Multimodal Prediction of Primary Open-Angle Glaucoma Using Polygenic Risk Scores and Clinical Features in a High-Risk African Ancestry Cohort

Primary open-angle glaucoma (POAG) disproportionately affects individuals of African ancestry, yet early detection tools remain limited. Using the largest African ancestry POAG cohort (N = 7,352), we developed machine learning models that integrate clinical features with four ancestry-matched polygenic risk scores (PRS): two genome-wide (PRS-CS POAAGG, PRS-CS MEGA) and two curated (MTAG PRS, MEGA PRS). Models trained on a clean cohort (N = 196) and tested in 1,013 suspects showed improved prediction with PRS, especially curated scores. MEGA PRS-based predictions correlated with IOP, CDR, and RNFL, validating biological relevance. Combining PRS with inter-eye asymmetry features further enhanced performance. These results support curated PRS such as MEGA PRS and MTAG PRS as interpretable, equitable tools for early glaucoma risk stratification.