Related Experiment Video
Updated: Sep 16, 2025
Author Spotlight: Anterior HR-OCT as a Non-Invasive Tool for Characterizing Ocular Surface Squamous Neoplasia
Published on: August 9, 2024
AI assistance enhances histopathologic distinction between sebaceous and squamous cell carcinoma of the eyelid
Jialu Geng1, Kai Zhang2, Li Dong1
1Beijing Tongren Eye Center, Beijing Key Laboratory of Intraocular Tumor Diagnosis and Treatment, Beijing Ophthalmology＆Visual Sciences Key Lab, Medical Artificial Intelligence Research and Verification Key Laboratory of the Ministry of Industry and Information Technology, Beijing Key Laboratory of Intelligent Diagnosis, Treatment and Prevention of Blinding Eye Diseases, Beijing Tongren Hospital, Capital Medical University, Beijing, China.
Abstract:
Sebaceous gland carcinoma (SGC) and some poorly differentiated squamous cell carcinomas (SC) of the eyelid may have overlapping clinical and histopathologic features, leading to potential misdiagnosis and delayed treatment. The authors developed a deep learning (DL)-based pathological analysis framework to classify SGC and SC automatically. In total, 282 whole slide images (WSIs) were used for training, validating and inner testing the DL framework and 36 WSIs were obtained from another hospital as an external testing dataset. In WSI level, the diagnostic accuracy of SGC and SC achieved 84.85% and 75.12%, respectively, in the internal testing set and reached 22.22% and 77.78%, respectively, in the external testing set. The accuracy of the pathologists could be improved with the AI framework (60.0 ± 9.8% vs. 76.8 ± 9.6%). This AI-based automatic pathological diagnostic framework achieved the performance of a well-experienced pathologist and can assist pathologists in making diagnoses more accurately, especially for non-ophthalmic pathologists.
