Foundation model-driven distributed learning for enhanced retinal age prediction.

The capacity of this framework to integrate data from underrepresented populations for training of retinal age prediction models could significantly enhance the accessibility of the RAG as an important disease biomarker.