文献速递眼底病与葡萄膜炎

AI增强三维OCT扫描对nAMD疾病活动度评估的影响:RAZORBILL研究

Impact of AI-enhanced three-dimensional OCT scans on disease activity assessment in patients with nAMD: the RAZORBILL study.
PubMed 文献 · Br J Ophthalmol · 2026-05-04

这项多中心观察性研究探讨了用分割信息增强OCT图像是否影响新生血管性年龄相关性黄斑变性(nAMD)患者的疾病活动度评估(DAA)。结果显示,增强与非增强OCT图像在DAA的比值比(OR)为1.078(P=0.229),无显著差异;评估者间一致性(Krippendorff's alpha)在增强组为0.416、非增强组为0.402,均较低。结论认为OCT图像增强未影响DAA的可能性或评估者间一致性,也未发现新的安全性信号。

要点

  • OCT图像增强与未增强组DAA的OR为1.078(P=0.229),无显著差异。
  • 评估者间一致性Krippendorff's alpha增强组0.416、非增强组0.402,均较低。
  • 研究未发现涉及产品的新安全性信号。

核心结论

  • OCT图像增强未影响充分检测疾病活动度的可能性,也未影响评估者间一致性。
  • 研究未发现涉及产品的新安全性信号。

研究设计与样本

  • 多中心观察性研究,纳入接受抗VEGF治疗的nAMD患者,比较增强与非增强OCT图像的DAA。
  • 采用广义线性混合模型评估OCT图像增强对DAA的影响,用Krippendorff's alpha评估评估者间一致性。

主要发现

  • 增强OCT图像的DAA比值比为0.759,非增强OCT图像为0.772。
  • 增强 versus 非增强OCT图像的OR为1.078,P值为0.229。
  • Krippendorff's alpha系数增强组为0.416,非增强组为0.402。

临床怎么用

  • 在nAMD患者的疾病活动度评估中,OCT图像增强未显示出优于非增强图像的效应。

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

BACKGROUND: The standard of care for neovascular age-related macular degeneration (nAMD) is repeated intravitreal injection with anti-vascular endothelial growth factor agents. Real world evidence suggests that some patients are undertreated, which might be linked with the retreatment decisions made by physicians. Therefore, this multicentre observational study aimed to investigate the extent to which enrichment of optical coherence tomography (OCT) images with segmentation information could influence and support disease activity assessment (DAA) in patients treated for nAMD. METHODS: Descriptive statistics were tabulated for the demographic and clinical characteristics and outcome variables. To assess the influence of automated OCT image enrichment with segmentation information on DAA, a generalised linear mixed model was employed. The degree of agreement in classification of disease activity across reviewers was assessed by Krippendorff's alpha. RESULTS: The odds estimate for DAA regarding enriched OCT images was 0.759 and for non-enriched OCT images 0.772. The OR for enriched versus non-enriched OCT images was 1.078 with a p value of 0.229. No difference in the odds of DAA between automated OCT image enrichment with segmentation and DAA was noted. Krippendorff's alpha coefficient was 0.416 for enriched and 0.402 for non-enriched OCT images. Thus, the inter-reviewer reliability/agreement was similarly low between enriched and non-enriched OCT images. CONCLUSIONS: OCT image enrichment did not appear to impact the likelihood of adequately detecting disease activity, nor did it have an impact on agreement between reviewers of images in this study. No new safety signals in products involved in the study were detected. TRIAL REGISTRATION NUMBER: NCT04662944.

来源:Br J Ophthalmol,查看原文。版权归原作者所有。

引用

  1. Kortuem KU, Zarranz-Ventura J, Bandello F, et al. Impact of AI-enhanced three-dimensional OCT scans on disease activity assessment in patients with nAMD: the RAZORBILL study. Br J Ophthalmol. 2026. PMID: 42082322. doi:10.1136/bjo-2024-327060

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。