Background/aims To assess the performance of deep-learning (DL) models for prediction of conversion to normal-tension glaucoma (NTG) in normotensive glaucoma suspect (GS) patients. Methods Datasets of 12 458 GS eyes were reviewed. Two hundred and ten eyes (105 eyes showing NTG conversion and 105 without conversion), followed up for a minimum of 7 years during which intraocular pressure (IOP) was lower than 21 mm Hg, were included. The features of two fundus images (optic disc photography and red-free retinal nerve fibre layer (RNFL) photography) were extracted by convolutional auto encoder. The extracted features as well as 15 clinical features including age, sex, IOP, spherical equivalent, central corneal thickness, axial length, average circumpapillary RNFL thickness, systolic/diastolic blood pressure and body mass index were used to predict NTG conversion. Prediction was performed using three machine-learning classifiers (ie, XGBoost, Random Forest, Gradient Boosting) with different feature combinations. Results All three algorithms achieved high diagnostic accuracy for NTG conversion prediction. The AUCs ranged from 0.987 (95% CI 0.978 to 1.000; Random Forest trained with both fundus images and clinical features) and 0.994 (95% CI 0.984 to 1.000; XGBoost trained with both fundus images and clinical features). XGBoost showed the best prediction performance for time to NTG conversion (mean squared error, 2.24). The top three important clinical features for time-to-conversion prediction were baseline IOP, diastolic blood pressure and average circumpapillary RNFL thickness. Conclusion DL models, trained with both fundus images and clinical data, showed the potential to predict whether and when normotensive GS patients will show conversion to NTG. All data relevant to the study are included in the article or uploaded as online supplemental information.

中文翻译：

---

基于深度学习的正常眼压青光眼疑似患者青光眼转化预测


背景/目的 评估深度学习 (DL) 模型预测正常眼压青光眼疑似 (GS) 患者转化为正常眼压青光眼 (NTG) 的性能。方法回顾 12 458 只 GS 眼的数据集。包括 210 只眼睛（105 只眼睛显示 NTG 转换，105 只眼睛未转换），随访至少 7 年，期间眼压 (IOP) 低于 21 mm Hg。通过卷积自动编码器提取两幅眼底图像（视盘摄影和无红视网膜神经纤维层（RNFL）摄影）的特征。提取的特征以及包括年龄、性别、眼压、球镜当量、中央角膜厚度、眼轴长度、平均周围乳头RNFL厚度、收缩压/舒张压和体重指数在内的15个临床特征用于预测NTG转换。使用具有不同特征组合的三种机器学习分类器（即 XGBoost、随机森林、梯度提升）进行预测。结果所有三种算法都实现了 NTG 转换预测的高诊断准确性。 AUC 范围为 0.987（95% CI 0.978 至 1.000；使用眼底图像和临床特征进行随机森林训练）和 0.994（95% CI 0.984 至 1.000；使用眼底图像和临床特征进行 XGBoost 训练）。 XGBoost 显示了时间到 NTG 转换的最佳预测性能（均方误差，2.24）。转换时间预测的三大重要临床特征是基线眼压、舒张压和平均乳头周围 RNFL 厚度。结论 使用眼底图像和临床数据进行训练的 DL 模型显示出预测血压正常的 GS 患者是否以及何时会转化为 NTG 的潜力。 与研究相关的所有数据都包含在文章中或作为在线补充信息上传。