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
A deep learning model predicts treatment outcomes for age-related macular degeneration with high accuracy, potentially helping doctors avoid unnecessary eye injections and improve patient adherence.
This prospective, nationwide, 18-centre Chinese study developed and externally validated the KongMing deep learning model for predicting visual and anatomical outcomes of anti-VEGF therapy in age-related macular degeneration, achieving AUCs exceeding 0.94 on both internal and external datasets. The model outperformed ophthalmologists across all experience levels and could reduce unnecessary injections while improving patient adherence.
What the study was
- Study design
- Prospective multicentre validation study
- Population
- Patients aged 50-85 with neovascular age-related macular degeneration (nAMD) receiving anti-VEGF therapy; 18 tertiary hospitals, 12 provinces in China
- Sample size
- 1,398 participants (1226 internal + 172 external validation)
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- The Lancet Digital Health
Why it surfaced
Published in Lancet Digital Health; prospective nationwide multicentre design with external validation (n=1398); AUC >0.94 substantially outperforms clinicians; addresses high-burden disease with direct treatment optimization potential; near-term clinical deployment pathway plausible.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.