Department of Ophthalmology of Federal University of São Paulo, São Paulo, Brazil; Rio de Janeiro Corneal Tomography and Biomechanical Study Group, Rio de Janeiro, Brazil; Instituto de Olhos Renato Ambrósio, Rio de Janeiro, Brazil; School of Engineering, University of Liverpool, Liverpool, United Kingdom.
Rio de Janeiro Corneal Tomography and Biomechanical Study Group, Rio de Janeiro, Brazil.
Am J Ophthalmol. 2018 Nov;195:223-232. doi: 10.1016/j.ajo.2018.08.005. Epub 2018 Aug 9.
To improve the detection of corneal ectasia susceptibility using tomographic data.
Multicenter case-control study.
Data from patients from 5 different clinics from South America, the United States, and Europe were evaluated. Artificial intelligence (AI) models were generated using Pentacam HR (Oculus, Wetzlar, Germany) parameters to discriminate the preoperative data of 3 groups: stable laser-assisted in situ keratomileusis (LASIK) cases (2980 patients with minimum follow-up of 7 years), ectasia susceptibility (71 eyes of 45 patients that developed post-LASIK ectasia [PLE]), and clinical keratoconus (KC; 182 patients). Model accuracy was independently tested in a different set of stable LASIK cases (298 patients with minimum follow-up of 4 years) and in 188 unoperated patients with very asymmetric ectasia (VAE); these patients presented normal topography (VAE-NT) in 1 eye and clinically diagnosed ectasia in the other (VAE-E). Accuracy was evaluated with ROC curves.
The random forest (RF) provided highest accuracy among AI models in this sample with 100% sensitivity for clinical ectasia (KC+VAE-E; cutoff 0.52), being named Pentacam Random Forest Index (PRFI). Considering all cases, the PRFI had an area under the curve (AUC) of 0.992 (94.2% sensitivity, 98.8% specificity; cutoff 0.216), being statistically higher than the Belin/Ambrósio deviation (BAD-D; AUC = 0.960, 87.3% sensitivity, 97.5% specificity; P = .006, DeLong's test). The optimized cutoff of 0.125 provided sensitivity of 85.2% for VAE-NT and 80% for PLE, with 96.6% specificity.
The PRFI enhances ectasia diagnosis. Further integrations with corneal biomechanical parameters and with the corneal impact from laser vision correction are needed for assessing ectasia risk.
利用断层扫描数据提高对角膜扩张易感性的检测。
多中心病例对照研究。
评估来自南美洲、美国和欧洲 5 个不同诊所的患者数据。使用 Pentacam HR（德国 Oculus，威茨拉尔）参数生成人工智能（AI）模型，以区分三组术前数据：稳定的激光辅助原位角膜磨镶术（LASIK）病例（2980 例，至少随访 7 年）、扩张易感性（45 例患者的 71 只眼发生 LASIK 后扩张[PLE]）和临床圆锥角膜（KC；182 例患者）。在另一组稳定的 LASIK 病例（298 例，至少随访 4 年）和 188 例未经手术的严重不对称扩张（VAE）患者中，对模型的准确性进行了独立测试；这些患者的 1 只眼表现为正常地形（VAE-NT），另 1 只眼诊断为临床扩张（VAE-E）。通过 ROC 曲线评估准确性。
在本样本中，随机森林（RF）在 AI 模型中提供了最高的准确性，对临床扩张（KC+VAE-E；截断值 0.52）的灵敏度为 100%，被命名为 Pentacam Random Forest Index（PRFI）。考虑到所有病例，PRFI 的曲线下面积（AUC）为 0.992（94.2%的敏感性，98.8%的特异性；截断值 0.216），统计学上高于 Belin/Ambrósio 偏差（BAD-D；AUC=0.960，87.3%的敏感性，97.5%的特异性；P=0.006，DeLong 检验）。优化的截断值为 0.125，对 VAE-NT 的敏感性为 85.2%，对 PLE 的敏感性为 80%，特异性为 96.6%。
PRFI 增强了扩张的诊断。为了评估扩张风险，需要进一步整合角膜生物力学参数和激光视力矫正对角膜的影响。
