GIM-ENDO: A Multimodal Endoscopic Image and Video Dataset for Gastric Intestinal Metaplasia Morphology and Pathology

#GIMENDO Dataset Gastric intestinal metaplasia (GIM) is a precursor lesion to gastric dysplasia and adenocarcinoma, and its early detection plays a critical role in interrupting the gastric carcinogenesis cascade. The dataset includes data from 23 patients (21 GIM-positive and 2 normal controls), acquired using the Olympus EVIS X1 system. Imaging modalities include white-light endoscopy (WLE) and image-enhanced endoscopy (IEE), specifically narrow-band imaging (NBI) and magnifying NBI (M-NBI). The dataset contains both images and video clips, along with demographic data, endoscopic findings, histopathological results, and Helicobacter pylori status. Detailed annotations are provided for six primary IEE endoscopic features: light blue crest (LBC), marginal turbid band (MTB), white opaque substance (WOS), TV Pattern (irregular/fusion), atrophy, and map-like erythema (MLE). In addition, two supplementary findings (AHP and GA) are annotated when present. Histopathological annotations include GIM subtypes (complete and incomplete) for all GIM-positive cases. OLGA and OLGIM staging are also provided where complete histological sampling was available. This dataset is intended to support the development, benchmarking, and validation of AI models for GIM detection, characterization, and multimodal analysis. For the latest updates, documentation, and additional information about this dataset, please visit:https://databiox.com/