圆锥角膜进展评估的变量与截断值研究
该回顾性队列研究纳入165例圆锥角膜患者的265只眼,间隔1年±3个月进行两次Pentacam测量,通过逻辑回归定义圆锥角膜进展指数(KPI),并用ROC曲线和约登指数确定截断值。结果显示D-index和KPI在检测进展时具有较高敏感性和特异性,建议用这两个参数评估进展。
要点
- 圆锥角膜进展评估应综合多个角膜变化参数,而非单一变量。
- D-index截断值0.4175,敏感性70.6%,特异性90%。
- KPI截断值-0.78196,敏感性84.7%,特异性90%。
核心结论
- 圆锥角膜进展应通过评估考虑多个角膜变化的参数来定义,建议用D-index和KPI检测进展。
研究设计与样本
- 回顾性队列研究(2010-2016),纳入165例圆锥角膜患者的265只眼,间隔1年±3个月进行两次Scheimpflug测量(Pentacam)。
主要发现
- 用于检测圆锥角膜的变量之间相关性较弱,例如RPImin与Kmax的相关系数r=0.245(p<0.001)。
- KPI由逻辑回归定义,包含Pachymin系数-0.78(p=0.001)、后表面最大高度系数0.27和后表面距最薄点3mm区域角膜曲率系数-12.44(后两者p<0.001)。
- ROC分析中约登指数最高的两个变量为D-index和KPI:D-index截断值0.4175(敏感性70.6%),约登指数0.606;KPI截断值-0.78196(敏感性84.7%),约登指数0.747;两者特异性均为90%。
关键数值与推荐
- D-index截断值0.4175,敏感性70.6%,特异性90%,约登指数0.606。
- KPI截断值-0.78196,敏感性84.7%,特异性90%,约登指数0.747。
临床怎么用
- 建议使用D-index和KPI这两个综合参数来检测圆锥角膜进展。
局限性
- 原文未述局限性
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
AIM: To define variables for the evaluation of keratoconus progression and to determine cut-off values. METHODS: In this retrospective cohort study (2010-2016), 265 eyes of 165 patients diagnosed with keratoconus underwent two Scheimpflug measurements (Pentacam) that took place 1 year apart ±3 months. Variables used for keratoconus detection were evaluated for progression and a correlation analysis was performed. By logistic regression analysis, a keratoconus progression index (KPI) was defined. Receiver-operating characteristic curve (ROC) analysis was performed and Youden Index calculated to determine cut-off values. RESULTS: Variables used for keratoconus detection showed a weak correlation with each other (eg, correlation r=0.245 between RPImin and Kmax, p<0.001). Therefore, we used parameters that took several variables into consideration (eg, D-index, index of surface variance, index for height asymmetry, KPI). KPI was defined by logistic regression and consisted of a Pachymin coefficient of -0.78 (p=0.001), a maximum elevation of back surface coefficient of 0.27 and coefficient of corneal curvature at the zone 3 mm away from the thinnest point on the posterior corneal surface of -12.44 (both p<0.001). The two variables with the highest Youden Index in the ROC analysis were D-index and KPI: D-index had a cut-off of 0.4175 (70.6% sensitivity) and Youden Index of 0.606. Cut-off for KPI was -0.78196 (84.7% sensitivity) and a Youden Index of 0.747; both 90% specificity. CONCLUSIONS: Keratoconus progression should be defined by evaluating parameters that consider several corneal changes; we suggest D-index and KPI to detect progression.
引用
- Shajari M, Steinwender G, Herrmann K, et al. Evaluation of keratoconus progression. Br J Ophthalmol. 2018. PMID: 29858179. doi:10.1136/bjophthalmol-2017-311651
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。