Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
A transformer-based deep learning model called KongMing predicts visual and anatomical outcomes of anti-VEGF therapy in neovascular AMD patients—outperforming experienced ophthalmologists across all accuracy metrics.
Neovascular age-related macular degeneration (nAMD) affects millions of older adults worldwide. While anti-VEGF therapy is the standard treatment, outcomes vary significantly between patients. The challenge? Knowing in advance who’ll benefit most. Accurate predictions before treatment could enable personalized care planning and improve patient expectations.
The KongMing Model was developed and validated using 29,772 OCT images from 1,226 patients across 18 Chinese tertiary hospitals. The transformer-based multitask model predicted best-corrected visual acuity (BCVA) changes—here’s what it achieved:
Clinicians get what they need: a non-invasive tool to deliver personalized outcome predictions before treatment. This enhances clinical decision-making, improves patient adherence through realistic expectations, and identifies patients unlikely to benefit—potentially reducing unnecessary interventions and healthcare burden.
One caveat: the study enrolled only treatment-naive patients from a single country, which limits how well the findings generalize to other populations and treatment-experienced cohorts. Broader external validation is needed.
Original paper: Development and validation of a deep learning model to predict visual and anatomical prognosis of anti-VEGF therapy for neovascular age-related macular degeneration (KongMing Study): a prospective, nationwide, multicentre study. — The Lancet. Digital health. 10.1016/j.landig.2027.100971