Tanito Masaki, Koyama Makoto
Department of Ophthalmology, Faculty of Medicine, Shimane University, Enya 89-1, Izumo 693-8501, Shimane, Japan.
Minamikoyasu Eye Clinic, 2-8-30 Minamikoyasu, Kimitsu 299-1162, Chiba, Japan.
Int J Mol Sci. 2025 May 15;26(10):4725. doi: 10.3390/ijms26104725.
Glaucoma is an age-related neurodegenerative disease characterized by progressive optic nerve damage. Accelerated biological aging, assessed using predicted age derived from fundus images, may serve as a biomarker for glaucoma progression. This study aimed to examine fundus-derived age acceleration among patients with primary open-angle glaucoma (POAG), exfoliation glaucoma (EXG), and controls, and to explore its biochemical basis through advanced glycation end products (AGEs). Fundus photographs from 237 participants (79 POAG, 79 EXG, and 79 age- and sex-matched controls) were analyzed using a deep learning model (EfficientNet) previously trained to predict biological age. AGE accumulation was assessed by measuring skin autofluorescence (sAF). Multivariate regression analyses were conducted to identify factors influencing predicted age acceleration, with stratification into age tertiles to control for age-related effects. EXG patients demonstrated significant accelerated biological aging compared to controls ( = 0.006), particularly evident in younger and middle-aged tertiles. AGE scores were significantly elevated in EXG relative to both POAG ( = 0.009) and control groups ( = 0.003). Predicted age and AGE scores were more strongly correlated than chronological age and AGEs, especially in the middle tertile ( = 0.002). Accelerated biological aging detected via fundus images occurs prominently in EXG, potentially reflecting underlying AGE accumulation. Fundus-derived predicted age could serve as a non-invasive biomarker for assessing glaucoma progression risk and warrants further exploration in clinical applications.
青光眼是一种与年龄相关的神经退行性疾病,其特征为进行性视神经损伤。使用从眼底图像得出的预测年龄评估的加速生物衰老,可能作为青光眼进展的生物标志物。本研究旨在检查原发性开角型青光眼(POAG)、剥脱性青光眼(EXG)患者及对照者的眼底源性年龄加速情况,并通过晚期糖基化终产物(AGEs)探索其生化基础。使用先前训练用于预测生物年龄的深度学习模型(EfficientNet),对237名参与者(79名POAG患者、79名EXG患者以及79名年龄和性别匹配的对照者)的眼底照片进行分析。通过测量皮肤自发荧光(sAF)评估AGEs积累情况。进行多变量回归分析以确定影响预测年龄加速的因素,并分层为年龄三分位数以控制年龄相关效应。与对照者相比,EXG患者表现出显著的加速生物衰老(P = 0.006),在年轻和中年三分位数中尤为明显。相对于POAG组(P = 0.009)和对照组(P = 0.003),EXG患者的AGE评分显著升高。预测年龄与AGE评分的相关性比实际年龄与AGEs的相关性更强,尤其是在中间三分位数(P = 0.002)。通过眼底图像检测到的加速生物衰老在EXG中尤为突出,可能反映了潜在的AGEs积累。眼底源性预测年龄可作为评估青光眼进展风险的非侵入性生物标志物,值得在临床应用中进一步探索。