机构地区:[1]北京医院超声科国家老年医学中心中国医学科学院老年医学研究院,100730 [2]北京医院心内科国家老年医学中心中国医学科学院老年医学研究院,100730 [3]北京医院核医学科国家老年医学中心中国医学科学院老年医学研究院,100730 [4]中国医学科学院阜外医院医学研究统计中心国家心血管病中心,北京100037 [5]北京医院血管外科国家老年医学中心中国医学科学院老年医学研究院,100730
出 处:《中华老年医学杂志》2021年第7期836-841,共6页Chinese Journal of Geriatrics
基 金:北京医院临床研究121工程资助项目(BJ-2018-198);中国医学科学院中央级公益性科研院所基本科研业务费项目(2019PT320012)。
摘 要:目的建立和验证用于评估老年肾动脉狭窄(RAS)预后的超声造影(CEUS)模型.方法单中心回顾性研究,纳入2017年10月至2020年7月北京医院就诊的324例老年RAS患者(按照1∶1比例随机分配建模组174例,验证组150例).收集患者的入院临床和影像学资料,包括一般情况、既往病史、血压、血肌酐、RAS程度和患肾的皮质血流灌注情况和随访资料,应用单因素和多因素Logistic回归建立CEUS评分系统模型,应用受试者工作特征曲线(ROC)和曲线下面积(AUC)评估预测准确度,应用决策曲线分析和列线图评估CEUS评分系统模型的临床应用价值.结果建模组和验证组患者的基线临床和影像学资料相关指标比较差异无统计学意义(均P>0.05).多因素Logistic回归分析结果显示,年龄[OR(95%CI):1.242(1.081~1.427),P<0.01]、糖尿病[OR(95%CI):1.545(1.107~2.156),P<0.05]、血压[OR(95%CI):1.328(1.056~1.670),P<0.05]、肾功能[OR(95%CI):2.374(1.216~3.887),P<0.01]和皮质血流灌注参数[OR(95%CI):2.646(1.553~6.369),P<0.01]是随访1年肾功能恶化的危险因素.根据多因素Logistic回归分析结果显示,绘制CEUS评分系统模型的列线图,其一致性指数C-Index为0.725(95%CI:0.653~0.776).建模组CEUS评分系统模型ROC曲线的AUC为0.824,约登指数为0.711(特异度为0.774,敏感度为0.837);验证组CEUS评分系统模型ROC曲线的AUC为0.853,约登指数为0.715(特异度为0.684,敏感度为0.889);两组ROC比较差异无统计学意义(D=1.387,P>0.05).此外,建模组和验证组模型的校准图显示CEUS评分系统模型校准曲线与标准曲线均接近,差异无统计学意义(均P>0.05).结论CEUS评分系统模型可用于预测老年RAS患者随访1年的肾功能恶化风险.Objective To establish and validate a prognostic model of a contrast-enhanced ultrasound scoring(CEUS)system for evaluating renal artery stenosis(RAS)in the elderly.Methods This was a single-center retrospective study.A total of 324 elderly RAS patients admitted to Beijing Hospital from October 2017 to July 2020 were randomly assigned into the model group(n=174)and the validation group(150)in a 1:1 ratio.Clinical and imaging data of patients on admission including general conditions,previous medical history,blood pressure,blood creatinine,renal artery stenosis and cortical blood perfusion in the affected kidney and renal function(GFR)at 1-year follow-up were collected.Univariate and multivariate logistic regression was used to establish a model of the CEUS scoring system.The receiver operating characteristic(ROC)curve and area under the ROC curve(AUC)were used to evaluate prediction accuracy.Clinical application value of the CEUS scoring system model was evaluated via decision curve analysis using a nomogram.Results Baseline clinical and radiomic data had no significant difference between the model group and the validation group(P>0.05).Multivariate logistic regression analysis results showed that age(OR=1.242,95%CI:1.081-1.427,P<0.01),diabetes(OR=1.545,95%CI:1.107-2.156,P<0.05),blood pressure(OR=1.328,95%CI:1.056-1.670,P<0.05),renal fimction(OR=2.374,95%CI:1.216-3.887,P<0.01)and cortical blood perfusion parameter(OR=2.646,95%CI:1.553-6.369,P<0.01)were risk factors for the deterioration of renal function during 1 year follow-up.Based on these results,a nomogram for the CEUS scoring system model was drawn,and its consistency index,the C-Index,was 0.725(95%CI:0.653-0.776).The AUC of the CEUS scoring system was 0.824 and the Youden index was 0.711 in the model group,with a specificity of 0.774 and a sensitivity of 0.837.The AUC of the CEUS scoring system was 0.853 and the Youden index was 0.715 m the validation group,with a specificity of 0.684 and a sensitivity of 0.889.There was no significant difference in ROC cu
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...