散光矫正型人工晶状体植入术后轴向旋转的预测研究

Predictors of axial rotation after astigmatism-correcting intraocular lens implantation

目的 研究散光矫正型人工晶状体(TIOL)植入术后轴向旋转的预测因素。方法 选取2019年9月至2022年6月我院收治的100例(200眼)行TIOL植入术的白内障患者,将患者随机分为模型建立组(60例120眼)、外部验证组(40例80眼),比较模型建立组、外部验证组术后发生轴向旋转(发生组)与未发生轴向旋转(未发生组)患者各项临床资料及数据的差异,采用Lasso回归和Logistic回归分析术后轴向旋转的预测因素,R语言建立预测模型,并对所建立模型进行评价和外部验证。结果 模型建立组:发生组患者眼轴长度、撕囊口直径均高于未发生组,植入张力环患者少于未发生组,前囊膜混浊分级与未发生组差异显著,以上差异均有统计学意义(均为P<0.05)。外部验证组:发生组患者眼轴长度、撕囊口直径均高于未发生组,前囊膜混浊分级与未发生组差异显著,以上差异均有统计学意义(均为P<0.05)。其余指标两组间差异均无统计学意义(均为P>0.05)。Lasso回归分析筛选出具有预测价值的4个术后轴向旋转因素:眼轴长度、撕囊口直径、前囊膜混浊分级、植入张力环,其系数分别为1.988、2.076、1.553、-0.762。Logistic回归分析结果显示,眼轴长度、撕囊口直径、前囊膜混浊分级4级是术后轴向旋转的相关危险因素,植入张力环是术后轴向旋转的相关保护因素(均为P<0.05)。针对模型建立组绘制的预测术后轴向旋转列线图模型的预测风险能力指数(C-index)为0.945,提示列线图预测能力较好。绘制模型建立组、外部验证组的校准曲线,结果显示,两者校准度分别为0.905、0.926,提示模型与实际观察结果有较好的一致性。绘制决策分析曲线(DCA)评价模型的临床效用,结果显示,术后轴向旋转的列线图预测模型具有明显的正向净收益。结论 眼轴长度、撕囊口直径、前囊膜混浊分级4级是白内障TIOL术后轴向旋转的相关危险因素,植入张力环是术后轴向旋转的相关保护因素,4项相关危险因素均可作为术后轴向旋转的预测因素,为临床早期预测和干预提供参考。

Objective To investigate the predictors of axial rotation after astigmatism-correcting toric intraocular lens(TIOL)implantation.Methods One hundred cataract patients(200 eyes)admitted to our hospital for TIOL implantation from September 2019 to June 2022 were selected and randomly divided into the model building group(60 patients,120 eyes)and the external validation group(40 patients,80 eyes),to analyze the clinical data of patients with axial rotation(occurrence group)and without axial rotation(non-occurrence group)after surgery in the two groups.Lasso and logistic regression were used to explore the predictors of postoperative axial rotation.A prediction model was developed in R language and then was evaluated and validated externally.Results Model building group:The axial length(AL)and the diameter of the capsulorhexis opening in the occurrence group were higher than those in the non-occurrence group,the number of patients implanted with tension rings was less than that in the non-occurrence group,and the anterior capsule opacification(ACO)grade was significantly different from that in the non-occurrence group(all P<0.05).External validation group:The AL and the diameter of the capsulorhexis opening in the occurrence group were higher than those in the non-occurrence group,and the ACO grade was significantly different from that in the non-occurrence group(all P<0.05).No significant differences were found in the rest indicators between the two groups(all P>0.05).Four factors were identified as having predictive value for postoperative axial rotation through Lasso regression analysis,namely,AL,diameter of the capsulorhexis opening,ACO grade,and tension ring implantation,with correlation coefficients of 1.988,2.076,1.553,and-0.762,respectively.Logistic regression analysis results showed that AL,diameter of the capsulorhexis opening,and ACO grade 4 were risk factors for postoperative axial rotation,while the tension ring implantation was a protective factor for postoperative axial rotation(all P<0.05).The concordance index(C\|index)of the nomogram drawn for the predication of postoperative axial rotation in the model building group was 0.945,indicating that the nomogram had good predictive effects.Calibration curves for the model building group and the external validation group showed that the calibration degrees of the two groups were 0.905 and 0.926,respectively,indicating good consistency between the model and the actual observations.A decision curve was drawn to evaluate the clinical effects of the model,and the results showed that the nomogram prediction model for postoperative axial rotation had significant positive net benefits.Conclusion AL,diameter of the capsulorhexis opening,and ACO grade 4 are risk factors associated with axial rotation after TIOL,and tension ring implantation is a protective factor associated with axial rotation after TIOL.These four factors can all be used as predictors of preoperative axial rotation,providing references for early clinical prediction and intervention.