A new pathway has opened to predict glaucoma risk following vitrectomy surgery, a core procedure for treating retinal diseases, using artificial intelligence.
A research team led by Professor Shin Young-in of the Department of Ophthalmology at Gachon University Gil Medical Center announced on the 27th that they have developed a machine learning-based predictive model for glaucoma risk. The system identifies high-risk patients in advance by integrating pre- and post-operative clinical data with imaging information.
Vitrectomy is an essential surgery for treating major retinal diseases including retinal detachment, diabetic retinopathy, and vitreous hemorrhage. The procedure removes the vitreous body inside the eye and directly treats the retina, with the number of procedures performed in Korea steadily increasing. However, elevated intraocular pressure and secondary glaucoma can develop after surgery, presenting an ongoing clinical challenge. Tools for quantitatively predicting high-risk patients have been limited until now.
The research team analyzed various clinical variables and surgery-related factors using a large-scale clinical database. The model's predictive performance was verified through cross-validation. The model enables early monitoring strategies for high-risk patients, customized follow-up intervals, and preemptive treatment interventions. Its clinical significance is substantial given its potential contribution to long-term vision preservation in surgical patients.
The study received the Poster Discussion Award at the 41st Asia-Pacific Academy of Ophthalmology Congress, the largest ophthalmology conference in the Asia-Pacific region.
"This research is meaningful in that it presents a practically applicable predictive model through collaboration between clinical ophthalmology and AI engineering," said Professor Shin. "We will establish a precision medicine-based glaucoma management system through validation with external cohorts and evaluation of clinical applicability."
The research was conducted in collaboration with the research team of Professor Kim Kwang-gi from the Department of Biomedical Engineering at Gachon University College of Medicine.