Gastrointestinal stromal tumors (GISTs) are the most common subtype of soft tissue sarcomas, with diagnosis based on a combination of clinical presentation, anatomical localization of the tumor, histological and immunohistochemical (IHC) studies. This paper explores the potential of artificial intelligence (AI) technologies for the automated detection of mitotic figures in histological images of GISTs. A literature review on the application of AI in morphological diagnostics of various tumors is presented. A test dataset of 220 digitized histological slides annotated in CVAT was created as part of the study. A convolutional neural network, YOLOv11, trained over 300 epochs, was used to analyze mitotic activity. Model performance evaluation showed that its accuracy was limited due to the small size of the training dataset. Future work will focus on expanding the dataset and improving the accuracy of pathological mitosis detection.
Musatov et al. (Thu,) studied this question.