基于临床及超声特征的颈动脉易损斑块模型的建立及分析

Establishment and Analysis of Carotid Artery Vulnerable Plaque Model based on Clinical and Ultrasonic Features

目的建立基于临床及超声特征的颈动脉易损斑块诊断模型,辅助临床进行辨别诊断。方法选取2021年6月~2022年2月于笔者医院检查发现颈动脉斑块患者86例,斑块86个,分为易损斑块组45个和非易损斑块组41个,记录患者临床及斑块超声特征,比较两组间各指标的差异并进行多因素Logistic回归分析;依据Fisher逐步判别分析法建立基于临床及超声特征的颈动脉易损斑块诊断模型,并对其诊断效能进行评价。结果两组间性别、年龄、体重指数(body mass index,BMI)、吸烟指数、脉压差、总胆固醇(total cholesterol,TC)、斑块最大厚径、内部有无强回声、内部强回声数量、内部强回声位置、超微血管成像技术(superb micro-vascular imaging,SMI)新生血管级别比较,差异均有统计学意义(P<0.05);对以上指标进行多因素Logistic回归分析,结果显示,性别、年龄、BMI、TC和SMI新生血管级别是颈动脉易损斑块的独立影响因素;诊断模型为Y_(1)=-108.759-7.527X_(1)+1.285X_(2)+4.457X_(3)+2.489X_(6)+4.332X_(11),Y_(2)=-84.538-3.418X_(1)+1.146X_(2)+3.910X_(3)+1.545X_(6)+3.081X_(11)(Y_(1)、Y_(2)、X_(1)、X_(2)、X_(3)、X_(6)、X_(11)分别代表易损斑块组、非易损斑块组、性别、年龄、BMI、TC、SMI新生血管级别),自身验证及交叉验证的正确率分别为83.7%和81.4%,ROC曲线分析显示其敏感度为84.4%,特异性为82.9%。结论依据Fisher逐步判别分析法建立的基于临床及超声特征的颈动脉易损斑块诊断模型可以为临床早期判断及诊断脑血管事件提供依据,有较高的应用价值。

Objective To establish a diagnostic model of carotid artery vulnerable plaque based on clinical and ultrasonic features,and to assist clinical discrimination diagnosis.Methods A total of 86 patients with carotid artery plaque who were hospitalized in Affiliated Hospital of North China University of Science and Technology from October 2021 to February 2020 were selected,there were 86 plaques,45 in vulnerable plaque group and 41 in non-vulnerable plaque group.The clinical and ultrasound characteristics of plaque were recorded and the differences of each index between the two groups were compared and multivariate Logistic regression analysis was performed.The diagnostic model of carotid artery vulnerable plaques based on Clinical and ultrasonic features was established by Fisher stepwise discriminant analysis,and its diagnostic efficacy was evaluated.Results There were statistically significant differences in gende,age,body mass index(BMI),smoking index,pulse pressure difference,total cholesterol(TC),maximum thickness of plaque,presence or absence of strong echo in plaque,number of strong echo in plaque,location of strong echo in plaque,the grade of SMI neovascularization between the two groups(P<0.05);the results of multivariate Logistic regression analysis showed that gender,age,BMI,TC and the grade of SMI neovascularization were independent influencing factors of vulnerable carotid plaque,the diagnostic model was established:Y_(1)=-108.759-7.527X_(1)+1.285X_(2)+4.457X_(3)+2.489X_(6)+4.332X_(11),Y_(2)=-84.538-3.418X_(1)+1.146X_(2)+3.910X_(3)+1.545X_(6)+3.081X_(11)(Y_(1),Y_(2),X_(1),X_(2),X_(3),X_(6),X_(11)represent vulnerable plaques,non-vulnerable plaques,gender,age,BMI,TC,the grade of SMI neovascularization),the accuracies of self-verification and cross-validation external data validation were 83.7%and 81.4%,respectively.The resluts of receiver operating characteristic curve analysis showed that the sensitivity and specifificity were 84.4%,82.9%,respectively.Conclusion The diagnostic model of vulnerable plaques based on Clinical and ultrasonic features established by Fisher stepwise discriminant analysis can provide basis for early clinical judgment and diagnosis,has a high application value.