文献速递白内障与屈光手术眼底病与葡萄膜炎

视网膜年龄差作为白内障术后患者满意度与视觉质量的生物标志物

Retinal Age Gap as a Biomarker for Patient Satisfaction and Visual Quality After Cataract Surgery.
PubMed 文献 · Am J Ophthalmol · 2025-11-13

本研究为回顾性连续干预病例系列,纳入2024年2月至2025年5月在町田市民病院接受白内障手术、目标术后正视的142只眼(平均年龄77.6±6.2岁),使用日本眼成像登记(JOIR)人工智能模型估算视网膜年龄并计算视网膜年龄差(RAG)。年龄匹配后比较RAG最低四分位组(Q1)与最高四分位组(Q4)的术后1个月结局,发现Q1组满意度更高、脱镜率更高、报告无眩光和无蜡样视觉的患者更多(均P<.01),较大的RAG独立预测较低的满意度(P<.05)。

要点

  • JOIR模型估算视网膜年龄的平均绝对误差为6.98年,均方根误差为8.97年。
  • 年龄匹配后,Q1组与Q4组多数视力和对比敏感度指标无差异,但Q1组满意度、脱镜率更高,报告无眩光和无蜡样视觉者更多(均P<.01)。
  • 较大的RAG独立预测较低的满意度,与矫正远视力无关(P<.05)。

核心结论

  • 较小的视网膜年龄差与更高的术后满意度相关,RAG可能作为个性化人工晶状体选择的实用生物标志物。

研究设计与样本

  • 回顾性、连续干预病例系列,纳入2024年2月至2025年5月在町田市民病院接受白内障手术、目标术后正视的连续病例。
  • 共纳入142只眼(平均年龄77.6±6.2岁),术前角膜散光<1屈光度且无既往眼部手术或其他眼病,植入单焦点人工晶状体或增强型单焦点人工晶状体。
  • 视网膜年龄采用日本眼成像登记(JOIR)人工智能模型估算,RAG为视网膜年龄减实际年龄,定义四分位(Q1-Q4),Q1与Q4年龄匹配后比较术后1个月结局。
  • 采用多变量逻辑回归评估与满意度相关的因素。

主要发现

  • JOIR模型估算视网膜年龄的平均绝对误差为6.98年,均方根误差为8.97年。
  • RAG与实际年龄呈负相关(Spearman ρ=-0.671,P<.001)。
  • 年龄匹配后,Q1组与Q4组多数视力和对比敏感度指标无差异,但Q1组满意度更高、脱镜率更高、报告无眩光和无蜡样视觉的患者更多(均P<.01)。
  • 较大的RAG独立预测较低的满意度,与矫正远视力无关(P<.05)。

关键数值与推荐

  • RAG定义为人工智能估算的视网膜年龄减去实际年龄,按四分位分为Q1至Q4。
  • 年龄匹配后比较Q1与Q4组术后1个月视力、对比敏感度、患者满意度、脱镜率和闪光感。
  • 多变量逻辑回归评估满意度相关因素,P<.05为统计学显著。

临床怎么用

  • 原文结论认为RAG可能作为个性化人工晶状体选择的实用生物标志物。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

PURPOSE: To evaluate whether the difference between artificial intelligence (AI)-estimated retinal biological age and chronological age-the retinal age gap (RAG)-is associated with postoperative visual function after cataract surgery. DESIGN: Retrospective, consecutive interventional case series. METHODS: Consecutive cataract surgeries at Machida Municipal Hospital (February 2024-May 2025) targeting postoperative emmetropia were included. Eligible eyes had preoperative corneal astigmatism <1 diopter and no prior ocular surgery or other ocular disease. Overall, 142 eyes (mean age, 77.6 ± 6.2 years) received a monofocal intraocular lens (IOL) or an enhanced monofocal IOL. Retinal age was estimated with the Japan Ocular Imaging Registry (JOIR) AI model, and RAG was computed as retinal age minus chronological age. Quartiles (Q1-Q4) were defined; after age matching between Q1 and Q4, 1-month outcomes-visual acuity, contrast sensitivity (CS), patient satisfaction, spectacle independence, and dysphotopsia-were compared. Multivariable logistic regression assessed factors associated with satisfaction. RESULTS: The JOIR model's mean absolute error was 6.98 years and root-mean-square error was 8.97 years. RAG correlated negatively with chronological age (Spearman's ρ = -0.671, P < .001). With age matching, most visual acuity and contrast sensitivity measures did not differ; however, Q1 showed higher satisfaction, higher spectacle independence, and more patients reporting no glare and no waxy vision (all P < .01). Larger RAG independently predicted lower satisfaction, irrespective of corrected distance visual acuity (P < .05). CONCLUSIONS: Smaller RAG was associated with greater postoperative satisfaction. RAG may serve as a practical biomarker to personalize intraocular lens selection.

来源:Am J Ophthalmol,查看原文。版权归原作者所有。

引用

  1. Komatsu K, Masuda Y, Sano K, et al. Retinal Age Gap as a Biomarker for Patient Satisfaction and Visual Quality After Cataract Surgery. Am J Ophthalmol. 2025. PMID: 41241358. doi:10.1016/j.ajo.2025.11.006

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。