Title of Article: Retinal age gap as a predictive biomarker for mortality risk
Key takeaway points from article:
This study aimed to develop a deep learning (DL) model that predicts age from fundus images and investigated the association between retinal age gap, defined as the difference between predicted and biological age, and mortality risk.
To conduct this study, fundus images of acceptable quality were taken from the UK Biobank, which totaled 80,169 images from 46,969 participants recruited between 2006 - 2010. Images were obtained at a 45-degree non-mydriatic, non-stereo fundus image, centred on the macula, including the optic disc. From this pool, 19,200 fundus images from 11,052 healthy subjects (no self-report of previous disease) with a mean age of 52.6 +/- 7.97 and 53.7% female, were used to train and validate the DL model for age prediction via five fold cross validation. From the remaining participants, 35,913 had mortality data available and were used as the analysis dataset to investigate the association between retinal age gap and mortality.
This trained DL model was found to be fairly accurate, achieving a strong correlation of 0.81 (P < 0.001) between predicted retinal age and chronological age (mean absolute error of 3.55 years). Regression analysis reveals that each 1-year increase in the retinal age gap was associated with the 2% increase in risk of all-cause mortality (HR = 1.02, 95% CI 1.00 to 1.03, P = 0.020) and a 3% increase of non-cardiovascular and non-cancer specific-mortality. However, the retinal age gap was not significantly shown to increase cardiovascular or cancer related mortality.
In conclusion, the retinal age gap may be a revolutionary biomarker of aging that is closely related to mortality risk, without the need for blood testing or blood pressure measurement. This also implies the possibility of using retinal images as a screening tool for risk stratification and management decisions.
Publication Date: January 18 2022.
Reference:
Zhu Z, Shi D, Guankai P, et al. Retinal age gap as a predictive biomarker for mortality risk. British Journal of Ophthalmology Published Online First: 18 January 2022. doi: 10.1136/bjophthalmol-2021-319807
Summary by: Neil Kamra