Department of Ophthalmology, the Second Affiliated Hospital of Zhejiang University, College of Medicine, Hangzhou, 310009, China.
Hangzhou Truth Medical Technology Ltd, Hangzhou, 311215, China.
Graefes Arch Clin Exp Ophthalmol. 2021 Aug;259(8):2401-2411. doi: 10.1007/s00417-021-05151-x. Epub 2021 Apr 12.
To detect the leakage points of central serous chorioretinopathy (CSC) automatically from dynamic images of fundus fluorescein angiography (FFA) using a deep learning algorithm (DLA).
The study included 2104 FFA images from 291 FFA sequences of 291 eyes (137 right eyes and 154 left eyes) from 262 patients. The leakage points were segmented with an attention gated network (AGN). The optic disk (OD) and macula region were segmented simultaneously using a U-net. To reduce the number of false positives based on time sequence, the leakage points were matched according to their positions in relation to the OD and macula.
With the AGN alone, the number of cases whose detection results perfectly matched the ground truth was only 37 out of 61 cases (60.7%) in the test set. The dice on the lesion level were 0.811. Using an elimination procedure to remove false positives, the number of accurate detection cases increased to 57 (93.4%). The dice on the lesion level also improved to 0.949.
Using DLA, the CSC leakage points in FFA can be identified reproducibly and accurately with a good match to the ground truth. This novel finding may pave the way for potential application of artificial intelligence to guide laser therapy.
使用深度学习算法（DLA）从眼底荧光血管造影（FFA）的动态图像中自动检测中心性浆液性脉络膜视网膜病变（CSC）的渗漏点。
该研究纳入了 262 例患者的 291 只眼（137 只右眼和 154 只左眼）的 291 个 FFA 序列的 2104 张 FFA 图像。使用注意门控网络（AGN）对渗漏点进行分割。使用 U-net 同时对视盘（OD）和黄斑区域进行分割。为了根据时间序列减少假阳性的数量，根据 OD 和黄斑的位置对渗漏点进行匹配。
仅使用 AGN，在测试集中，有 61 个病例中的检测结果与金标准完全匹配的仅有 37 个（60.7%）。病变级别的骰子系数为 0.811。使用消除假阳性的程序，准确检测到的病例数增加到 57 个（93.4%）。病变级别的骰子系数也提高到了 0.949。
使用 DLA，可以准确、可重复地识别 FFA 中的 CSC 渗漏点，与金标准的匹配良好。这一新颖的发现可能为人工智能在指导激光治疗中的潜在应用铺平道路。
