Biopsy Sampling in Upper Gastrointestinal Endoscopy: A Survey from 10 Tertiary Referral Centres Across Europe.

BACKGROUND: Guidelines give robust recommendations on which biopsies should be taken when there is endoscopic suggestion of gastric inflammation. Adherence to these guidelines often seems arbitrary. This study aimed to give an overview on current practice in tertiary referral centres across Europe. METHODS: Data were collected at 10 tertiary referral centres. Demographic data, the indication for each procedure, endoscopic findings, and the number and sampling site of biopsies were recorded. Findings were compared between centres, and factors influencing the decision to take biopsies were explored. RESULTS: Biopsies were taken in 56.6% of 9,425 procedures, with significant variation between centres (p < 0.001). Gastric biopsies were taken in 43.8% of all procedures. Sampling location varied with the procedure indication (p < 0.001) without consistent pattern across the centres. Fewer biopsies were taken in centres which routinely applied the updated Sydney classification for gastritis assessment (46.0%), compared to centres where this was done only upon request (75.3%, p < 0.001). This was the same for centres stratifying patients according to the OLGA system (51.8 vs. 73.0%, p < 0.001). More biopsies were taken in centres following the MAPS guidelines on stomach surveillance (68.1 vs. 37.1%, p < 0.001). Biopsy sampling was more likely in younger patients in 8 centres (p < 0.05), but this was not true for the whole cohort (p = 0.537). The percentage of procedures with biopsies correlated directly with additional costs charged in case of biopsies (r = 0.709, p = 0.022). CONCLUSION: Adherence to guideline recommendations for biopsy sampling at gastroscopy was inconsistent across the participating centres. Our data suggest that centre-specific policies are applied instead.

Deep learning for detection and segmentation of artefact and disease instances in gastrointestinal endoscopy.

The Endoscopy Computer Vision Challenge (EndoCV) is a crowd-sourcing initiative to address eminent problems in developing reliable computer aided detection and diagnosis endoscopy systems and suggest a pathway for clinical translation of technologies. Whilst endoscopy is a widely used diagnostic and treatment tool for hollow-organs, there are several core challenges often faced by endoscopists, mainly: 1) presence of multi-class artefacts that hinder their visual interpretation, and 2) difficulty in identifying subtle precancerous precursors and cancer abnormalities. Artefacts often affect the robustness of deep learning methods applied to the gastrointestinal tract organs as they can be confused with tissue of interest. EndoCV2020 challenges are designed to address research questions in these remits. In this paper, we present a summary of methods developed by the top 17 teams and provide an objective comparison of state-of-the-art methods and methods designed by the participants for two sub-challenges: i) artefact detection and segmentation (EAD2020), and ii) disease detection and segmentation (EDD2020). Multi-center, multi-organ, multi-class, and multi-modal clinical endoscopy datasets were compiled for both EAD2020 and EDD2020 sub-challenges. The out-of-sample generalization ability of detection algorithms was also evaluated. Whilst most teams focused on accuracy improvements, only a few methods hold credibility for clinical usability. The best performing teams provided solutions to tackle class imbalance, and variabilities in size, origin, modality and occurrences by exploring data augmentation, data fusion, and optimal class thresholding techniques.