文献速递眼底病与葡萄膜炎

使用深度学习自动分割糖尿病视网膜病变的超广角荧光素血管造影图像

Automated segmentation of ultra-widefield fluorescein angiography of diabetic retinopathy using deep learning.
PubMed 文献 · Br J Ophthalmol · 2023-11-22

本研究为回顾性横断面病历审查,收集严重非增殖期或增殖期糖尿病视网膜病变患者的951张超广角荧光素血管造影图像,标注非灌注区、新生血管、视盘、背景和外部区域,用卷积神经网络训练和验证(80%)并测试(20%)自动分割模型。深度学习模型检测新生血管和非灌注区以及视盘和外界性能良好,有助于糖尿病视网膜病变门诊医师并克服分级者主观性。

要点

  • 深度学习模型在超广角荧光素血管造影上分割非灌注区和新生血管性能良好。
  • 各标签最佳准确率:非灌注区0.8208、新生血管0.8338、视盘0.9801。
  • 自动分割有助于糖尿病视网膜病变门诊医师并克服分级者主观性。

核心结论

  • 深度学习模型能以良好性能在超广角荧光素血管造影上检测新生血管和非灌注区以及视盘和外界。
  • 这种对超广角荧光素血管造影重要特征的自动分割将有助于糖尿病视网膜病变门诊医师并克服分级者主观性。

研究设计与样本

  • 回顾性横断面病历审查研究。
  • 共收集951张来自严重非增殖期糖尿病视网膜病变或增殖期糖尿病视网膜病变患者的超广角荧光素血管造影图像。
  • 每张图像均分割并标注非灌注区、新生血管、视盘、背景和外部区域。
  • 使用标注图像,通过卷积神经网络训练和验证(80%)深度学习模型用于自动分割,并在测试集(20%)上测试。
  • 评估每个模型和每个标签的准确率。

主要发现

  • 卷积神经网络模型对非灌注区、新生血管、视盘、背景和外部区域各标签的最佳准确率分别为0.8208、0.8338、0.9801、0.9253和0.9766。
  • 各标签最佳交并比分别为0.6806、0.5675、0.7107、0.8551和0.924,平均平均边界F1分数分别为0.6702、0.8742、0.9092、0.8103和0.9006。

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

BACKGROUND/AIMS: Retinal capillary non-perfusion (NP) and neovascularisation (NV) are two of the most important angiographic changes in diabetic retinopathy (DR). This study investigated the feasibility of using deep learning (DL) models to automatically segment NP and NV on ultra-widefield fluorescein angiography (UWFA) images from patients with DR. METHODS: Retrospective cross-sectional chart review study. In total, 951 UWFA images were collected from patients with severe non-proliferative DR (NPDR) or proliferative DR (PDR). Each image was segmented and labelled for NP, NV, disc, background and outside areas. Using the labelled images, DL models were trained and validated (80%) using convolutional neural networks (CNNs) for automated segmentation and tested (20%) on test sets. Accuracy of each model and each label were assessed. RESULTS: The best accuracy from CNN models for each label was 0.8208, 0.8338, 0.9801, 0.9253 and 0.9766 for NP, NV, disc, background and outside areas, respectively. The best Intersection over Union for each label was 0.6806, 0.5675, 0.7107, 0.8551 and 0.924 and mean mean boundary F1 score (BF score) was 0.6702, 0.8742, 0.9092, 0.8103 and 0.9006, respectively. CONCLUSIONS: DL models can detect NV and NP as well as disc and outer margins on UWFA with good performance. This automated segmentation of important UWFA features will aid physicians in DR clinics and in overcoming grader subjectivity.

来源:Br J Ophthalmol,查看原文。版权归原作者所有。

引用

  1. Lee PK, Ra H, Baek J, et al. Automated segmentation of ultra-widefield fluorescein angiography of diabetic retinopathy using deep learning. Br J Ophthalmol. 2023. PMID: 36241374. doi:10.1136/bjo-2022-321063

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。