Predictive Potential of Retina-Based Biological Age in Assessing Chronic Obstructive Pulmonary Disease Risk.

Area of Science:

• Ophthalmology
• Pulmonology
• Artificial Intelligence

Background:

• A deep-learning algorithm, RetiAGE, was previously developed to predict biological age from retinal photographs.
• RetiAGE has demonstrated associations with future morbidity and mortality risks.

Purpose of the Study:

• To evaluate the performance of RetiAGE in predicting the future risk of chronic obstructive pulmonary disease (COPD).

Main Methods:

• RetiAGE scores were generated from retinal images of UK Biobank participants.
• Cox proportional hazards models assessed the association between RetiAGE and incident COPD, adjusting for confounders.
• Generalised linear models examined the association between RetiAGE and baseline respiratory function (FEV1/FVC, PEF).

Main Results:

• Participants in moderate and high RetiAGE risk tertiles had lower baseline respiratory function (p < 0.05).
• Higher RetiAGE risk tertiles were associated with a significantly increased risk of incident COPD (HR = 1.60).
• Incorporating RetiAGE improved the predictive performance of multivariable risk models (p < 0.001).

Conclusions:

• The deep-learning-based retinal aging biomarker, RetiAGE, shows potential for stratifying COPD development risk.