ICORG, Imperial College London, London, UK.
Western Eye Hospital Imperial College Healthcare NHS Trust, London, UK.
Expert Rev Mol Diagn. 2021 Jan;21(1):109-118. doi: 10.1080/14737159.2020.1865806. Epub 2020 Dec 28.
To assess a recently described CNN (convolutional neural network) DARC (Detection of Apoptosing Retinal Cells) algorithm in predicting new Subretinal Fluid (SRF) formation in Age-related-Macular-Degeneration (AMD).
Anonymized DARC, baseline and serial OCT images (n = 427) from 29 AMD eyes of Phase 2 clinical trial (ISRCTN10751859) were assessed with CNN algorithms, enabling the location of each DARC spot on corresponding OCT slices (n = 20,629). Assessment of DARC in a rabbit model of angiogenesis was performed in parallel.
A CNN DARC count >5 at baseline was significantly (p = 0.0156) related to development of new SRF throughout 36 months. Prediction rate of eyes using unique DARC spots overlying new SRF had positive predictive values, sensitivities and specificities >70%, with DARC count significantly (p < 0.005) related to the magnitude of SRF accumulation at all time points. DARC identified earliest stages of angiogenesis in-vivo.
DARC was able to predict new wet-AMD activity. Using only an OCT-CNN definition of new SRF, we demonstrate that DARC can identify early endothelial neovascular activity, as confirmed by rabbit studies. Although larger validation studies are required, this shows the potential of DARC as a biomarker of wet AMD, and potentially saving vision-loss.
评估最近描述的 CNN（卷积神经网络）DARC（检测凋亡视网膜细胞）算法在预测年龄相关性黄斑变性（AMD）中新的视网膜下液（SRF）形成中的作用。
对 29 只 AMD 眼的第 2 阶段临床试验（ISRCTN10751859）的匿名 DARC、基线和系列 OCT 图像（n=427）进行了 CNN 算法评估，从而能够在相应的 OCT 切片上定位每个 DARC 点（n=20,629）。同时平行评估了 DARC 在兔血管生成模型中的作用。
基线时 CNN DARC 计数>5 与 36 个月内新 SRF 的发展显著相关（p=0.0156）。在使用覆盖新 SRF 的独特 DARC 点的眼睛中，预测率具有 >70%的阳性预测值、灵敏度和特异性，DARC 计数与所有时间点 SRF 积累的幅度显著相关（p<0.005）。DARC 在体内识别出了血管生成的最早阶段。
DARC 能够预测新的湿性 AMD 活动。仅使用 OCT-CNN 对新的 SRF 进行定义，我们证明 DARC 可以识别早期的内皮新生血管活动，这得到了兔研究的证实。尽管需要更大的验证研究，但这表明 DARC 作为湿性 AMD 的生物标志物具有潜力，并可能挽救视力丧失。
