[目的]探讨基于光学相干断层扫描血管成像技术(OCTA),筛查糖尿病肾脏病(DKD)高风险人群的临床预测模型.[方法]本研究以567例糖尿病患者为研究对象,在逐步logistic回归分析的基础上,运用随机森林算法筛选纳入建模的指标,构建基于OCTA糖尿病肾脏病临床预测模型.通过受试者工作特征曲线评价模型区分度,通过决策曲线分析评估模型临床有效性.[结果]构建基于OCTA的DKD临床预测模型,ROC曲线下面积为0.878,Bri-er=0.11.[结论]本研究构建了基于OCTA结果进行糖尿病肾脏病临床预测的列线图预测模型,并从多维度验证模型,从而达到早期预警、提前实施干预的目的.
[Objective]To construct and validate a clinical prediction model for diabetic kidney disease(DKD)based on optical coherence tomography angiography(OCTA).[Methods]This study enrolled 567 diabetes patients.The random forest algorithm as well as logistic regression analysis were applied to construct the prediction model.The model discrimina-tion and clinical usefulness were evaluated by receiver operating characteristic curve(ROC)and decision curve analysis(DCA),respectively.[Results]The clinical prediction model for DKD based on OCTA was constructed with area under the curve(AUC)of 0.878 and Brier score of 0.11.[Conclusions]Through multidimensional verification,the clinical pre-diction nomogram model based on OCTA allowed for early warning and advanced intervention of DKD.
陆丽娇;徐楠;刘鑫鑫;杜芳芳;郑枞;彭鸿钧;曹明哲;艾诗蓓
中山大学附属第七医院眼科,广东 深圳 518000
临床医学
光学相干断层扫描血管成像技术;糖尿病肾脏病;临床预测模型;受试者工作特征曲线;决策曲线分析
optical coherence tomography angiography(OCTA);diabetic kidney disease(DKD);clinical predic-tion model;receiver operating characteristic curve(ROC);decision curve analysis(DCA)
《中山大学学报（医学科学版）》 2024 (002)
253-260 / 8
评论