He Vision Group Collaborates to Publish the Latest AI Research Results, Offering New Scheme for Early Glaucoma Screening in Grid-Based Communities
As the world’s leading cause of irreversible blindness, early screening and intervention of glaucoma has always been a core challenge in public health. With its insidious onset and hidden disease progression, glaucoma is known as the “the silent thief of sight” because it often damages vision without patient’s awareness, and irreversible and severe damage to the visual function has already occurred when obvious symptoms appear.
Recently, Dr. Dai Guangzheng and his team members Zhang Chenguang, Yan Hairu, and Dang Yu from Visual Intelligence Center of He University collaborated with researchers from He Eye Specialist Hospital, He Vision Industrial Group, and Liaoning DRAGONFLEYE Health Technology to publish a groundbreaking study titled “Intelligent Screening of Narrow Anterior Chamber Angle Based on Portable Slit Lamp” in the international top journal npj Digital Medicine (IF: 15.1).
The study evaluates the feasibility of using portable slit lamp combined with advanced artificial intelligence algorithms to detect narrow anterior chamber angle (NACA) from anterior segment images, providing a new solution for early screening of primary angle-closure glaucoma (PACG).
Technical Breakthrough
AI + Portable Slit Lamp: Making Glaucoma Screening More Efficient and Convenient
Glaucoma is a leading cause of irreversible blindness worldwide. It is categorized into three main types: primary glaucoma, secondary glaucoma and pediatric glaucoma. Primary glaucoma is the most common type, which is subdivided into primary open-angle glaucoma (PAOG) and primary angle-closure glaucoma (PACG). Among them, PACG is the type of glaucoma with an extremely high blindness rate, characterized by shallow anterior chamber and narrow angle.
Statistics show that the prevalence rate of PACG among people aged 40 and above worldwide is approximately 0.6%, with a higher prevalence rate in Asia. In areas such as China and Japan, the prevalence rate ranges from 0.75% to 1.19%, with the proportion of female patients being significantly higher than that of males, and the prevalence increasing with age. Regular screening is crucial for people over 40 years old or high-risk individuals for early detection and prevention of severe visual impairment.
“Screening for glaucoma must be shifted forward. Identifying high-risk groups before the angle of the anterior chamber is completely closed is the critical window to prevent blindness.” emphasized Dr. Dai Guangzheng.
Notably, traditional glaucoma screening has many limitations. Gonioscopy, the gold standard for detecting anterior chamber angle status, is highly invasive and complext to operate, making it only applicable in specialized ophthalmic settings and not for large-scale community screening. Other non-contact detection methods either rely on subjective judgment, are expensive, or mostly require operation by professionals.
To address these issues, the research team conducted an in-depth exploration ranging from the construction of a glaucoma dataset to algorithm development. They used a modified mobile camera to capture anterior segment images and developed an algorithm for identifying NACA (Narrow Angle Closure Attack). By preprocessing the images, segmenting the corneal and iris light bands, and fitting the central anterior chamber section, eight biological parameters were calculated.
In the training datasets, the accuracy of NACA identification using these parameters reached up to 0.85, with the AUC reaching up to 0.90. The internal test dataset also showed promising performance, with a maximum accuracy of 0.88, a highest sensitivity of 0.95, and a maximum specificity of 1.00.
Additionally, the integrated model, which integrates the eight biological parameters, demonstrated excellent diagnostic performance in the internal test dataset: a sensitivity of 0.90 and a specificity of 0.85. The comprehensive diagnostic efficacy is comparable to professional Scheimpflug imaging devices. The portable slit lamp system equipped with advanced algorithms shows significant potential in glaucoma risk screening application.
Fig. 8 | Portable slit lamp. The image shows the compact design and key features of the portable slit lamp used in the study.
Fig. 9 | Anterior segment slit lamp photos captured for one eye.
Notably, the portable slit lamp (ISPECTOR MINI HE 010-21) used in this study was co-developed by He University and Shenyang Eyerobo Co., Ltd. This portable slit lamp is compact, weighing only 100 grams, and easy to operate. It can complete image acquisition when paired with a smartphone, effectively addressing the issues of high cost and poor portability of the traditional equipment.
This study fully demonstrates the application prospects of integrating AI with portable medical devices for ophthalmic screening, meaning that convenient and efficient glaucoma screening may be accessible to primary medical institutions, communities, and even remote areas in the future, allowing more people to benefit from early detection and intervention.
Of course, this device is still “growing”. Although the model’s performance declined on external datasets, facing generalization challenges primarily due to factors like handheld shooting stability and ambient light—the research team has already formulated improvement plans, such as designing a specialized eye mask to enhance stability and optimizing light band segmentation and anterior chamber angle fitting algorithms.
Technology for All
Promoting More Innovative Achievements to Benefit the Public
In recent years, artificial intelligence (AI) technology has been increasingly applied in the medical field, demonstrating significant potential, especially in ophthalmic diagnosis.
“This research pioneers a groundbreaking approach for the early screening of PACG. This achievement represents not only a milestone of technological innovation, but also a successful model of in-depth integration of medical, education, industry, and research,” emphasized Dr. Guangzheng Dai. With the continuous optimization and expansion of technology, AI-powered smart screening solutions are expected to reshape the pattern of glaucoma prevention and control in community and rural grassroots levels.
As a digital eye health ecosystem industrial group driven by innovative technology, He Eye Specialist Hospital has provided solid support for this research. Moving forward, aligned with the “Healthy China” strategic direction, He Eye Specialist Hospital will fully leverage its eye health whole-industry-chain ecosystem platform, and with the help of cutting-edge technologies such as AI, gene therapy, and stem cells to promote the transformation of eye health services towards intelligence, precision, and inclusiveness, striving to make more technological innovations benefit the people and bring them well-being.
