National Clinical Research Center for Ocular Diseases, Eye Hospital, Wenzhou Medical University, Wenzhou, 325027, China; State Key Laboratory of Ophthalmology, Optometry and Visual Science, Eye Hospital, Wenzhou Medical University, Wenzhou, 325027, China; Faculty of Medicine and Health Sciences, University of Antwerp, Wilrijk, Belgium; ELZA Institute, Dietikon, Switzerland.
Faculty of Medicine and Health Sciences, University of Antwerp, Wilrijk, Belgium; Department of Ophthalmology, Antwerp University Hospital, Edegem, Belgium.
Cont Lens Anterior Eye. 2023 Jun;46(3):101840. doi: 10.1016/j.clae.2023.101840. Epub 2023 Apr 12.
To determine whether combinations of devices with different measuring principles, supported by artificial intelligence (AI), can improve the diagnosis of keratoconus (KC).
Scheimpflug tomography, spectral-domain optical coherence tomography (SD-OCT), and air-puff tonometry were performed in all eyes. The most relevant machine-derived parameters to diagnose KC were determined using feature selection. The normal and forme fruste KC (FFKC) eyes were divided into training and validation datasets. The selected features from a single device or different combinations of devices were used to develop models based on random forest (RF) or neural networks (NN) trained to distinguish FFKC from normal eyes. The accuracy was determined using receiver operating characteristic (ROC) curves, area under the curve (AUC), sensitivity, and specificity.
271 normal eyes, 84 FFKC eyes, 85 early KC eyes, and 159 advanced KC eyes were included. A total of 14 models were built. Air-puff tonometry had the highest AUC for detecting FFKC using a single device (AUC = 0.801). Among all two-device combinations, the highest AUC was accomplished using RF applied to selected features from SD-OCT and air-puff tonometry (AUC = 0.902), followed by the three-device combination with RF (AUC = 0.871) with the best accuracy.
Existing parameters can precisely diagnose early and advanced KC, but their diagnostic ability for FFKC could be optimized. Applying an AI algorithm to a combination of air-puff tonometry with Scheimpflug tomography or SD-OCT could improve FFKC diagnostic ability. The improvement in diagnostic ability by combining three devices is modest.
确定不同测量原理的设备与人工智能（AI）相结合是否能提高对圆锥角膜（KC）的诊断能力。
对所有眼睛进行Scheimpflug 断层扫描、光谱域光相干断层扫描（SD-OCT）和空气脉冲眼压测量。使用特征选择确定最相关的用于诊断 KC 的机器衍生参数。将正常和未定型 KC（FFKC）眼分为训练和验证数据集。使用来自单个设备或不同设备组合的选定特征来开发基于随机森林（RF）或神经网络（NN）的模型，这些模型经过训练可区分 FFKC 与正常眼。使用接收器工作特征（ROC）曲线、曲线下面积（AUC）、灵敏度和特异性来确定准确性。
纳入了 271 只正常眼、84 只 FFKC 眼、85 只早期 KC 眼和 159 只晚期 KC 眼。共建立了 14 个模型。使用单个设备进行空气脉冲眼压测量时，ROC 曲线的 AUC 最高，用于检测 FFKC（AUC=0.801）。在所有两种设备组合中，使用 RF 应用于 SD-OCT 和空气脉冲眼压测量的选定特征的 AUC 最高（AUC=0.902），其次是使用 RF 的三种设备组合（AUC=0.871），准确性最高。
现有的参数可以精确诊断早期和晚期 KC，但它们对 FFKC 的诊断能力可以进一步优化。应用 AI 算法对空气脉冲眼压与 Scheimpflug 断层扫描或 SD-OCT 的组合可以提高 FFKC 的诊断能力。三种设备组合对诊断能力的改善效果不大。
