[Application of artificial intelligence in pathological diagnosis of ocular tumors]

Ocular tumors encompass ocular surface tumors, orbital tumors, and intraocular tumors, characterized by high heterogeneity and complex classifications, which pose considerable challenges to pathological diagnosis. The scarcity of specialized ophthalmic pathologists further exacerbates the difficulty in diagnosing these tumors. In recent years, artificial intelligence (AI) has penetrated diverse areas of healthcare, particularly in departments with intensive medical imaging and image-based workflows, such as radiology, pathology, and ultrasonography. With the advent of the digital pathology era, the role of AI in aiding pathological diagnosis has become increasingly prominent. In recent years, it has been tentatively applied to several ocular tumors, including eyelid sebaceous carcinoma, basal cell carcinoma, malignant melanoma, and orbital lymphoproliferative disorders, exerting an auxiliary diagnostic function. AI can also assist in predicting the prognosis of uveal malignant melanoma and retinoblastoma. In the future, with the increasing popularization and refinement of AI-assisted pathological diagnosis technologies, it will not only enhance the diagnostic accuracy and work efficiency of ocular tumors, better guide clinical treatment, but also alleviate the issues of inadequate medical resources and a shortage of specialized ophthalmic pathologists.