基于剪切波弹性成像参数的慢性肾病肾纤维化评估预测模型的开发

Development of a Prediction Model for Assessing Renal Fibrosis in Chronic Kidney Disease Based on Shear-wave Elastic Imaging Parameters

目的:探讨基于剪切波弹性成像(SWE)参数构建的预测模型在评估慢性肾脏病(CKD)患者肾脏纤维化程度中的应用价值。方法:选取2022年3月至2024年3月我院收治的126例CKD患者,根据肾脏病理等级分为轻度纤维化组(n=82)和中重度纤维化组(n=44),同时选取同期进行健康体检者68例为对照组。所有研究对象均接受SWE检查,于肾长轴切面肾实质中下极区域定位ROI,评估SWE值。采用多因素Logistic回归分析影响CKD患者肾脏纤维化程度的相关因素,使用R软件的glmpath包构建CKD患者肾脏纤维化程度预测列线图模型并计算一致性指数(C-index),绘制决策曲线评价预测模型的临床适用性。另选取80例CKD患者作为独立数据集对预测模型的效能进行验证。结果:轻度组高血压患病率、血尿素氮水平显著低于中重度组,而肾小球滤过率(eGFR)显著高于中重度组(P<0.05);三组SWE水平对比差异具有统计学意义(F=42.570,P<0.05),其中对照组SWE参数[(6.11±1.89)kPa]低于轻度组SWE参数[(8.00±2.76)kPa],轻度组显著低于中重度组[(8.00±2.76)kPa](P<0.05);多因素Logistic回归分析显示,高血压、eGFR、血尿素氮及SWE值是影响CKD肾脏纤维化程度的独立因素。基于上述因素构建的CKD肾脏纤维化程度预测列线图模型,一致性指数达0.807,显示出良好的区分度。决策曲线分析表明,当阈值概率<78%时,模型的临床净获益率>0,验证了其良好的临床适用性。ROC曲线显示该模型预测的AUC为0.877(95%CI 0.772~0.932),特异度为72.88%,敏感度为95.24%。结论:基于SWE参数的CKD肾脏纤维化预测模型具有较高的区分度和临床适用性,有助于精准识别中重度肾脏纤维化风险患者。

Objective:To develop and evaluate a prediction model for assessing the degree of renal fibrosis in chronic kidney disease(CKD)patients based on shear-wave elastic imaging(SWE)parameters.Methods:A total of 126 patients with CKD admitted to our hospital from March 2022 to March 2024 were selected.According to renal pathological grade,renal fibrosis was divided into mild group(n=82)and moderate and severe group(n=44).At the same time,68 healthy subjects were selected as the control group.After comparing the general data of three groups,all participants received SWE examination,with a region of interest(ROI)positioned in the middle and lower pole of the renal parenchyma in the long axis view to evaluate SWE values.Multivariate Logistic regression was performed to analyze the related factors affecting the degree of renal fibrosis in CKD patients.Using the glmpath package in R software to construct a prediction Nomogram model,and the decision curve was used to analyze the clinical applicability of the Nomogram model.Another 80 patients with CKD were selected as independent data sets to verify the efficacy of the prediction model.Results:The proportion of hypertension and blood urea nitrogen in mild group were lower than those in moderate and severe group,and the level of glomerular filtration rate(eGFR)was higher than those in moderate and severe group(P<0.05).The SWE values were significantly different among the three groups(F=42.570,P<0.05),with the control group having the lowest[(6.11±1.89)kPa],followed by the mild fibrosis group[(8.00±2.76)kPa],and the moderate-to-severe fibrosis group having the highest(values significantly higher than 8.00 kPa,P<0.05).Multivariate Logistic regression analysis showed that hypertension,eGFR,blood urea nitrogen and SWE were the factors affecting the degree of renal fibrosis in CKD patients(all P<0.05).Based on the results of multivariate Logistic regression analysis,a Nomogram model of renal fibrosis degree in CKD patients was constructed,which showed that the C-index of the model was 0.807,indicating good discriminative ability.Moreover,the decision curve showed that when the threshold probability was<78%,the Nomogram model had the clinical net benefit rate>0,suggesting a good clinical applicability of the model.The ROC curve showed that the AUC predicted by the model was 0.877(95%CI 0.772~0.932),with the specificity of 72.88%,and the sensitivity of 95.24%.Conclusion:The Nomogram prediction model of kidney fibrosis in CKD based on SWE parameters has good differentiation and clinical applicability,which is helpful to screen out patients with moderate to severe kidney fibrosis.