Gastrointestinal tissue-samples from population-based endoscopy studies

Illustration (click to hide):

Project Description

The project consists of gastrointestinal tissue-samples from 3 population-based endoscopy studies. In all three studies a random selection from the general population was included. In the Kalixanda upper endoscopy study includes biopsies from 1000 individuals, in LongGERD 388 individuals had an upper endoscopy with biopsies and in the PopCol study colonoscopies was performed on 745 individuals.

During endoscopy, tissue-samples were collected. In upper endoscopies a total of 7 samples were collected from the esophagus (2 locations), the stomach (3 locations) and the duodenum (2 locations). In colonoscopies 5 samples were collected, one from the ileum and 4 locations in the colon/rectum. Hematoxylin-eosin stained slides have been prepared from each biopsy and all slides have been scanned and digitalized as high-resolution ndpi-files. Most endoscopies are without endoscopic or histological findings. Among endoscopic and histological findings are reflux esophagitis, eosinophilic esophagitis, gastric ulcer, duodenal ulcer, atrophic gastritis, Hp-associated gastritis, duodenitis, celiac disease, Crohn’s disease, ulcerative colitis, microscopic colitis, diverticulosis, adenoma, hyperplastic polyps among others.

The subjects have also answered health- and symptom-related questionnaires.
The aim of this project is to structure the digitalized samples, to annotate samples and to automate the process of finding and quantifying tissue abnormalities such as inflammatory cell infiltration or epithelial barrier disruptions and relate the abnormalities to diseases/symptoms. The goal is to be able to use data-driven analyses to identify find previously unknown differences between specified symptom/diseases and tissue abnormalities.