frame workflow recognition during endoscopic submucosal dissection (ESD) using artificial intelligence (AI) T2 - Endoscopy N2 - Aims  Precise surgical phase recognition and evaluation may improve our understanding of complex endoscopic procedures. Furthermore, quality control measurements and endoscopy training could benefit from objective descriptions of surgical phase distributions. Therefore, we aimed to develop an artificial intelligence algorithm for frame-by-frame operational phase recognition during endoscopic submucosal dissection (ESD). Methods  Full length ESD-videos from 31 patients comprising 6.297.782 single images were collected retrospectively. Videos were annotated on a frame-by-frame basis for the operational macro-phases diagnostics, marking, injection, dissection and bleeding. Further subphases were the application of electrical current, visible injection of fluid into the submucosal space and scope manipulation, leading to 11 phases in total. 4.975.699 frames (21 patients) were used for training of a video swin transformer using uniform frame sampling for temporal information. Hyperparameter tuning was performed with 897.325 further frames (6 patients), while 424.758 frames (4 patients) were used for validation. Results  The overall F1 scores on the test dataset for the macro-phases and all 11 phases were 0.96 and 0.90, respectively. The recall values for diagnostics, marking, injection, dissection and bleeding were 1.00, 1.00, 0.95, 0.96 and 0.93, respectively. Conclusions  The algorithm classified operational phases during ESD with high accuracy. A precise evaluation of phase distribution may allow for the development of objective quality metrics for quality control and training. Y1 - 2025 U6 - https://doi.org/10.1055/s-0045-1806324 VL - 57 IS - S 02 SP - S511 PB - Thieme CY - Stuttgart ER - TY - GEN A1 - Zellmer, Stephan A1 - Rauber, David A1 - Probst, Andreas A1 - Weber, Tobias A1 - Braun, Georg A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Schnoy, Elisabeth A1 - Birzle, Lisa A1 - Aehling, Niklas A1 - Schulz, Dominik Andreas Helmut Otto A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Künstliche Intelligenz als Hilfsmittel zur Detektion der Papilla duodeni major und des papillären Ostiums während der ERCP T2 - Zeitschrift für Gastroenterologie N2 - Einleitung  Die Endoskopische Retrograde Cholangiopankreatikographie (ERCP) ist der Goldstandard in der endoskopischen Therapie von Erkrankungen des pankreatobiliären Trakts. Allerdings ist sie technisch anspruchsvoll, schwer zu erlernen und mit einer relativ hohen Komplikationsrate assoziiert. Daher soll in der vorliegenden Machbarkeitsstudie geprüft werden, ob mithilfe eines Deeplearning- Algorithmus die Papille und das Ostium zuverlässig detektiert werden können und dieser für Endoskopiker, insbesondere in der Ausbildungssituation, ein geeignetes Hilfsmittel darstellen könnte. Material und Methodik Insgesamt wurden 1534 ERCP-Bilder von 134 Patienten analysiert, wobei sowohl die Papilla duodeni major als auch das Ostium segmentiert wurden. Anschließend erfolgte das Training eines neuronalen Netzes unter Verwendung eines Deep-Learning-Algorithmus. Für den Test des Algorithmus erfolgte eine fünffache Kreuzvalidierung. Ergebnisse  Auf den 1534 gelabelten Bildern wurden für die Klasse Papille ein F1-Wert von 0,7996, eine Sensitivität von 0,8488 und eine Spezifität von 0,9822 erzielt. Für die Klasse Ostium ergaben sich ein F1-Wert von 0,5198, eine Sensitivität von 0,5945 und eine Spezifität von 0,9974. Klassenübergreifend (Klasse Papille und Klasse Ostium) betrug der F1-Wert 0,6593, die Sensitivität 0,7216 und für die Spezifität 0,9898. Zusammenfassung  In der vorliegenden Machbarkeitsstudie zeigte das neuronale Netz eine hohe Sensitivität und eine sehr hohe Spezifität bei der Identifikation der Papilla duodeni major. Die Detektion des Ostiums erfolgte hingegen mit einer deutlich geringeren Sensitivität. Zukünftig ist eine Erweiterung des Trainingsdatensatzes um Videos und klinische Daten vorgesehen, um die Leistungsfähigkeit des Netzwerks zu verbessern. Hierdurch könnte langfristig ein geeignetes Assistenzsystem für die ERCP, insbesondere in der Ausbildungssituation etabliert werden. Y1 - 2025 U6 - https://doi.org/10.1055/s-0045-1806882 VL - 63 IS - 5 SP - e295 PB - Thieme CY - Stuttgart ER - TY - JOUR A1 - Ebigbo, Alanna A1 - Mendel, Robert A1 - Rückert, Tobias A1 - Schuster, Laurin A1 - Probst, Andreas A1 - Manzeneder, Johannes A1 - Prinz, Friederike A1 - Mende, Matthias A1 - Steinbrück, Ingo A1 - Faiss, Siegbert A1 - Rauber, David A1 - Souza Jr., Luis Antonio de A1 - Papa, João Paulo A1 - Deprez, Pierre A1 - Oyama, Tsuneo A1 - Takahashi, Akiko A1 - Seewald, Stefan A1 - Sharma, Prateek A1 - Byrne, Michael F. A1 - Palm, Christoph A1 - Messmann, Helmut T1 - Endoscopic prediction of submucosal invasion in Barrett’s cancer with the use of Artificial Intelligence: A pilot Study JF - Endoscopy N2 - Background and aims: The accurate differentiation between T1a and T1b Barrett’s cancer has both therapeutic and prognostic implications but is challenging even for experienced physicians. We trained an Artificial Intelligence (AI) system on the basis of deep artificial neural networks (deep learning) to differentiate between T1a and T1b Barrett’s cancer white-light images. Methods: Endoscopic images from three tertiary care centres in Germany were collected retrospectively. A deep learning system was trained and tested using the principles of cross-validation. A total of 230 white-light endoscopic images (108 T1a and 122 T1b) was evaluated with the AI-system. For comparison, the images were also classified by experts specialized in endoscopic diagnosis and treatment of Barrett’s cancer. Results: The sensitivity, specificity, F1 and accuracy of the AI-system in the differentiation between T1a and T1b cancer lesions was 0.77, 0.64, 0.73 and 0.71, respectively. There was no statistically significant difference between the performance of the AI-system and that of human experts with sensitivity, specificity, F1 and accuracy of 0.63, 0.78, 0.67 and 0.70 respectively. Conclusion: This pilot study demonstrates the first multicenter application of an AI-based system in the prediction of submucosal invasion in endoscopic images of Barrett’s cancer. AI scored equal to international experts in the field, but more work is necessary to improve the system and apply it to video sequences and in a real-life setting. Nevertheless, the correct prediction of submucosal invasion in Barret´s cancer remains challenging for both experts and AI. KW - Maschinelles Lernen KW - Neuronales Netz KW - Speiseröhrenkrebs KW - Diagnose KW - Artificial Intelligence KW - Machine learning KW - Adenocarcinoma KW - Barrett’s cancer KW - submucosal invasion Y1 - 2021 U6 - https://doi.org/10.1055/a-1311-8570 VL - 53 IS - 09 SP - 878 EP - 883 PB - Thieme CY - Stuttgart ER - TY - JOUR A1 - Mendel, Robert A1 - Rauber, David A1 - Souza Jr., Luis Antonio de A1 - Papa, João Paulo A1 - Palm, Christoph T1 - Error-Correcting Mean-Teacher: Corrections instead of consistency-targets applied to semi-supervised medical image segmentation JF - Computers in Biology and Medicine N2 - 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. KW - Semi-supervised Segmentation KW - Mean-Teacher KW - Pseudo-labels KW - Medical Imaging Y1 - 2023 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:898-opus4-57790 SN - 0010-4825 N1 - Corresponding author der OTH Regensburg: Robert Mendel VL - 154 IS - March PB - Elsevier ER - TY - CHAP A1 - Mendel, Robert A1 - Souza Jr., Luis Antonio de A1 - Rauber, David A1 - Papa, João Paulo A1 - Palm, Christoph T1 - Semi-supervised Segmentation Based on Error-Correcting Supervision T2 - Computer vision - ECCV 2020: 16th European conference, Glasgow, UK, August 23-28, 2020, Proceedings, Part XXIX N2 - Pixel-level classification is an essential part of computer vision. For learning from labeled data, many powerful deep learning models have been developed recently. In this work, we augment such supervised segmentation models by allowing them to learn from unlabeled data. Our semi-supervised approach, termed Error-Correcting Supervision, leverages a collaborative strategy. Apart from the supervised training on the labeled data, the segmentation network is judged by an additional network. The secondary correction network learns on the labeled data to optimally spot correct predictions, as well as to amend incorrect ones. As auxiliary regularization term, the corrector directly influences the supervised training of the segmentation network. On unlabeled data, the output of the correction network is essential to create a proxy for the unknown truth. The corrector’s output is combined with the segmentation network’s prediction to form the new target. We propose a loss function that incorporates both the pseudo-labels as well as the predictive certainty of the correction network. Our approach can easily be added to supervised segmentation models. We show consistent improvements over a supervised baseline on experiments on both the Pascal VOC 2012 and the Cityscapes datasets with varying amounts of labeled data. KW - Semi-Supervised Learning KW - Machine Learning Y1 - 2020 SN - 978-3-030-58525-9 U6 - https://doi.org/10.1007/978-3-030-58526-6_9 SP - 141 EP - 157 PB - Springer CY - Cham ER - TY - JOUR A1 - Römmele, Christoph A1 - Mendel, Robert A1 - Barrett, Caroline A1 - Kiesl, Hans A1 - Rauber, David A1 - Rückert, Tobias A1 - Kraus, Lisa A1 - Heinkele, Jakob A1 - Dhillon, Christine A1 - Grosser, Bianca A1 - Prinz, Friederike A1 - Wanzl, Julia A1 - Fleischmann, Carola A1 - Nagl, Sandra A1 - Schnoy, Elisabeth A1 - Schlottmann, Jakob A1 - Dellon, Evan S. A1 - Messmann, Helmut A1 - Palm, Christoph A1 - Ebigbo, Alanna T1 - An artificial intelligence algorithm is highly accurate for detecting endoscopic features of eosinophilic esophagitis JF - Scientific Reports N2 - The endoscopic features associated with eosinophilic esophagitis (EoE) may be missed during routine endoscopy. We aimed to develop and evaluate an Artificial Intelligence (AI) algorithm for detecting and quantifying the endoscopic features of EoE in white light images, supplemented by the EoE Endoscopic Reference Score (EREFS). An AI algorithm (AI-EoE) was constructed and trained to differentiate between EoE and normal esophagus using endoscopic white light images extracted from the database of the University Hospital Augsburg. In addition to binary classification, a second algorithm was trained with specific auxiliary branches for each EREFS feature (AI-EoE-EREFS). The AI algorithms were evaluated on an external data set from the University of North Carolina, Chapel Hill (UNC), and compared with the performance of human endoscopists with varying levels of experience. The overall sensitivity, specificity, and accuracy of AI-EoE were 0.93 for all measures, while the AUC was 0.986. With additional auxiliary branches for the EREFS categories, the AI algorithm (AI-EoEEREFS) performance improved to 0.96, 0.94, 0.95, and 0.992 for sensitivity, specificity, accuracy, and AUC, respectively. AI-EoE and AI-EoE-EREFS performed significantly better than endoscopy beginners and senior fellows on the same set of images. An AI algorithm can be trained to detect and quantify endoscopic features of EoE with excellent performance scores. The addition of the EREFS criteria improved the performance of the AI algorithm, which performed significantly better than endoscopists with a lower or medium experience level. KW - Artificial Intelligence KW - Smart Endoscopy KW - eosinophilic esophagitis Y1 - 2022 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:898-opus4-46928 VL - 12 PB - Nature Portfolio CY - London ER - TY - GEN A1 - Römmele, Christoph A1 - Mendel, Robert A1 - Rauber, David A1 - Rückert, Tobias A1 - Byrne, Michael F. A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Endoscopic Diagnosis of Eosinophilic Esophagitis Using a deep Learning Algorithm T2 - Endoscopy N2 - Aims Eosinophilic esophagitis (EoE) is easily missed during endoscopy, either because physicians are not familiar with its endoscopic features or the morphologic changes are too subtle. In this preliminary paper, we present the first attempt to detect EoE in endoscopic white light (WL) images using a deep learning network (EoE-AI). Methods 401 WL images of eosinophilic esophagitis and 871 WL images of normal esophageal mucosa were evaluated. All images were assessed for the Endoscopic Reference score (EREFS) (edema, rings, exudates, furrows, strictures). Images with strictures were excluded. EoE was defined as the presence of at least 15 eosinophils per high power field on biopsy. A convolutional neural network based on the ResNet architecture with several five-fold cross-validation runs was used. Adding auxiliary EREFS-classification branches to the neural network allowed the inclusion of the scores as optimization criteria during training. EoE-AI was evaluated for sensitivity, specificity, and F1-score. In addition, two human endoscopists evaluated the images. Results EoE-AI showed a mean sensitivity, specificity, and F1 of 0.759, 0.976, and 0.834 respectively, averaged over the five distinct cross-validation runs. With the EREFS-augmented architecture, a mean sensitivity, specificity, and F1-score of 0.848, 0.945, and 0.861 could be demonstrated respectively. In comparison, the two human endoscopists had an average sensitivity, specificity, and F1-score of 0.718, 0.958, and 0.793. Conclusions To the best of our knowledge, this is the first application of deep learning to endoscopic images of EoE which were also assessed after augmentation with the EREFS-score. The next step is the evaluation of EoE-AI using an external dataset. We then plan to assess the EoE-AI tool on endoscopic videos, and also in real-time. This preliminary work is encouraging regarding the ability for AI to enhance physician detection of EoE, and potentially to do a true “optical biopsy” but more work is needed. KW - Eosinophilic Esophagitis KW - Endoscopy KW - Deep Learning Y1 - 2021 U6 - https://doi.org/10.1055/s-0041-1724274 VL - 53 IS - S 01 PB - Georg Thieme Verlag CY - Stuttgart ER - TY - JOUR A1 - Scheppach, Markus W. A1 - Rauber, David A1 - Stallhofer, Johannes A1 - Muzalyova, Anna A1 - Otten, Vera A1 - Manzeneder, Carolin A1 - Schwamberger, Tanja A1 - Wanzl, Julia A1 - Schlottmann, Jakob A1 - Tadic, Vidan A1 - Probst, Andreas A1 - Schnoy, Elisabeth A1 - Römmele, Christoph A1 - Fleischmann, Carola A1 - Meinikheim, Michael A1 - Miller, Silvia A1 - Märkl, Bruno A1 - Stallmach, Andreas A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Detection of duodenal villous atrophy on endoscopic images using a deep learning algorithm JF - Gastrointestinal Endoscopy N2 - 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. KW - celiac disease KW - villous atrophy KW - endoscopy detection KW - artificial intelligence Y1 - 2023 U6 - https://doi.org/10.1016/j.gie.2023.01.006 PB - Elsevier ER - TY - JOUR A1 - Scheppach, Markus W. A1 - Mendel, Robert A1 - Muzalyova, Anna A1 - Rauber, David A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Yip, Hon Chi A1 - Lau, Louis Ho Shing A1 - Gölder, Stefan Karl A1 - Schmidt, Arthur A1 - Kouladouros, Konstantinos A1 - Abdelhafez, Mohamed A1 - Walter, Benjamin M. A1 - Meinikheim, Michael A1 - Chiu, Philip Wai Yan A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Artificial intelligence improves submucosal vessel detection during third space endoscopy JF - Endoscopy N2 - Background and study aims: While artificial intelligence (AI) shows high potential in decision support for diagnostic gastrointestinal endoscopy, its role in therapeutic endoscopy remains unclear. Third space endoscopic procedures pose the risk of intraprocedural bleeding. Therefore, we aimed to develop an AI algorithm for intraprocedural blood vessel detection. Patients and Methods: Using a test dataset with 101 standardized video clips containing 200 predefined submucosal blood vessels, 19 endoscopists were evaluated for the vessel detection rate (VDR) and time (VDT) with and without support of an AI algorithm. Test subjects were grouped according to experience in ESD. Results: With AI support, endoscopists VDR increased from 56.4% [CI 54.1–58.6] to 72.4% [CI 70.3–74.4]. Endoscopists‘ VDT dropped from 6.7sec [CI 6.2-7.1] to 5.2sec [CI 4.8-5.7]. False positive (FP) readings appeared in 4.5% of frames and were marked significantly shorter than true positives (6.0sec [CI 5.28-6.70] vs. 0.7sec [CI 0.55-0.87]). Conclusions: AI improved the vessel detection rate and time of endoscopists during third space endoscopy. While these data need to be corroborated by clinical trials, AI may prove to be an invaluable tool for the improvement of endoscopic interventions. KW - Artificial Intelligence KW - Third Space Endoscopy Y1 - 2025 U6 - https://doi.org/10.1055/a-2534-1164 PB - Thieme CY - Stuttgart ER - TY - INPR A1 - Allan, Max A1 - Kondo, Satoshi A1 - Bodenstedt, Sebastian A1 - Leger, Stefan A1 - Kadkhodamohammadi, Rahim A1 - Luengo, Imanol A1 - Fuentes, Felix A1 - Flouty, Evangello A1 - Mohammed, Ahmed A1 - Pedersen, Marius A1 - Kori, Avinash A1 - Alex, Varghese A1 - Krishnamurthi, Ganapathy A1 - Rauber, David A1 - Mendel, Robert A1 - Palm, Christoph A1 - Bano, Sophia A1 - Saibro, Guinther A1 - Shih, Chi-Sheng A1 - Chiang, Hsun-An A1 - Zhuang, Juntang A1 - Yang, Junlin A1 - Iglovikov, Vladimir A1 - Dobrenkii, Anton A1 - Reddiboina, Madhu A1 - Reddy, Anubhav A1 - Liu, Xingtong A1 - Gao, Cong A1 - Unberath, Mathias A1 - Kim, Myeonghyeon A1 - Kim, Chanho A1 - Kim, Chaewon A1 - Kim, Hyejin A1 - Lee, Gyeongmin A1 - Ullah, Ihsan A1 - Luna, Miguel A1 - Park, Sang Hyun A1 - Azizian, Mahdi A1 - Stoyanov, Danail A1 - Maier-Hein, Lena A1 - Speidel, Stefanie T1 - 2018 Robotic Scene Segmentation Challenge N2 - In 2015 we began a sub-challenge at the EndoVis workshop at MICCAI in Munich using endoscope images of exvivo tissue with automatically generated annotations from robot forward kinematics and instrument CAD models. However, the limited background variation and simple motion rendered the dataset uninformative in learning about which techniques would be suitable for segmentation in real surgery. In 2017, at the same workshop in Quebec we introduced the robotic instrument segmentation dataset with 10 teams participating in the challenge to perform binary, articulating parts and type segmentation of da Vinci instruments. This challenge included realistic instrument motion and more complex porcine tissue as background and was widely addressed with modfications on U-Nets and other popular CNN architectures [1]. In 2018 we added to the complexity by introducing a set of anatomical objects and medical devices to the segmented classes. To avoid over-complicating the challenge, we continued with porcine data which is dramatically simpler than human tissue due to the lack of fatty tissue occluding many organs. KW - Minimally invasive surgery KW - Robotic KW - Minimal-invasive Chirurgie KW - Robotik Y1 - 2020 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:898-opus4-50049 UR - https://arxiv.org/abs/2001.11190 ER - TY - GEN A1 - Scheppach, Markus W. A1 - Rauber, David A1 - Mendel, Robert A1 - Palm, Christoph A1 - Byrne, Michael F. A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Detection Of Celiac Disease Using A Deep Learning Algorithm T2 - Endoscopy N2 - Aims Celiac disease (CD) is a complex condition caused by an autoimmune reaction to ingested gluten. Due to its polymorphic manifestation and subtle endoscopic presentation, the diagnosis is difficult and thus the disorder is underreported. We aimed to use deep learning to identify celiac disease on endoscopic images of the small bowel. Methods Patients with small intestinal histology compatible with CD (MARSH classification I-III) were extracted retrospectively from the database of Augsburg University hospital. They were compared to patients with no clinical signs of CD and histologically normal small intestinal mucosa. In a first step MARSH III and normal small intestinal mucosa were differentiated with the help of a deep learning algorithm. For this, the endoscopic white light images were divided into five equal-sized subsets. We avoided splitting the images of one patient into several subsets. A ResNet-50 model was trained with the images from four subsets and then validated with the remaining subset. This process was repeated for each subset, such that each subset was validated once. Sensitivity, specificity, and harmonic mean (F1) of the algorithm were determined. Results The algorithm showed values of 0.83, 0.88, and 0.84 for sensitivity, specificity, and F1, respectively. Further data showing a comparison between the detection rate of the AI model and that of experienced endoscopists will be available at the time of the upcoming conference. Conclusions We present the first clinical report on the use of a deep learning algorithm for the detection of celiac disease using endoscopic images. Further evaluation on an external data set, as well as in the detection of CD in real-time, will follow. However, this work at least suggests that AI can assist endoscopists in the endoscopic diagnosis of CD, and ultimately may be able to do a true optical biopsy in live-time. KW - Celiac Disease KW - Deep Learning Y1 - 2021 U6 - https://doi.org/10.1055/s-0041-1724970 N1 - Digital poster exhibition VL - 53 IS - S 01 PB - Georg Thieme Verlag CY - Stuttgart ER - TY - CHAP A1 - Rauber, David A1 - Mendel, Robert A1 - Scheppach, Markus W. A1 - Ebigbo, Alanna A1 - Messmann, Helmut A1 - Palm, Christoph T1 - Analysis of Celiac Disease with Multimodal Deep Learning T2 - Bildverarbeitung für die Medizin 2022: Proceedings, German Workshop on Medical Image Computing, Heidelberg, June 26-28, 2022 N2 - Celiac disease is an autoimmune disorder caused by gluten that results in an inflammatory response of the small intestine.We investigated whether celiac disease can be detected using endoscopic images through a deep learning approach. The results show that additional clinical parameters can improve the classification accuracy. In this work, we distinguished between healthy tissue and Marsh III, according to the Marsh score system. We first trained a baseline network to classify endoscopic images of the small bowel into these two classes and then augmented the approach with a multimodality component that took the antibody status into account. KW - Deep Learning KW - Endoscopy Y1 - 2022 U6 - https://doi.org/10.1007/978-3-658-36932-3_25 SP - 115 EP - 120 PB - Springer Vieweg CY - Wiesbaden ER - TY - CHAP A1 - Mendel, Robert A1 - Rauber, David A1 - Palm, Christoph T1 - Exploring the Effects of Contrastive Learning on Homogeneous Medical Image Data T2 - Bildverarbeitung für die Medizin 2023: Proceedings, German Workshop on Medical Image Computing, July 2– 4, 2023, Braunschweig N2 - 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. Y1 - 2023 U6 - https://doi.org/10.1007/978-3-658-41657-7 SP - 128 EP - 13 PB - Springer Vieweg CY - Wiesbaden ER - TY - GEN A1 - Mendel, Robert A1 - Souza Jr., Luis Antonio de A1 - Rauber, David A1 - Papa, João Paulo A1 - Palm, Christoph T1 - Abstract: Semi-supervised Segmentation Based on Error-correcting Supervision T2 - Bildverarbeitung für die Medizin 2021. Proceedings, German Workshop on Medical Image Computing, Regensburg, March 7-9, 2021 N2 - Pixel-level classification is an essential part of computer vision. For learning from labeled data, many powerful deep learning models have been developed recently. In this work, we augment such supervised segmentation models by allowing them to learn from unlabeled data. Our semi-supervised approach, termed Error-Correcting Supervision, leverages a collaborative strategy. Apart from the supervised training on the labeled data, the segmentation network is judged by an additional network. KW - Deep Learning Y1 - 2021 SN - 978-3-658-33197-9 U6 - https://doi.org/10.1007/978-3-658-33198-6_43 SP - 178 PB - Springer Vieweg CY - Wiesbaden ER - TY - GEN A1 - Rückert, Tobias A1 - Rieder, Maximilian A1 - Rauber, David A1 - Xiao, Michel A1 - Humolli, Eg A1 - Feussner, Hubertus A1 - Wilhelm, Dirk A1 - Palm, Christoph T1 - Augmenting instrument segmentation in video sequences of minimally invasive surgery by synthetic smoky frames T2 - International Journal of Computer Assisted Radiology and Surgery KW - Surgical instrument segmentation KW - smoke simulation KW - unpaired image-to-image translation KW - robot-assisted surgery Y1 - 2023 U6 - https://doi.org/10.1007/s11548-023-02878-2 VL - 18 IS - Suppl 1 SP - S54 EP - S56 PB - Springer Nature ER - TY - JOUR A1 - Maerkl, Raphaela A1 - Rueckert, Tobias A1 - Rauber, David A1 - Gutbrod, Max A1 - Weber Nunes, Danilo A1 - Palm, Christoph T1 - Enhancing generalization in zero-shot multi-label endoscopic instrument classification JF - International Journal of Computer Assisted Radiology and Surgery N2 - Purpose Recognizing previously unseen classes with neural networks is a significant challenge due to their limited generalization capabilities. This issue is particularly critical in safety-critical domains such as medical applications, where accurate classification is essential for reliability and patient safety. Zero-shot learning methods address this challenge by utilizing additional semantic data, with their performance relying heavily on the quality of the generated embeddings. Methods This work investigates the use of full descriptive sentences, generated by a Sentence-BERT model, as class representations, compared to simpler category-based word embeddings derived from a BERT model. Additionally, the impact of z-score normalization as a post-processing step on these embeddings is explored. The proposed approach is evaluated on a multi-label generalized zero-shot learning task, focusing on the recognition of surgical instruments in endoscopic images from minimally invasive cholecystectomies. Results The results demonstrate that combining sentence embeddings and z-score normalization significantly improves model performance. For unseen classes, the AUROC improves from 43.9% to 64.9%, and the multi-label accuracy from 26.1% to 79.5%. Overall performance measured across both seen and unseen classes improves from 49.3% to 64.9% in AUROC and from 37.3% to 65.1% in multi-label accuracy, highlighting the effectiveness of our approach. Conclusion These findings demonstrate that sentence embeddings and z-score normalization can substantially enhance the generalization performance of zero-shot learning models. However, as the study is based on a single dataset, future work should validate the method across diverse datasets and application domains to establish its robustness and broader applicability. KW - Generalized zero-shot learning KW - Sentence embeddings KW - Z-score normalization KW - Multi-label classification KW - Surgical instruments Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:898-opus4-85674 N1 - Corresponding author der OTH Regensburg: Raphaela Maerkl VL - 20 SP - 1577 EP - 1587 PB - Springer Nature ER - TY - GEN A1 - Scheppach, Markus W. A1 - Weber Nunes, Danilo A1 - Rauber, David A1 - Arizi, X. A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Ebigbo, Alanna A1 - Palm, Christoph A1 - Messmann, Helmut T1 - Künstliche Intelligenz-basierte Erkennung von interventionellen Phasen bei der endoskopischen Submukosadissektion T2 - Zeitschrift für Gastroenterologie N2 - Einleitung: Die endoskopische Submukosadissektion (ESD) ist ein komplexes endoskopisches Verfahren, das technische Expertise erfordert. Objektive Methoden zur Analyse von interventionellen Abläufen bei ESD könnten für Qualitätssicherung und Ausbildung, wie auch eine automatische Befunderstellung von Nutzen sein. Ziele: In dieser Studie wurde ein KI-Algorithmus für die Erkennung und Klassifizierung der interventionellen Phasen der ESD entwickelt, um die technische Basis für eine standardisierte Leistungsbewertung und automatische Befunderstellung zu schaffen. Methodik: Vollständige ESD-Videoaufnahmen von 49 Patienten wurden retrospektiv zusammengestellt. Der Datensatz umfasste 6.390.151 Einzelbilder, die alle für die folgenden interventionellen Phasen annotiert wurden: Diagnostik, Markierung, Injektion, Dissektion und Hämostase. 3.973.712 Bilder (28 Patienten) wurden für das Training eines Video-Swin-Transformers genutzt. Dabei wurde temporale Information durch standardisierte BIldextraktion in festgelegten zeitlichen Abständen zum analysierten Bild inkorporiert. 2.416.439 separate Bilder (21 Patienten) wurden für eine interne Validierung genutzt. Ergebnis: Bei der internen Evaluation erreichte das System insgesamt einen F1-Wert von 0,88. Es wurden F1-Werte von 0,99, 0,89, 0,89, 0,91 und 0,52 für Diagnostik, Markierung, Injektion, Dissektion bzw. Blutungsmanagement gemessen. Die Sensitivitäten für dieselben Parameter betrugen 1,00, 0,80, 0,94, 0,89 und 0,67, die Spezifitäten lagen bei 1,00, 1,00, 0,98, 0,88 und 0,93. Positive prädiktive Werte wurden mit 0,98, 1,00, 0,85, 0,94 und 0,43 gemessen. Schlussfolgerung: In dieser vorläufigen Studie zeigte ein KI-Algorithmus eine hohe Leistungsfähigkeit für die Einzelbild-Erkennung von Verfahrensphasen während der ESD. Die vergleichsweise niedrige Leistung für die Blutungsphase wurde auf das seltene Auftreten von Blutungsepisoden im Trainingsdatensatz zurückgeführt, der zu diesem Zeitpunkt nur Videos in voller Länge umfasste. Die zukünftige Entwicklung des Algorithmus wird sich auf die Reduzierung von Klassenungleichgewichten durch selektive Annotationsprotokolle konzentrieren. Y1 - 2025 U6 - https://doi.org/10.1055/s-0045-1811093 VL - 63 IS - 08 SP - e612 EP - e613 PB - Thieme CY - Stuttgart ER - TY - GEN A1 - Scheppach, Markus W. A1 - Rauber, David A1 - Zingler, C. A1 - Weber Nunes, Danilo A1 - Probst, Andreas A1 - Römmele, Christoph A1 - Nagl, Sandra A1 - Ebigbo, Alanna A1 - Palm, Christoph A1 - Messmann, Helmut T1 - Instrumentenerkennung während der endoskopischen Submukosadissektion mittels künstlicher Intelligenz T2 - Zeitschrift für Gastroenterologie N2 - Einleitung: Die endoskopische Submukosadissektion (ESD) ist eine komplexe Technik zur Resektion gastrointestinaler Frühneoplasien. Dabei werden für die verschiedenen Schritte der Intervention spezifische endoskopische Instrumente verwendet. Die präzise und automatische Erkennung und Abgrenzung der verwendeten Instrumente (Injektionsnadeln, elektrochirurgische Messer mit unterschiedlichen Konfigurationen, hämostatische Zangen) könnte wertvolle Informationen über den Fortschritt und die Verfahrensmerkmale der ESD liefern und eine automatische standardisierte Berichterstattung ermöglichen. Ziele: Ziel dieser Studie war die Entwicklung eines KI-Algorithmus zur Erkennung und Delineation von endoskopischen Instrumenten bei der ESD. Methodik: 17 ESD-Videos (9×rektal, 5×ösophageal, 3×gastrisch) wurden retrospektiv zusammengestellt. Auf 8530 Einzelbilder dieser Videos wurden durch 2 Studienmitarbeiter die folgenden Klassen eingezeichnet: Hakenmesser – Spitze, Hakenmesser – Katheter, Nadelmesser – Spitze und – Katheter, Injektionsnadel -Spitze und – Katheter sowie hämostatische Zange – Spitze und – Katheter. Der annotierte Datensatz wurde zum Training eines DeepLabV3+-Deep-Learning-Algorithmus mit ConvNeXt-Backbone zur Erkennung und Abgrenzung der genannten Klassen verwendet. Die Evaluation erfolgte durch 5-fache interne Kreuzvalidierung. Ergebnis: Die Validierung auf Einzelpixelbasis ergab insgesamt einen F1-Score von 0,80, eine Sensitivität von 0,81 und eine Spezifität von 1,00. Es wurden F1-Scores von 1,00, 0,97, 0,80, 0,98, 0,85, 0,97, 0,80, 0,51 bzw. 0,85 für die Klassen Hakenmesser – Katheter und – Spitze, Nadelmesser – Katheter und – Spitze, Injektionsnadel – Katheter und – Spitze, hämostatische Zange – Katheter und – Spitze gemessen. Schlussfolgerung: In dieser Studie wurden die wichtigsten endoskopischen Instrumente, die während der ESD verwendet werden, mit hoher Genauigkeit erkannt. Die geringere Leistung bei der hämostatische Zange – Katheter kann auf die Unterrepräsentation dieser Klassen in den Trainingsdaten zurückgeführt werden. Zukünftige Studien werden sich auf die Erweiterung der Instrumentenklassen sowie auf die Ausbalancierung der Trainingsdaten konzentrieren. Y1 - 2025 U6 - https://doi.org/10.1055/s-0045-1811092 VL - 63 IS - 8 PB - Thieme ER - TY - CHAP A1 - Klausmann, Leonard A1 - Rueckert, Tobias A1 - Rauber, David A1 - Maerkl, Raphaela A1 - Yildiran, Suemeyye R. A1 - Gutbrod, Max A1 - Palm, Christoph T1 - DIY challenge blueprint: from organization to technical realization in biomedical image analysis T2 - Medical Image Computing and Computer Assisted Intervention - MICCAI 2025 ; Proceedings Part XI N2 - Biomedical image analysis challenges have become the de facto standard for publishing new datasets and benchmarking different state-of-the-art algorithms. Most challenges use commercial cloud-based platforms, which can limit custom options and involve disadvantages such as reduced data control and increased costs for extended functionalities. In contrast, Do-It-Yourself (DIY) approaches have the capability to emphasize reliability, compliance, and custom features, providing a solid basis for low-cost, custom designs in self-hosted systems. Our approach emphasizes cost efficiency, improved data sovereignty, and strong compliance with regulatory frameworks, such as the GDPR. This paper presents a blueprint for DIY biomedical imaging challenges, designed to provide institutions with greater autonomy over their challenge infrastructure. Our approach comprehensively addresses both organizational and technical dimensions, including key user roles, data management strategies, and secure, efficient workflows. Key technical contributions include a modular, containerized infrastructure based on Docker, integration of open-source identity management, and automated solution evaluation workflows. Practical deployment guidelines are provided to facilitate implementation and operational stability. The feasibility and adaptability of the proposed framework are demonstrated through the MICCAI 2024 PhaKIR challenge with multiple international teams submitting and validating their solutions through our self-hosted platform. This work can be used as a baseline for future self-hosted DIY implementations and our results encourage further studies in the area of biomedical image analysis challenges. KW - Biomedical challenges KW - Image analysis KW - Blueprint KW - Do-It-Yourself KW - Self-hosting Y1 - 2025 SN - 978-3-032-05141-7 U6 - https://doi.org/10.1007/978-3-032-05141-7_9 SP - 85 EP - 95 PB - Springer CY - Cham ER - TY - INPR A1 - Gutbrod, Max A1 - Rauber, David A1 - Weber Nunes, Danilo A1 - Palm, Christoph T1 - OpenMIBOOD: Open Medical Imaging Benchmarks for Out-Of-Distribution Detection N2 - The growing reliance on Artificial Intelligence (AI) in critical domains such as healthcare demands robust mechanisms to ensure the trustworthiness of these systems, especially when faced with unexpected or anomalous inputs. This paper introduces the Open Medical Imaging Benchmarks for Out-Of-Distribution Detection (OpenMIBOOD), a comprehensive framework for evaluating out-of-distribution (OOD) detection methods specifically in medical imaging contexts. OpenMIBOOD includes three benchmarks from diverse medical domains, encompassing 14 datasets divided into covariate-shifted in-distribution, near-OOD, and far-OOD categories. We evaluate 24 post-hoc methods across these benchmarks, providing a standardized reference to advance the development and fair comparison of OOD detection methods. Results reveal that findings from broad-scale OOD benchmarks in natural image domains do not translate to medical applications, underscoring the critical need for such benchmarks in the medical field. By mitigating the risk of exposing AI models to inputs outside their training distribution, OpenMIBOOD aims to support the advancement of reliable and trustworthy AI systems in healthcare. The repository is available at this https URL. Y1 - 2025 U6 - https://doi.org/10.48550/arXiv.2503.16247 N1 - Der Aufsatz wurde peer-reviewed veröffentlicht und ist ebenfalls in diesem Repositorium verzeichnet unter: https://opus4.kobv.de/opus4-oth-regensburg/8467 ER - TY - GEN A1 - Scheppach, Markus W. A1 - Mendel, Robert A1 - Probst, Andreas A1 - Rauber, David A1 - Rückert, Tobias A1 - Meinikheim, Michael A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Real-time detection and delineation of tissue during third-space endoscopy using artificial intelligence (AI) T2 - Endoscopy N2 - Aims  AI has proven great potential in assisting endoscopists in diagnostics, however its role in therapeutic endoscopy remains unclear. Endoscopic submucosal dissection (ESD) is a technically demanding intervention with a slow learning curve and relevant risks like bleeding and perforation. Therefore, we aimed to develop an algorithm for the real-time detection and delineation of relevant structures during third-space endoscopy. Methods  5470 still images from 59 full length videos (47 ESD, 12 POEM) were annotated. 179681 additional unlabeled images were added to the training dataset. Consequently, a DeepLabv3+ neural network architecture was trained with the ECMT semi-supervised algorithm (under review elsewhere). Evaluation of vessel detection was performed on a dataset of 101 standardized video clips from 15 separate third-space endoscopy videos with 200 predefined blood vessels. Results  Internal validation yielded an overall mean Dice score of 85% (68% for blood vessels, 86% for submucosal layer, 88% for muscle layer). On the video test data, the overall vessel detection rate (VDR) was 94% (96% for ESD, 74% for POEM). The median overall vessel detection time (VDT) was 0.32 sec (0.3 sec for ESD, 0.62 sec for POEM). Conclusions  Evaluation of the developed algorithm on a video test dataset showed high VDR and quick VDT, especially for ESD. Further research will focus on a possible clinical benefit of the AI application for VDR and VDT during third-space endoscopy. KW - Speiseröhrenkrankheit KW - Künstliche Intelligenz KW - Artificial Intelligence Y1 - 2023 U6 - https://doi.org/10.1055/s-0043-1765128 VL - 55 IS - S02 SP - S53 EP - S54 PB - Thieme ER - TY - JOUR A1 - Souza Jr., Luis Antonio de A1 - Passos, Leandro A. A1 - Santana, Marcos Cleison S. A1 - Mendel, Robert A1 - Rauber, David A1 - Ebigbo, Alanna A1 - Probst, Andreas A1 - Messmann, Helmut A1 - Papa, João Paulo A1 - Palm, Christoph T1 - Layer-selective deep representation to improve esophageal cancer classification JF - Medical & Biological Engineering & Computing N2 - Even though artificial intelligence and machine learning have demonstrated remarkable performances in medical image computing, their accountability and transparency level must be improved to transfer this success into clinical practice. The reliability of machine learning decisions must be explained and interpreted, especially for supporting the medical diagnosis.For this task, the deep learning techniques’ black-box nature must somehow be lightened up to clarify its promising results. Hence, we aim to investigate the impact of the ResNet-50 deep convolutional design for Barrett’s esophagus and adenocarcinoma classification. For such a task, and aiming at proposing a two-step learning technique, the output of each convolutional layer that composes the ResNet-50 architecture was trained and classified for further definition of layers that would provide more impact in the architecture. We showed that local information and high-dimensional features are essential to improve the classification for our task. Besides, we observed a significant improvement when the most discriminative layers expressed more impact in the training and classification of ResNet-50 for Barrett’s esophagus and adenocarcinoma classification, demonstrating that both human knowledge and computational processing may influence the correct learning of such a problem. KW - Multistep training KW - Barrett’s esophagus detection KW - Convolutional neural networks KW - Deep learning Y1 - 2024 U6 - https://doi.org/10.1007/s11517-024-03142-8 VL - 62 SP - 3355 EP - 3372 PB - Springer Nature CY - Heidelberg ER - TY - CHAP A1 - Weber Nunes, Danilo A1 - Rauber, David A1 - Palm, Christoph ED - Palm, Christoph ED - Breininger, Katharina ED - Deserno, Thomas M. ED - Handels, Heinz ED - Maier, Andreas ED - Maier-Hein, Klaus H. ED - Tolxdorff, Thomas T1 - Self-supervised 3D Vision Transformer Pre-training for Robust Brain Tumor Classification T2 - Bildverarbeitung für die Medizin 2025: Proceedings, German Conference on Medical Image Computing, Regensburg March 09-11, 2025 N2 - Brain tumors pose significant challenges in neurology, making precise classification crucial for prognosis and treatment planning. This work investigates the effectiveness of a self-supervised learning approach–masked autoencoding (MAE)–to pre-train a vision transformer (ViT) model for brain tumor classification. Our method uses non-domain specific data, leveraging the ADNI and OASIS-3 MRI datasets, which primarily focus on degenerative diseases, for pretraining. The model is subsequently fine-tuned and evaluated on the BraTS glioma and meningioma datasets, representing a novel use of these datasets for tumor classification. The pre-trained MAE ViT model achieves an average F1 score of 0.91 in a 5-fold cross-validation setting, outperforming the nnU-Net encoder trained from scratch, particularly under limited data conditions. These findings highlight the potential of self-supervised MAE in enhancing brain tumor classification accuracy, even with restricted labeled data. Y1 - 2025 U6 - https://doi.org/10.1007/978-3-658-47422-5_69 SP - 298 EP - 303 PB - Springer Vieweg CY - Wiesbaden ER - TY - CHAP A1 - Gutbrod, Max A1 - Geisler, Benedikt A1 - Rauber, David A1 - Palm, Christoph ED - Maier, Andreas ED - Deserno, Thomas M. ED - Handels, Heinz ED - Maier-Hein, Klaus H. ED - Palm, Christoph ED - Tolxdorff, Thomas T1 - Data Augmentation for Images of Chronic Foot Wounds T2 - Bildverarbeitung für die Medizin 2024: Proceedings, German Workshop on Medical Image Computing, March 10-12, 2024, Erlangen N2 - Training data for Neural Networks is often scarce in the medical domain, which often results in models that struggle to generalize and consequently showpoor performance on unseen datasets. Generally, adding augmentation methods to the training pipeline considerably enhances a model’s performance. Using the dataset of the Foot Ulcer Segmentation Challenge, we analyze two additional augmentation methods in the domain of chronic foot wounds - local warping of wound edges along with projection and blurring of shapes inside wounds. Our experiments show that improvements in the Dice similarity coefficient and Normalized Surface Distance metrics depend on a sensible selection of those augmentation methods. Y1 - 2024 U6 - https://doi.org/10.1007/978-3-658-44037-4_71 SP - 261 EP - 266 PB - Springer CY - Wiesbaden ER - TY - CHAP A1 - Gutbrod, Max A1 - Rauber, David A1 - Weber Nunes, Danilo A1 - Palm, Christoph T1 - OpenMIBOOD: Open Medical Imaging Benchmarks for Out-Of-Distribution Detection T2 - 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 10.-17. June 2025, Nashville N2 - The growing reliance on Artificial Intelligence (AI) in critical domains such as healthcare demands robust mechanisms to ensure the trustworthiness of these systems, especially when faced with unexpected or anomalous inputs. This paper introduces the Open Medical Imaging Benchmarks for Out-Of-Distribution Detection (OpenMIBOOD), a comprehensive framework for evaluating out-of-distribution (OOD) detection methods specifically in medical imaging contexts. OpenMIBOOD includes three benchmarks from diverse medical domains, encompassing 14 datasets divided into covariate-shifted in-distribution, nearOOD, and far-OOD categories. We evaluate 24 post-hoc methods across these benchmarks, providing a standardized reference to advance the development and fair comparison of OODdetection methods. Results reveal that findings from broad-scale OOD benchmarks in natural image domains do not translate to medical applications, underscoring the critical need for such benchmarks in the medical field. By mitigating the risk of exposing AI models to inputs outside their training distribution, OpenMIBOOD aims to support the advancement of reliable and trustworthy AI systems in healthcare. The repository is available at https://github.com/remic-othr/OpenMIBOOD. KW - Benchmark testing KW - Reliability KW - Trustworthiness KW - out-of-distribution Y1 - 2025 UR - https://openaccess.thecvf.com/content/CVPR2025/html/Gutbrod_OpenMIBOOD_Open_Medical_Imaging_Benchmarks_for_Out-Of-Distribution_Detection_CVPR_2025_paper.html SN - 979-8-3315-4364-8 U6 - https://doi.org/10.1109/CVPR52734.2025.02410 N1 - Die Preprint-Version ist ebenfalls in diesem Repositorium verzeichnet unter: https://opus4.kobv.de/opus4-oth-regensburg/8059 SP - 25874 EP - 25886 PB - IEEE ER - TY - INPR A1 - Rückert, Tobias A1 - Rauber, David A1 - Maerkl, Raphaela A1 - Klausmann, Leonard A1 - Yildiran, Suemeyye R. A1 - Gutbrod, Max A1 - Nunes, Danilo Weber A1 - Moreno, Alvaro Fernandez A1 - Luengo, Imanol A1 - Stoyanov, Danail A1 - Toussaint, Nicolas A1 - Cho, Enki A1 - Kim, Hyeon Bae A1 - Choo, Oh Sung A1 - Kim, Ka Young A1 - Kim, Seong Tae A1 - Arantes, Gonçalo A1 - Song, Kehan A1 - Zhu, Jianjun A1 - Xiong, Junchen A1 - Lin, Tingyi A1 - Kikuchi, Shunsuke A1 - Matsuzaki, Hiroki A1 - Kouno, Atsushi A1 - Manesco, João Renato Ribeiro A1 - Papa, João Paulo A1 - Choi, Tae-Min A1 - Jeong, Tae Kyeong A1 - Park, Juyoun A1 - Alabi, Oluwatosin A1 - Wei, Meng A1 - Vercauteren, Tom A1 - Wu, Runzhi A1 - Xu, Mengya A1 - an Wang, A1 - Bai, Long A1 - Ren, Hongliang A1 - Yamlahi, Amine A1 - Hennighausen, Jakob A1 - Maier-Hein, Lena A1 - Kondo, Satoshi A1 - Kasai, Satoshi A1 - Hirasawa, Kousuke A1 - Yang, Shu A1 - Wang, Yihui A1 - Chen, Hao A1 - Rodríguez, Santiago A1 - Aparicio, Nicolás A1 - Manrique, Leonardo A1 - Lyons, Juan Camilo A1 - Hosie, Olivia A1 - Ayobi, Nicolás A1 - Arbeláez, Pablo A1 - Li, Yiping A1 - Khalil, Yasmina Al A1 - Nasirihaghighi, Sahar A1 - Speidel, Stefanie A1 - Rückert, Daniel A1 - Feussner, Hubertus A1 - Wilhelm, Dirk A1 - Palm, Christoph T1 - Comparative validation of surgical phase recognition, instrument keypoint estimation, and instrument instance segmentation in endoscopy: Results of the PhaKIR 2024 challenge N2 - Reliable recognition and localization of surgical instruments in endoscopic video recordings are foundational for a wide range of applications in computer- and robot-assisted minimally invasive surgery (RAMIS), including surgical training, skill assessment, and autonomous assistance. However, robust performance under real-world conditions remains a significant challenge. Incorporating surgical context - such as the current procedural phase - has emerged as a promising strategy to improve robustness and interpretability. To address these challenges, we organized the Surgical Procedure Phase, Keypoint, and Instrument Recognition (PhaKIR) sub-challenge as part of the Endoscopic Vision (EndoVis) challenge at MICCAI 2024. We introduced a novel, multi-center dataset comprising thirteen full-length laparoscopic cholecystectomy videos collected from three distinct medical institutions, with unified annotations for three interrelated tasks: surgical phase recognition, instrument keypoint estimation, and instrument instance segmentation. Unlike existing datasets, ours enables joint investigation of instrument localization and procedural context within the same data while supporting the integration of temporal information across entire procedures. We report results and findings in accordance with the BIAS guidelines for biomedical image analysis challenges. The PhaKIR sub-challenge advances the field by providing a unique benchmark for developing temporally aware, context-driven methods in RAMIS and offers a high-quality resource to support future research in surgical scene understanding. Y1 - 2025 N1 - Der Aufsatz wurde peer-reviewed veröffentlicht und ist ebenfalls in diesem Repositorium verzeichnet unter: https://opus4.kobv.de/opus4-oth-regensburg/frontdoor/index/index/start/0/rows/10/sortfield/score/sortorder/desc/searchtype/simple/query/10.1016%2Fj.media.2026.103945/docId/8846 ER - TY - GEN A1 - Scheppach, Markus W. A1 - Mendel, Robert A1 - Muzalyova, Anna A1 - Rauber, David A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Yip, Hon Chi A1 - Lau, Louis Ho Shing A1 - Gölder, Stefan Karl A1 - Schmidt, Arthur A1 - Kouladouros, Konstantinos A1 - Abdelhafez, Mohamed A1 - Walter, B. A1 - Meinikheim, Michael A1 - Chiu, Philip Wai Yan A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Künstliche Intelligenz erhöht die Gefäßerkennung von Endoskopikern bei third space Endoskopie T2 - Zeitschrift für Gastroenterologie N2 - Einleitung: Künstliche Intelligenz (KI)-Algorithmen unterstützen Endoskopiker bei der Erkennung und Charakterisierung von Kolonpolypen in der klinischen Praxis und führen zu einer Erhöhung der Adenomdetektionsrate. Auch bei therapeutischen Maßnahmen wie der endoskopischen Submukosadissektion (ESD) könne relevante anatomische Strukturen durch KI mit hoher Genauigkeit erkannt und im endoskopischen Bild in Echtzeit markiert werden. Der Effekt einer solchen Applikation auf die Gefäßdetektion von Endoskopikern ist bislang nicht erforscht. Ziele:  In dieser Studie wurde der Effekt eines KI-Algorithmus zur Echtzeit-Gefäßmarkierung bei ESD auf die Gefäßdetektionsrate von Endoskopikern untersucht. Methodik:  59 third space Endoskopievideos wurde aus der Datenbank des Universitätsklinikums Augsburg extrahiert. Auf 5470 Einzelbildern dieser Untersuchungen wurde submukosale Blutgefäße annotiert. Zusammen mit weiteren 179681 unmarkierten Bildern wurde ein DeepLabV3+ neuronales Netzwerk mit einer semi-supervised learning Methode darin trainiert, submukosale Blutgefäße auf dem endoskopischen Bild zu erkennen und in Echtzeit einzuzeichnen. Anhand eines Videotests mit 101 Videoclips und 200 vordefinierten Blutgefäßen wurden 19 Endoskopiker mit und ohne KI Unterstützung getestet. Ergebnis:  Der Algorithmus erkannte in dem Videotest 93.5% der Gefäße in einer Detektionszeit von im Median 0,3 Sekunden. Die Gefäßdetektionsrate von Endoskopikern erhöhte sich durch KI Unterstützung von 56,4% auf 72,4% (p<0.001). Die Gefäßdetektionszeit reduzierte sich durch KI-Unterstützung von 6,7 auf 5.2 Sekunden (p<0.001). Der Algorithmus zeigte eine Rate an falsch positiven Detektionen in 4.5% der Einzelbilder. Falsch positiv erkannte Strukturen wurde kürzer detektiert, als richtig positive (0.7 und 6.0 Sekunden, p<0.001). Schlussfolgerung:  KI Unterstützung führte zu einer erhöhten Gefäßdetektionsrate und schnelleren Gefäßdetektionszeit von Endoskopikern. Ein möglicher klinischer Effekt auf die intraprozedurale Komplikationsrate oder Operationszeit könnte in prospektiven Studien ermittelt werden. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1790087 VL - 62 IS - 09 SP - e830 PB - Georg Thieme Verlag KG ER - TY - GEN A1 - Scheppach, Markus W. A1 - Nunes, Danilo Weber A1 - Arizi, X. A1 - Rauber, David A1 - Probst, Andreas A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Intraoperative Phasenerkennung bei endoskopischer Submukosadissektion mit Hilfe von künstlicher Intelligenz T2 - Zeitschrift für Gastroenterologie N2 - Einleitung:  Künstliche Intelligenz (KI) wird in der Endoskopie des Gastrointestinaltraktes zur Erkennung und Charakterisierung von Kolonpolypen eingesetzt. Die Rolle von KI bei therapeutischen Maßnahmen wurde noch nicht eingehend untersucht. Eine intraprozedurale Phasenerkennung bei endoskopischer Submukoasdissektion (ESD) könnte die Erhebung von Qualitätsindikatoren ermöglichen. Weiterhin könnte diese Technologie zu einem tieferen Verständnis über die Eigenschaften der Prozedur führen und weiterführende Applikationen zur automatischen Dokumentation oder standardisiertem Training vorbereiten. Ziele: Ziel dieser Studie war die Entwicklung eines KI Algorithmus zur intraprozeduralen Phasenerkennung bei endoskopischer Submukosadissektion. Methodik:  2071546 Einzelbilder aus 27 ESD Videos in voller Länge wurden für die übergeordneten Klassen Diagnostik, Markierung, Nadelinjektion, Dissektion und Blutung, sowie die untergeordneten Klassen Endoskop-Manipulation, Injektion und Applikation von elektrischem Strom annotiert. Mit einem Trainingsdatensatz (898440 Einzelbilder, 17 ESDs) wurde ein Video Swin Transformer mit uniformer Stichprobenentnahme trainiert und intern validiert (769523 Einzelbilder, 6 ESDs). Neben der internen Validierung wurde der Algorithmus anhand von einem separaten Testdatensatz (403583 Einzelbilder, 4 ESDs) evaluiert. Ergebnis:  Der F1 Score des Algorithmus für alle Klassen lag in der internen Validierung bei 83%, in dem separaten Test bei 90%. Anhand des separaten Tests wurden true positive (TP)-Raten für Diagnostik, Markierung, Nadelinjektion, Dissektion und Blutung von 100%, 100%, 96%, 97% und 93% ermittelt. Für Endoskopmanipulation, Injektion und Applikation von Elektrizität lagen die TP-Raten bei 92%, 98% und 91%. Schlussfolgerung:  Der entwickelte Algorithmus klassifizierte ESD Videos in voller Länge und anhand jedes einzelnen Bildes mit hoher Genauigkeit. Zukünftige Forschungsvorhaben könnten intraoperative Qualitätsindikatioren auf Basis dieser Informationen entwickeln und eine automatisierte Dokumentation ermöglichen. Y1 - 2024 U6 - https://doi.org/10.1055/s-0044-1790084 VL - 62 IS - 09 SP - e828 PB - Georg Thieme Verlag KG ER - TY - GEN A1 - Zellmer, Stephan A1 - Rauber, David A1 - Probst, Andreas A1 - Weber, Tobias A1 - Nagl, Sandra A1 - Römmele, Christoph A1 - Schnoy, Elisabeth A1 - Palm, Christoph A1 - Messmann, Helmut A1 - Ebigbo, Alanna T1 - Verwendung künstlicher Intelligenz bei der Detektion der Papilla duodeni major T2 - Zeitschrift für Gastroenterologie N2 - Einleitung Die Endoskopische Retrograde Cholangiopankreatikographie (ERCP) ist der Goldstandard in der Diagnostik und Therapie von Erkrankungen des pankreatobiliären Trakts. Jedoch ist sie technisch sehr anspruchsvoll und weist eine vergleichsweise hohe Komplikationsrate auf. Ziele  In der vorliegenden Machbarkeitsstudie soll geprüft werden, ob mithilfe eines Deep-learning-Algorithmus die Papille und das Ostium zuverlässig detektiert werden können und somit für Endoskopiker mit geringer Erfahrung ein geeignetes Hilfsmittel, insbesondere für die Ausbildungssituation, darstellen könnten. Methodik Wir betrachteten insgesamt 606 Bilddatensätze 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 für die Klasse Papille eine F1-Wert von 0,7908, eine Sensitivität von 0,7943 und eine Spezifität von 0,9785 erreicht werden, für die Klasse Ostium eine F1-Wert von 0,5538, eine Sensitivität von 0,5094 und eine Spezifität von 0,9970 (vgl. [Tab. 1]). Unabhängig von der Klasse zeigte sich gemittelt (Klasse Papille und Klasse Ostium) ein F1-Wert von 0,6673, eine Sensitivität von 0,6519 und eine Spezifität von 0,9877 (vgl. [Tab. 2]). Schlussfolgerung  In vorliegende Machbarkeitsstudie konnte das neuronale Netz die Papilla duodeni major mit einer hohen Sensitivität und sehr hohen Spezifität identifizieren. Bei der Detektion des Ostiums war die Sensitivität deutlich geringer. Zukünftig soll das das neuronale Netz mit mehr Daten trainiert werden. Außerdem ist geplant, den Algorithmus auch auf Videos anzuwenden. Somit könnte langfristig ein geeignetes Hilfsmittel für die ERCP etabliert werden. KW - Künstliche Intelligenz Y1 - 2023 UR - https://www.thieme-connect.de/products/ejournals/abstract/10.1055/s-0043-1772000 U6 - https://doi.org/10.1055/s-0043-1772000 VL - 61 IS - 08 SP - e539 EP - e540 PB - Thieme CY - Stuttgart ER - TY - JOUR A1 - Rueckert, Tobias A1 - Rauber, David A1 - Maerkl, Raphaela A1 - Klausmann, Leonard A1 - Yildiran, Suemeyye R. A1 - Gutbrod, Max A1 - Nunes, Danilo Weber A1 - Moreno, Alvaro Fernandez A1 - Luengo, Imanol A1 - Stoyanov, Danail A1 - Toussaint, Nicolas A1 - Cho, Enki A1 - Kim, Hyeon Bae A1 - Choo, Oh Sung A1 - Kim, Ka Young A1 - Kim, Seong Tae A1 - Arantes, Gonçalo A1 - Song, Kehan A1 - Zhu, Jianjun A1 - Xiong, Junchen A1 - Lin, Tingyi A1 - Kikuchi, Shunsuke A1 - Matsuzaki, Hiroki A1 - Kouno, Atsushi A1 - Manesco, João Renato Ribeiro A1 - Papa, João Paulo A1 - Choi, Tae-Min A1 - Jeong, Tae Kyeong A1 - Park, Juyoun A1 - Alabi, Oluwatosin A1 - Wei, Meng A1 - Vercauteren, Tom A1 - Wu, Runzhi A1 - Xu, Mengya A1 - Wang, An A1 - Bai, Long A1 - Ren, Hongliang A1 - Yamlahi, Amine A1 - Hennighausen, Jakob A1 - Maier-Hein, Lena A1 - Kondo, Satoshi A1 - Kasai, Satoshi A1 - Hirasawa, Kousuke A1 - Yang, Shu A1 - Wang, Yihui A1 - Chen, Hao A1 - Rodríguez, Santiago A1 - Aparicio, Nicolás A1 - Manrique, Leonardo A1 - Palm, Christoph A1 - Wilhelm, Dirk A1 - Feussner, Hubertus A1 - Rueckert, Daniel A1 - Speidel, Stefanie A1 - Nasirihaghighi, Sahar A1 - Al Khalil, Yasmina A1 - Li, Yiping A1 - Arbeláez, Pablo A1 - Ayobi, Nicolás A1 - Hosie, Olivia A1 - Lyons, Juan Camilo T1 - Comparative validation of surgical phase recognition, instrument keypoint estimation, and instrument instance segmentation in endoscopy: Results of the PhaKIR 2024 challenge JF - Medical Image Analysis N2 - Reliable recognition and localization of surgical instruments in endoscopic video recordings are foundational for a wide range of applications in computer- and robot-assisted minimally invasive surgery (RAMIS), including surgical training, skill assessment, and autonomous assistance. However, robust performance under real-world conditions remains a significant challenge. Incorporating surgical context – such as the current procedural phase – has emerged as a promising strategy to improve robustness and interpretability. To address these challenges, we organized the Surgical Procedure Phase, Keypoint, and Instrument Recognition (PhaKIR) sub-challenge as part of the Endoscopic Vision (EndoVis) challenge at MICCAI 2024. We introduced a novel, multi-center dataset comprising thirteen full-length laparoscopic cholecystectomy videos collected from three distinct medical institutions, with unified annotations for three interrelated tasks: surgical phase recognition, instrument keypoint estimation, and instrument instance segmentation. Unlike existing datasets, ours enables joint investigation of instrument localization and procedural context within the same data while supporting the integration of temporal information across entire procedures. We report results and findings in accordance with the BIAS guidelines for biomedical image analysis challenges. The PhaKIR sub-challenge advances the field by providing a unique benchmark for developing temporally aware, context-driven methods in RAMIS and offers a high-quality resource to support future research in surgical scene understanding. KW - Surgical phase recognition KW - Instrument keypoint estimation KW - Instrument instance segmentation KW - Robot-assisted surgery Y1 - 2026 U6 - https://doi.org/10.1016/j.media.2026.103945 SN - 1361-8415 N1 - Corresponding author der OTH Regensburg: Tobias Rueckert Die Preprint-Version ist ebenfalls in diesem Repositorium verzeichnet unter: https://opus4.kobv.de/opus4-oth-regensburg/solrsearch/index/search/start/0/rows/10/sortfield/score/sortorder/desc/searchtype/simple/query/2507.16559 VL - 109 PB - Elsevier ER - TY - GEN A1 - Rueckert, Tobias A1 - Rauber, David A1 - Klausmann, Leonard A1 - Gutbrod, Max A1 - Rueckert, Daniel A1 - Feussner, Hubertus A1 - Wilhelm, Dirk A1 - Palm, Christoph T1 - PhaKIR Dataset - Surgical Procedure Phase, Keypoint, and Instrument Recognition [Data set] N2 - Note: A script for extracting the individual frames from the video files while preserving the challenge-compliant directory structure and frame-to-mask naming conventions is available on GitHub and can be accessed here: https://github.com/remic-othr/PhaKIR_Dataset. The dataset is described in the following publications: Rueckert, Tobias et al.: Comparative validation of surgical phase recognition, instrument keypoint estimation, and instrument instance segmentation in endoscopy: Results of the PhaKIR 2024 challenge. arXiv preprint, https://arxiv.org/abs/2507.16559. 2025. Rueckert, Tobias et al.: Video Dataset for Surgical Phase, Keypoint, and Instrument Recognition in Laparoscopic Surgery (PhaKIR). arXiv preprint, https://arxiv.org/abs/2511.06549. 2025. The proposed dataset was used as the training dataset in the PhaKIR challenge (https://phakir.re-mic.de/) as part of EndoVis-2024 at MICCAI 2024 and consists of eight real-world videos of human cholecystectomies ranging from 23 to 60 minutes in duration. The procedures were performed by experienced physicians, and the videos were recorded in three hospitals. In addition to existing datasets, our annotations provide pixel-wise instance segmentation masks of surgical instruments for a total of 19 categories, coordinates of relevant instrument keypoints (instrument tip(s), shaft-tip transition, shaft), both at an interval of one frame per second, and specifications regarding the intervention phases for a total of eight different phase categories for each individual frame in one dataset and thus comprehensively cover instrument localization and the context of the operation. Furthermore, the provision of the complete video sequences offers the opportunity to include the temporal information regarding the respective tasks and thus further optimize the resulting methods and outcomes. Y1 - 2025 U6 - https://doi.org/10.5281/zenodo.15740620 ER - TY - GEN A1 - Gutbrod, Max A1 - Rauber, David A1 - Weber Nunes, Danilo A1 - Palm, Christoph T1 - A cleaned subset of the first five CATARACTS test videos [Data set] N2 - This dataset is a subset of the original CATARACTS test dataset and is used by the OpenMIBOOD framework to evaluate a specific out-of-distribution setting. When using this dataset, it is mandatory to cite the corresponding publication (OpenMIBOOD (10.1109/CVPR52734.2025.02410)) and follow the acknowledgement and citation requirements of the original dataset (CATARACTS). The original CATARACTS dataset (associated publication,Homepage) consists of 50 videos of cataract surgeries, split into 25 train and 25 test videos. This subset contains the frames of the first 5 test videos. Further, black frames at the beginning of each video were removed. Y1 - 2025 U6 - https://doi.org/10.5281/zenodo.14924735 N1 - Related works: Is derived from: Dataset: 10.21227/ac97-8m18 (DOI) Software: Repository URL: https://github.com/remic-othr/OpenMIBOOD ER - TY - GEN A1 - Gutbrod, Max A1 - Rauber, David A1 - Weber Nunes, Danilo A1 - Palm, Christoph T1 - Cropped single instrument frames subset from Cholec80 [Data set] N2 - This dataset is a subset of the original Cholec80 dataset and is used by the OpenMIBOOD framework to evaluate a specific out-of-distribution setting. When using this dataset, it is mandatory to cite the corresponding publication (OpenMIBOOD) and to follow the acknowledgement and citation requirements of the original dataset (Cholec80). The original Cholec80 dataset (associated paper,Homepage) consists of 80 cholecystectomy surgery videos recorded at 25 fps, performed by 13 surgeons. It includes phase annotations (25 fps) and tool presence labels (1 fps), with phase definitions provided by a senior surgeon. A tool is considered present if at least half of its tip is visible. The dataset categorizes tools into seven types: Grasper, Bipolar, Hook, Scissors, Clipper, Irrigator, and Specimen bag. Multiple tools may be present in each frame. Additionally, 76 of the 80 videos exhibit a strong black vignette. For this dataset subset, frames were extracted based on tool presence labels, selecting only those containing Grasper, Bipolar, Hook, or Clipper while ensuring that only a single tool appears per frame. To enhance visual consistency, the black vignette was removed by extracting an inner rectangular region, where applicable. KW - Tool Presence Detection KW - Cholecystectomy KW - Laparoscopic KW - Deep Learning KW - Out-Of-Distribution Detection Y1 - 2025 U6 - https://doi.org/10.5281/zenodo.14921670 N1 - Related works Is derived from Journal article: 10.1109/TMI.2016.2593957 Software Repository URL https://github.com/remic-othr/OpenMIBOOD ER - TY - GEN A1 - Gutbrod, Max A1 - Rauber, David A1 - Weber Nunes, Danilo A1 - Palm, Christoph T1 - OpenMIBOOD's classification models for the MIDOG, PhaKIR, and OASIS-3 benchmarks [Data set] N2 - These models are provided for evaluating post-hoc out-of-distribution methods on the three OpenMIBOOD benchmarks: MIDOG, PhaKIR, and OASIS-3. When using these models, make sure to give appropriate credit and cite the OpenMIBOOD publication. Y1 - 2025 U6 - https://doi.org/10.5281/zenodo.14982267 N1 - Software Repository URL https://github.com/remic-othr/OpenMIBOOD ER - TY - GEN A1 - Klausmann, Leonard A1 - Rueckert, Tobias A1 - Rauber, David A1 - Maerkl, Raphaela A1 - Yildiran, Suemeyye R. A1 - Gutbrod, Max A1 - Palm, Christoph ED - Handels, Heinz ED - Breininger, Katharina ED - Deserno, Thomas M. ED - Maier, Andreas ED - Maier-Hein, Klaus H. ED - Palm, Christoph ED - Tolxdorff, Thomas T1 - Abstract: DIY Challenge Blueprint BT - from organization to technical implementation in Biomedical Image Analysis T2 - Bildverarbeitung für die Medizin 2025: Proceedings, German Conference on Medical Image Computing, Lübeck March 15-17, 2026 N2 - The high cost of challenge platforms prevents many people from organizing their own competitions. The do-it-yourself (DIY) challenge blueprint [1] allows you to host your own biomedical AI benchmark challenge. Our DIY approach circumvents the current constraints of commercial challenge platforms. A sovereign, extensible and cost-efficient deployment is provided via containerised, identity-managed and reproducible pipelines. Focus lies on GDPR-compliant hosting via infrastructure-as-code, automated evaluation, modular orchestration, and role-based identity and access management. The framework integrates Docker-based execution and standardised interfaces for task definitions, dataset curation and evaluation. All in all it is designed to be flexible and modular, as demonstrated in the MICCAI 2024 PhaKIR challenge [2, 3]. In this case study, different medical tasks on a multicentre laparoscopic dataset with framewise labels for phases and spatial annotations for instruments across fulllength videos were supported. This case study empirically validates the DIY challenge blueprint as a reproducible and customizable challenge-hosting infrastructure. The full code can be found at https://github.com/remic-othr/PhaKIR_DIY. KW - Bildverarbeitung Y1 - 2026 U6 - https://doi.org/10.1007/978-3-658-51100-5_27 SP - 131 EP - 131 PB - Springer Vieweg CY - Wiesbaden ER - TY - CHAP A1 - Gutbrod, Max A1 - Rauber, David A1 - Palm, Christoph ED - Handels, Heinz ED - Breininger, Katharina ED - Deserno, Thomas M. ED - Maier, Andreas ED - Maier-Hein, Klaus H. ED - Palm, Christoph ED - Tolxdorff, Thomas T1 - Improving Generalization in Mitotic Cell Detection via Domain Transformations T2 - Bildverarbeitung für die Medizin 2025: Proceedings, German Conference on Medical Image Computing, Lübeck March 15-17, 2026 N2 - We address domain generalization (DG) in mitotic-cell (MC) detection by combining a β-variational autoencoder (VAE) for domain transformations with feature-space alignment together with an object detector. The β-VAE synthesizes domain-transformed images, and the detector is trained to map originals and their transformed counterparts to equal representations. On the MIDOG++ dataset, this approach improves out-of-domain detection F1 scores by 7 and 3 percentage points compared to the color-variation augmentation and stain-normalization baselines. Results further suggest that morphology shifts hinder generalization more than stain shifts. KW - Künstliche Intelligenz KW - Bildverarbeitung Y1 - 2026 U6 - https://doi.org/10.1007/978-3-658-51100-5_71 SP - 362 EP - 367 PB - Springer Vieweg CY - Wiesbaden ER -