An npj Digital Medicine study has identified the retinal age gap – the difference between a person’s retinal age as predicted by artificial intelligence (AI) and their chronological age – as a non-invasive indicator of reproductive aging in women.
Using a deep learning model known as FLEX (Frozen and Learning Ensemble Crossover), the China-based team of researchers analyzed 2,560 retinal images from 1,294 healthy Chinese women to predict retinal age and compare them to each subject’s chronological age. The difference between the two showed a significant inverse correlation with Anti-Müllerian Hormone (AMH) levels, especially in women aged 40–50, a group at critical reproductive transition.
This novel metric demonstrated a compelling association: for women aged 45–50, each additional year in the retinal age gap was linked to a 20 percent increased likelihood of low AMH levels, suggesting accelerated reproductive aging. The study also linked a higher retinal age gap to earlier menopause and elevated levels of follicle-stimulating hormone (FSH) and prolactin, further substantiating its relevance in endocrine and reproductive health.
The study’s visual saliency maps revealed that vascular regions, particularly around the optic disc and temporal arcades, were most influential in age prediction – consistent with known vascular aging patterns. Furthermore, a genome-wide association study (GWAS) highlighted shared genetic pathways between retinal aging and ovarian function, including progesterone-mediated oocyte maturation and oxytocin signaling.
Perhaps most notably, the study demonstrated that predictive accuracy of AMH levels improved when retinal images were combined with genetic data, indicating potential for multimodal precision medicine applications.
The study underscores an expanding role in cross-disciplinary diagnostics, say the authors, reinforcing the retina’s emerging role as a window into systemic aging. Fundus photography, already standard in ocular health assessments, could soon serve a dual purpose in screening for reproductive and vascular aging. As AI tools become more and more integrated into clinical workflows, retinal imaging could soon grow to inform fertility counseling, menopause timing, and age-related endocrine risks, highlighting an emerging intersection of ocular imaging and women's health.