Ophthalmology Department, Rothschild Foundation Hospital, Paris, France.
Humanitas University, Department of Biomedical Sciences, Milan, Italy.
Acta Ophthalmol. 2024 Nov;102(7):e984-e993. doi: 10.1111/aos.16693. Epub 2024 Apr 29.
To elaborate a deep learning (DL) model for automatic prediction of late recurrence (LR) of rhegmatogenous retinal detachment (RRD) using pseudocolor and fundus autofluorescence (AF) ultra-wide field (UWF) images obtained preoperatively and postoperatively.
We retrospectively included patients >18 years who underwent either scleral buckling (SB) or pars plana vitrectomy (PPV) for primary or recurrent RRD with a post-operative follow-up >2 years. Records of RRD recurrence between 6 weeks and 2 years after surgery served as a ground truth for the training of the deep learning (DL) models. Four separate DL models were trained to predict LR within the 2 postoperative years (binary outputs) using, respectively, UWF preoperative and postoperative pseudocolor images and UWF preoperative and postoperative AF images.
A total of 412 eyes were included in the study (332 eyes treated with PPV and 80 eyes with SB). The mean follow-up was 4.0 ± 2.1 years. The DL models based on preoperative and postoperative pseudocolor UWF imaging predicted recurrence with 85.6% (sensitivity 86.7%, specificity 85.4%) and 90.2% accuracy (sensitivity 87.0%, specificity 90.8%) in PPV-treated eyes, and 87.0% (sensitivity 86.7%, specificity 87.0%) and 91.1% (sensitivity 88.2%, specificity 91.9%) in SB-treated eyes, respectively. The DL models using preoperative and postoperative AF-UWF imaging predicted recurrence with 87.6% (sensitivity 84.0% and specificity 88.3%) and 91.0% (sensitivity 88.9%, specificity 91.5%) accuracy in PPV eyes, and 86.5% (sensitivity 87.5%; specificity 86.2%) and 90.6% (sensitivity 90.0%, specificity 90.7%) in SB eyes, respectively. Among the risk factors detected with visualisation methods, potential novel ones were extensive laser retinopexy and asymmetric staphyloma.
DL can accurately predict the LR of RRD based on UWF images (especially postoperative ones), which can help refine follow-up strategies. Saliency maps might provide further insight into the dynamics of RRD recurrence.
利用术前和术后获得的假彩色和眼底自发荧光（AF）超广角（UWF）图像，阐述一种用于预测孔源性视网膜脱离（RRD）术后晚期复发（LR）的深度学习（DL）模型。
我们回顾性纳入了年龄大于 18 岁的患者，这些患者因原发性或复发性 RRD 接受了巩膜扣带术（SB）或经睫状体平坦部玻璃体切除术（PPV）治疗，术后随访时间大于 2 年。术后 6 周到 2 年内的 RRD 复发记录被用作训练深度学习（DL）模型的真实数据。分别使用 UWF 术前和术后的假彩色图像以及 UWF 术前和术后的 AF 图像训练 4 个独立的 DL 模型，以预测术后 2 年内的 LR（二进制输出）。
共纳入 412 只眼（332 只眼接受了 PPV 治疗，80 只眼接受了 SB 治疗）。平均随访时间为 4.0±2.1 年。基于术前和术后 UWF 假彩色成像的 DL 模型在接受 PPV 治疗的眼中，预测复发的准确率为 85.6%（灵敏度 86.7%，特异性 85.4%）和 90.2%（灵敏度 87.0%，特异性 90.8%），在接受 SB 治疗的眼中，预测复发的准确率为 87.0%（灵敏度 86.7%，特异性 87.0%）和 91.1%（灵敏度 88.2%，特异性 91.9%）。使用术前和术后 AF-UWF 成像的 DL 模型在接受 PPV 治疗的眼中预测复发的准确率为 87.6%（灵敏度 84.0%，特异性 88.3%）和 91.0%（灵敏度 88.9%，特异性 91.5%），在接受 SB 治疗的眼中，预测复发的准确率为 86.5%（灵敏度 87.5%，特异性 86.2%）和 90.6%（灵敏度 90.0%，特异性 90.7%）。在可视化方法检测到的风险因素中，有一些新的潜在风险因素，如广泛的激光光凝和不对称性葡萄肿。
基于 UWF 图像（尤其是术后图像），DL 可以准确预测 RRD 的 LR，这有助于完善随访策略。显著图可能为 RRD 复发的动态提供进一步的见解。
