深度学习在胃组织病理诊断中的应用进展

Application progress of deep learning in pathological diagnosis for gastric tissues

深度学习在组织病理中的应用主要包括对组织病理进行诊断和分级、评估生物标志物以及患者预后。为了使深度学习模型更具泛化性和鲁棒性,常通过颜色归一化和数据增强等对图像进行预处理。目前,深度学习在胃组织病理诊断中的研究也在不断探索,主要包括利用深度学习识别癌变区域,或者对胃组织病变进行分类(包括对病变进行良恶性分类的二分类法,或者区分正常黏膜、胃炎或者胃癌等的三分类法),以及判定胃癌是否出现了淋巴结转移。此外,也有少数研究将人工智能应用于辅助病理医师进行胃组织病变的诊断,这不仅提高了病理医师诊断的准确率,而且缩短了诊断时间。

The application of deep learning in tissue pathology mainly includes diagnosis and grading of tissue pathology,assessment of biomarkers,and patient prognosis.In order to make deep learning models more generalizable and robust,image preprocessing techniques such as color normalization and data augmentation are commonly used.Currently,research on deep learning in gastric tissue pathological diagnosis is also constantly being explored.This mainly includes using deep learning to identify cancerous regions,classifying gastric tissue lesions(including binary classification method for benign and malignant lesions,or ternary classification method for distinguishing between normal mucosa,gastritis,or gastric cancer),and determining the presence of lymph node metastasis in gastric cancer.Additionally,there are also a few studies that use artificial intelligence to assist pathologists in diagnosing gastric tissue lesions,which not only improves the accuracy of pathological diagnosis but also shortens the diagnostic time.