Multi-scale Deformable Transformer for the Classification of Gastric Glands: The IMGL Dataset

Gastric cancer is one of the most common cancers and a leading cause of cancer-related death worldwide. Among the risk factors of gastric cancer, the gastric intestinal metaplasia (IM) has been found to increase the risk of gastric cancer and is considered as one of the precancerous lesions. Therefore, early detection of IM could allow risk stratification regarding the possibility of progression to cancer. To this end, accurate classification of gastric glands from the histological images plays an important role in the diagnostic confirmation of IM. To date, although many gland segmentation approaches have been proposed, no general model has been proposed for the identification of IM glands. Thus, in this paper, we propose a model for gastric glands' classification. More specifically, we propose a multi-scale deformable transformer-based network for glands' classification into normal and IM gastric glands. To evaluate the efficiency of the proposed methodology we created the IMGL dataset consisting of 1000 gland images, including both intestinal metaplasia and normal cases received from 20 Whole Slide Images (WSI). The results showed that the proposed approach achieves an F1 score equal to 0.94, showing great potential for the gastric glands' classification.