有晶状体眼人工晶状体植入患者术后异常拱高的眼部参数预测研究
回顾性病例系列研究纳入332例(664只眼)行后房型有晶状体眼人工晶状体植入术的近视患者,按术后3个月拱高是否在250~750 μm分为低、正常、高拱高组。低拱高的独立预测参数为晶状体厚度(LT)、水平睫状沟直径与角膜水平直径差值(HCS-WTW)及前房深度(ACD),高拱高无独立预测参数;Logistic回归模型可提高异常拱高的预测效力。
要点
- 低拱高独立预测参数为LT、HCS-WTW和ACD,高拱高无独立预测参数。
- LT预测低拱高的AUC为0.829,截断值3.745 mm。
- HCS-WTW和ACD预测低拱高的AUC分别为0.745和0.730。
核心结论
- 术后低拱高的预测参数包括LT、HCS-WTW及ACD,高拱高无独立预测参数。
- Logistic回归方程能提高异常拱高的预测效力。
研究设计与样本
- 回顾性病例系列研究,纳入2020年11月至2021年11月于上海普瑞眼科医院和无锡华夏眼科医院行pIOL植入术的近视眼患者332例(664只眼),女性246例(492只眼),男性86例(172只眼),年龄中位数28岁。
- 收集术前眼部参数包括SE、IOP、HCS、VCS、WTW、K1/K2、CCT、ACD、LT及眼轴长度,测量术后3个月拱高。
- 根据术后拱高是否处于250~750 μm理想范围分为低拱高组、正常拱高组及高拱高组,以正常拱高组为基准进行ROC曲线分析并建立多元Logistic回归模型。
主要发现
- 3个组WTW、HCS与WTW差值、ACD及LT的差异有统计学意义;高拱高组WTW与其余两组差异有统计学意义(P<0.01),低拱高组HCS与WTW差值与其余两组差异有统计学意义(P<0.01),ACD与LT在3个组间差异均有统计学意义(P<0.01)。
- LT预测低拱高的AUC、灵敏度、特异度分别为0.829、0.952、0.561,截断值3.745 mm;HCS-WTW分别为0.745、0.857、0.644,截断值0.020 mm;ACD分别为0.730、0.619、0.853,截断值2.975 mm。
- 各眼部参数对高拱高状态的预测性较差(均AUC<0.7)。
- 低拱高预测模型Logistic(V1)=-10.067+5.328·HCS-3.620·WTW+6.263·LT,高拱高预测模型Logistic(V2)=6.232+1.323·WTW-3.358·LT。
- 2个预测模型参数对低、高拱高的AUC、灵敏度、特异度分别为0.884、0.810、0.824和0.736、0.810、0.554。
关键数值与推荐
- 术后拱高理想范围为250~750 μm,据此分为低拱高组、正常拱高组及高拱高组。
- 低拱高独立预测参数LT的截断值为3.745 mm,HCS-WTW的截断值为0.020 mm,ACD的截断值为2.975 mm。
临床怎么用
- 术前可依据LT、HCS-WTW及ACD评估pIOL植入术后低拱高风险。
- Logistic回归模型参数可用于提高低拱高和高拱高的预测诊断效率。
局限性
- 原文未报告局限性。
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
Objective: To analyze the relationships between preoperative ocular parameters and postoperative anomalous vaults, and research their predictive diagnostic value. Methods: In this retrospective case series study, 664 eyes from 332 patients underwent posterior chamber phakic intraocular lens (pIOL) implantation at Shanghai Bright Eye Hospital and Wuxi Huaxia Eye Hospital from November 2020 to November 2021. Preoperative ocular parameters, including spherical equivalent, intraocular pressure, horizontal/vertical ciliary sulcus diameters (HCS/VCS), white-to-white diameters (WTW), corneal steep/flat curvature, central corneal thickness, anterior chamber depth (ACD), lens thickness (LT), and axial length were collected. The pIOL vaults were measured 3 months after surgery. Patients were categorized into low vault group, optimal vault group, and high vault group based on whether the vault fell within the ideal range (250 to 750 μm). Using the optimal vault group as a benchmark, receiver operating characteristic (ROC) curves were drawn for each ocular parameter of the low and high vault groups to analyze diagnostic efficiency and cut-off values for abnormal vaults after pIOL operation. Each ocular parameter was used as an independent variable to establish a multivariate logistic regression model for two different vault anomalies. ROC curves were drawn and analyzed again based on the regression results. Results: Statistically significant differences were observed in WTW, HCS-WTW, ACD, and LT among the three groups. Comparisons between each pair of groups indicated that WTW in the high vault group significantly differed from the other two groups (P<0.05), HCS-WTW in the low vault group significantly differed from the other groups (P<0.05), and ACD and LT explained statistical differences among the three groups (P<0.05), while other parameters showed no differences. ROC curves illustrated that independent ocular parameters such as LT, HCS-WTW, and ACD had clinical predictive diagnostic significance for low vault abnormalities. The area under the curve (AUC), sensitivity, and specificity for these parameters were 0.829(0.952, 0.561), 0.745(0.857, 0.644), and 0.730(0.619, 0.853), respectively. The diagnostic cut-off values were 3.745, 0.020, and 2.975 mm, respectively. The clinical predictive significance of independent ocular parameters in diagnosing the high vault group was poor (AUC<0.7). The predictive Logistic model equation for low vault was Logistic(V1)=-10.067+5.328·HCS-3.620·WTW+6.263·LT, and the predictive model for high vault was Logistic(V2)=6.232+1.323·WTW-3.358·LT. The new parameters in the predictive equation significantly improved the diagnostic efficiency of low and high vault abnormalities, reaching 0.884(0.810, 0.824) and 0.736(0.810, 0.554), respectively. Conclusions: Preoperative predictive diagnostic parameters for postoperative low vault group included LT, HCS-WTW, and ACD, while the high vault group had no independent predictive diagnostic parameters. Logistic regression improved the predictive diagnostic efficiency of abnormal vaults. 目的: 探讨术前眼部参数与有晶状体眼后房型人工晶状体(pIOL)植入术后拱高异常状态的关系及其预测价值。 方法: 回顾性病例系列研究。收集2020年11月至2021年11月于上海普瑞眼科医院和无锡华夏眼科医院行pIOL植入术的近视眼患者332例(664只眼),其中女性246例(492只眼),男性86例(172只眼);年龄中位数为28岁。收集患者术前眼部参数,包括等效球镜度数(SE)、眼压(IOP)、水平睫状沟直径(HCS)、垂直睫状沟直径(VCS)、角膜水平直径(WTW)、陡峭轴曲率(K1)/平坦轴曲率(K2)、角膜中央厚度(CCT)、前房深度(ACD)、晶状体厚度(LT)以及眼轴长度,并测量患者术后3个月的拱高。根据术后拱高是否处于250~750 μm的理想范围分为低拱高组、正常拱高组及高拱高组,对低拱高和高拱高眼的各参数进行受试者工作特征(ROC)曲线分析,再将各眼部参数作为协变量建立多元Logistic回归模型,根据回归结果获取新参数再进行ROC曲线分析。 结果: 3个组WTW、HCS与WTW差值、ACD及LT的差异有统计学意义,组间比较显示高拱高组的WTW与其余两组的差异有统计学意义(P<0.01),低拱高组HCS与WTW差值与其余两组差异有统计学意义(P<0.01),ACD与LT在3个组间的差异均有统计学意义(P<0.01)。ROC曲线分析各参数对预测低拱高的曲线下面积(AUC)及灵敏度、特异度:LT分别为0.829、0.952和0.561,HCS与WTW差值分别为0.745、0.857和0.644,ACD则为0.730、0.619和0.853;这3个参数的截断值分别为3.745、0.020和2.975 mm。而各眼部参数对高拱高状态的预测性较差(均AUC<0.7)。低拱高预测模型参数Logistic(V1)=-10.067+5.328·HCS-3.620·WTW+6.263·LT,高拱高预测模型参数为Logistic(V2)=6.232+1.323·WTW-3.358·LT;2个预测模型参数分别对低、高拱高的预测性均有显著提高,AUC、灵敏度、特异度分别为0.884、0.810、0.824和0.736、0.810、0.554。 结论: 术后低拱高的预测参数包括LT、HCS与WTW差值及ACD,高拱高无独立预测参数;Logistic回归方程能提高异常拱高的预测效力。.
引用
- Yu P, Huang L, Li Q, et al. [Investigating ocular parameters for predicting anomalous vault among phakic intraocular lens patients]. Zhonghua Yan Ke Za Zhi. 2023. PMID: 38061901. doi:10.3760/cma.j.cn112142-20231024-00167
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。