A cross population study of retinal aging biomarkers with longitudinal pre-training and label distribution learning

作     者：Yu, Zhen Chen, Ruiye Gui, Peng Wang, Wei Razzak, Imran Alinejad-Rokny, Hamid Zeng, Xiaomin Shang, Xianwen Zhang, Lei Yang, Xiaohong Yu, Honghua Huang, Wenyong Lu, Huimin van Wijngaarden, Peter He, Mingguang Zhu, Zhuoting Ge, Zongyuan 

作者机构：Monash Univ AIM Hlth Lab Melbourne Vic Australia Monash Univ Fac Informat Technol Melbourne Vic Australia Univ Melbourne Ctr Eye Res Australia Melbourne Vic Australia Univ Melbourne Dept Surg Ophthalmol Melbourne Vic Australia Wuhan Inst Technol Sch Comp Sci & Engn Artificial Intelligence Wuhan Peoples R China Sun Yat Sen Univ Zhongshan Ophthalm Ctr State Key Lab Ophthalmol Guangzhou Peoples R China MBZUAI Dept Computat Biol Abu Dhabi U Arab Emirates Univ New South Wales Grad Sch Biomed Engn Sydney NSW Australia Hong Kong Polytech Univ Sch Optometry Hong Kong Peoples R China Monash Univ Sch Translat Med Melbourne Vic Australia Guangdong Prov Peoples Hosp Guangdong Acad Med Sci Dept Ophthalmol Guangzhou Peoples R China Southeast Univ Sch Automat Nanjing Peoples R China Monash Airdoc Res Ctr Melbourne Vic Australia 

出 版 物：《NPJ DIGITAL MEDICINE》 (npj Digit. Med.)

年 卷 期：2025年第8卷第1期

页      面：1-14页

核心收录：

基　　金：NHMRC Investigator Grant [APP2010072, APP2041559] NVIDIA AI Technology Centre Global STEM Professorship Scheme [P0046113] Fundamental Research Funds of the State Key Laboratory of Ophthalmology, Project of Investigation on Health Status of Employees in Financial Industry in Guangzhou, China [Z012014075] Victorian State Government Melbourne Research Scholarship 

主　　题：Risk management 

摘      要：Retinal age has emerged as a promising biomarker of aging, offering a non-invasive and accessible assessment tool. We developed a deep learning model to estimate retinal age with enhanced accuracy, leveraging retinal images from diverse populations. Our approach integrates self-supervised learning to capture chronological information from both snapshot and sequential images, alongside a progressive label distribution learning module to model biological aging variability. Trained and validated on healthy cohorts (34,433 participants from the UK Biobank and three Chinese cohorts), the model achieved a mean absolute error of 2.79 years, surpassing previous methods. When applied to broader populations, analysis of the retinal age gap-the difference between retina-predicted and chronological age-revealed associations with increased risks of all-cause mortality and multiple age-related diseases. These findings highlight the potential of retinal age as a reliable biomarker for predicting survival and aging outcomes, supporting targeted risk management and precision health interventions.