aceuse of bone mineral density in machine learning models to predict age-related macular degeneration risk
Source article: Decoding the Bone-Eye Axis: Machine Learning for Age-Related Macular Degeneration Risk Prediction
Abstract: Age-related macular degeneration (AMD) is a leading cause of irreversible vision loss, yet systemic determinants of its risk remain incompletely understood. Bone mineral density (BMD), a marker of skeletal and biological aging, may reflect shared pathways linking systemic and retinal degeneration. We investigated the association between BMD and AMD using a multilayered framework integrating epidemiological analyses, Mendelian randomization (MR), proteomic and metabolomic profiling, machine learning, and an exper…
Contested: both sides are scored from claims and sources, not community votes.
"CRE-003 OphaGo - Retinal Imaging Adapter - Huced Despro ITS" by Djoko Kuswanto is marked with CC0 1.0. To view the terms, visit https://creativecommons.org/publicdomain/zero/1.0/.
Researchers examined whether bone mineral density relates to age-related macular degeneration using data from UK Biobank, NHANES, and a Tianjin hospital cohort, plus genetic Mendelian randomization, proteomics, metabolomics, machine learning, and a low-BMD rat model. By the September 2026 publication date they reported lower BMD was consistently associated with higher AMD risk and that machine learning models identified BMD as a recurrent predictive contributor alongside age.
The finding matters because BMD is an accessible marker of systemic aging that could help flag retinal vulnerability, but the source itself notes the incremental clinical value is not yet formally tested, measurement methods varied across cohorts, and animal retinal alterations are not direct AMD validation. Whether BMD improves real-world screening or points to actionable pathways remains uncertain.
- Analysis spanned 3 cohorts: UK Biobank, National Health and Nutrition Examination Survey, and a hospital-based Tianjin cohort.
- Two-sample Mendelian randomization provided supportive genetic evidence consistent with a modest potential contribution of higher BMD to lower AMD risk.
- UK Biobank proteomic and metabolomic analyses pointed to extracellular matrix remodeling, lipid transport, amino acid metabolism, and inflammatory pathways.
- Two-step MR prioritized granzyme A, collagen type II alpha 1 chain, and NEL-like protein 1 as candidate molecular intermediates.
Across UK Biobank, NHANES, and a Tianjin hospital cohort, machine learning models flagged lower bone mineral density as a recurrent contributor to age-related macular degeneration risk prediction alongside age.
The same machine learning finding lacks proven incremental clinical utility, relies on cohorts with differing AMD ascertainment and BMD measurement, and animal retinal changes cannot be read as direct AMD validation.
The rundown
The work combined epidemiological analysis across three cohorts with two-sample Mendelian randomization, proteomic and metabolomic profiling from UK Biobank, machine learning modeling, and a glucocorticoid-induced low-BMD rat model.
In rats, researchers observed outer retinal thinning, vascular narrowing, and delayed visual-spatial performance, while molecular analyses highlighted overlapping signatures and prioritized circulating proteins as candidate intermediates rather than established mediators.
Clinical utility of BMD in prediction models remains unproven, cohort methods were heterogeneous, and animal findings do not directly validate AMD pathology.
Sources
- Peer-reviewedCyborg and Bionic Systems2026-09-14
How should this claim be treated?
The debate