Artificial intelligence model for detecting duodenal endoscopic changes on images of functional dyspepsia

Background: Recently, it has been suggested that the duodenum may be the pathological locus of functional dyspepsia (FD). Additionally, an image-based artificial intelligence (AI) model was shown to discriminate colonoscopy images of irritable bowel syndrome from healthy subjects with an area under the curve (AUC) 0.95.
Aim: To evaluate an AI model to distinguish duodenal images of FD patients from healthy subjects.
Methods: Duodenal images were collected from hospital records and labeled as \"functional dyspepsia\" or non-FD in electronic medical records. Helicobacter pylori (HP) infection status was obtained from the Japan Endoscopy Database. Google Cloud AutoML Vision was used to classify four groups: FD/HP current infection (n = 32), FD/HP uninfected (n = 35), non-FD/HP current infection (n = 39), and non-FD/HP uninfected (n = 33). Patients with organic diseases (e.g., cancer, ulcer, postoperative abdomen, reflux) and narrow-band or dye-spread images were excluded. Sensitivi..., , , # Data from: Artificial intelligence model for detecting duodenal endoscopic changes on images of functional dyspepsia

[https://doi.org/10.5061/dryad.pvmcvdntc](https://doi.org/10.5061/dryad.pvmcvdntc)

We will post 70 duodenal images from 32 patients who were treated for Functional Dyspepsia (FD) and had Helicobacter pylori infection, and 141 images from 39 asymptomatic individuals with Helicobacter pylori infection. Using the presence or absence of symptoms as training data, we constructed an image AI to distinguish between the two groups.

## Description of the data and file structure

In the FDHP box, there are duodenal images in jpg format of patients who were treated for Functional Dyspepsia (FD) and had Helicobacter pylori infection. In the nonFDHPpositive box, there are duodenal images in jpg format of asymptomatic individuals with Helicobacter pylori infection.
,