László Madácsy

Aims Our aim was to compare the Ankon magnetically controlled capsule endoscopy system (MCCE) and traditional gastroscopy in the evaluation of gastric disorders in outpatients who underwent both examinations on the same day. Methods Between August 2018 and October 2022, consecutive outpatients were enrolled to study who underwent MCCE and was referred to traditional gastroscopy and biopsies due to the found pathologies on capsule endoscopy. UBT test was done just before MCCE. Results Overall 62 outpatients (36 male, 26 female, 45 years old in average) underwent gastroscopy and MCCE on the same day in our Endoscopy Unit. UBT test were positive in 13 patients (21%), H.pylori positivity were found in 16 patients (25.8%) during histopathology. MCCE revealed no pathologies in 11 patients. By MCCE, focal lesions were found in overall 15 patients (1 gastric polyp, 5 erosions, 3 ulcer, 6 foveolar hyperplasia) and gastritis were described in overall 50 patients (39 distal gastritis, 10 pangastritis, 1 other gastritis). By gastroscopy, focal lesions were found in 8 cases (1 polyp, 4 erosions, 3 ulcers, 3 foveolar hyperplasia) and gastritis were found in 61 patients (51 lower, 9 pangastritis, 1 other gastritis). Gastroscopy was negative in 2 patients Regarding focal lesions, the gastric polyp, 3 foveolar hyperplasias and 3 gastric ulcers were visualised by both endoscopy techniques, from the 2 out of the 4 gastric erosions were visualised both techniques. Conclusions MCCE is an effective and safe diagnostic method to evaluate upper GI mucosal lesions. Useful diagnostic method to evaluate gastric mucosa and it can be the future non-invasive screening tool to decrease morbidity and mortality of upper GI disorders. Article published online: 14 April 2023 © 2023. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany

Aims Background: Precise recognition and characterization of polyps is essential in the prevention of CRC. Artificial intelligence deep learning algorithm-based Decision Support System (AI-DSS) was developed to achieve high accuracy in polyp characterization, in advance to assist in everyday practice for non-expert endoscopists. The advanced version of our AI-DSS (Polypbrain) has an automatized lesion segmentation function of the endoscopic images, so the AI-DSS can predict the histology during colonoscopy. Aim: Comparing the differentiation performance and accuracy between adenomas and hyperplastic polyps of our AI-DSS (Polypbrain) to seven colonoscopic experts. Methods AI-DSS was trained on selected pictures of our anonymous electronic database , for which there were histological diagnosis. For this analysis, 596 histologically identified sub-centimetric colorectal polyps and 2309 HD, electronic chromo-endoscopic images were used. The test set contained 211 pictures taken by white light and virtual chromoendoscopy techniques from 123 randomly selected polyps. AI-DSS was classified the polyps to 3 categories (hyperplastic polyp, adenoma and non-polyp) after auto segmentation of the images. We compared the AI-DSS result to the summarized decision of the expert endoscopists ([Fig. 1]). Results Polypbrain AI-DSS segmentation performance: IOU: 86.5%, DSC: 91.8%, Sensitivity: 95.1%, Specificity: 95.8%, Accuracy: 95.5% Conclusions Polypbrain AI-DSS is just as accurate tool for real-time polyp segmentation and histology prediction as high quality expert endoscopists. Article published online: 14 April 2023 © 2023. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany

Aims Recognition of polyps is essential in colonoscopies to prevent CRC formation. Currently it is mainly based only in experts’ decisions. AI-based decision support systems (AI-DSS) were suggested to be helpful, but randomized, prospective, real-time results are still limited. In the present study we aimed to compare the effectiveness an accuracy of real-time AI-DSS and standard high-definition WLI in the detection of polyps during colonoscopies Methods One thousand two hundred and fifty-nine outpatients were consecutively enrolled between July 2021 and May 2022. Patients were randomized before colonoscopy to age and sex matched groups for AI-DSS (Fujifilm CAD EYE) or standard WLI colonoscopy (Fujfilm 6000 and 7000 Eluxeo system with 720 and 760 colonoscopes). The number of polyp negative and positive results were compared as well as the number of discovered polyps. Results There were 644 (M/F 319/325, mean age: 55 years) subjects in the AI-DSS and 615 (M/F 286/329 mean age: 54 years) in the standard WLI group. Significantly higher number of patients [190 (29.5%) vs. 148 (24.1%), p=0.03] had at least one polyp in the AI-DSS group compared to the controls evaluated by WLI only. The mean number of polyps were also higher in AI-DSS group: 0.46±0.89 (mean±SD) vs. controls 0.38±0.80, p=0.04.No significant differences were depicted between the two groups as we compared the colonoscope withdrawal times: 498.2±215.7 sec. vs. 490.9±215,1 sec., respectively. Conclusions Real-time AI-DSS assisted colonoscopy seems to be an effective but not time-consuming tool during the daily clinical practice to improve the detection of colonic polyps. The acceptance of the software could be further increased with the reduction of false positive alarm rate. Article published online: 14 April 2023 © 2023. European Society of Gastrointestinal Endoscopy. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany

First prospective European study for the feasibility and safety of magnetically controlled capsule endoscopy in gastric mucosal abnormalities

Linked color imaging (LCI) is a new endoscopic technology that may increase colorectal adenoma detection rate (ADR) and polyp detection rate (PDR) by virtual chromoendoscopy. Aim of the present stu...

Objectives: To analyze risk factors of post-ERCP pancreatitis (PEP) in high risk patients in whom prophylactic pancreatic stents (PPSs) were intended to use to prevent endoscopic retrograde cholangiography (ERCP)-related complications. Patients and methods: Three hundred and seventeen high-risk patients for developing PEP were considered for preventive pancreatic stent placement in our endoscopy unit over 5 years. 5 Fr, 3–5 cm long PPSs were used. All data were collected in a predefined database. Development of PEP despite PPS placement was analyzed. Results: PEP occurred in 29 of 288 successfully stented patients (10.07%; 24 mild, four moderate, one severe). PPS was protective against all risks factors except for sphincter of Oddi dysfunction (SOD). PPS related complication rate was 2.78% (n = 8). Unsuccessful stenting occurred in 29 patients (9.15%), PEP developed in 41.38% (n = 12; seven mild, four moderate, one severe). Those patients who had more patient related risk factors were more likely to develop PEP despite preventive measures with PPS. On the contrary, PPS placement was less successful in patients who had more procedure related risk factors. Conclusions: PPS is protective against all risks factors of PEP except for SOD in high-risk patients. More vulnerable patients who have more patient-related risk factors are more likely to develop PEP despite PPS is used, while more complex procedures predispose to unsuccessful PPS placement in patients with more procedure-related risk factors. PPS insertion in high-risk patients is effective and safe preventive method and the procedure related complication rate is reasonably low.