The Incremental Value Of Carotid Plaque For Estimating Pre-Test Probability In Patients With Stable Coronary Artery Disease

Background and Objective In the latest guidelines,calculation of pre-test probability(PTP)has been emphasized as a preliminary procedure to decide the most appropriate diagnosis examination,and the updated Diamond-Forrester method(UDFM)and Duke clinical score(DCS)were suggested to be the advanced solution for calculating PTP.However,several observational studies suggested that both prediction models are unsatisfactory in the patients referred for coronary angiography(CAG)or coronary computed tomography angiography(CTA).In this study,the UDFM model was used as a reference to establish the basic model,and carotid plaque indicators were added it to explore its incremental value in calculating SCAD-PTP.Materials and methods1 Research object and data collection: This study was conducted on 270 inpatients suspected for CAD who underwent both CAG and carotid ultrasound from January 2017 to March 2018 at the Second Affiliated Hospital of Zhengzhou University.The result of CAG and carotid ultrasound,clinical history,laboratory examination and types of chest pain were retrospectively collected.Stable coronary artery disease is defined as the diameter stenosis of one of major coronary arteries more than 50%.All patients were divided into two groups(CAD group and no-CAD group)according to the result of angiography.Carotid plaque parameters including MPT and TNP were directly obtained from carotid ultrasound reports.PS was calculated by summing up all plaque thickness regardless length of each plaque.2 Statistical methods: Differences in patient characteristics were compared with Mann‐ Whitney U-test or chi-square test as appropriate.Multivariate logistic regression model was used to establish prediction models with SCAD as outcome: the basic prediction model including the all UDFM variables: sex,age,type of chest pain,and updated prediction model,which in addition included carotid plaque indictors.The discrimination and calibration ability of models was compared by calculating the area under the receiver operating characteristics(ROC)curve(AUC)and Hosmer-Lemeshow(H-L)test.Reclassification improvement between models was described by categorical net reclassification index(NRI).We also performed decision curve analysis(DCA)to evaluate the clinical benefits.P value was two-sided with a statistical significance level of less than 0.05.Result(1)The univariate analysis demonstrated that sex(p value = 0.018),age(p value < 0.001),hypertension(p value < 0.001),uric acid(p value = 0.019),chest pain(p value < 0.001)and PS,MPT,and TNP(all p value < 0.001)were significantly different between CAD group and non-CAD group.(2)The basic model based on sex,age,types of chest pain,had moderate discrimination ability(AUC = 0.729,95% CI 0.667-0.790)but superior calibration ability(HL test: χ2 = 9.672,p = 0.289).Adding carotid plaque indicators(PS,MTP,TNP)separately to basic model could also increase the AUC(all p value < 0.001)and reclassification ability,but decrease the calibration ability.Additionally,the highest increase of AUC was adding PS variable and NRI was adding TNP.(3)Decision curve showed the higher net clinical benefit of updated model compared with basic model.Conclusion(1)All three carotid plaque indicators were closely correlated with the presence of SCAD.(2)Carotid plaque indicators significantly increase the accuracy of patient’s PTP of SCAD.