To Establish And Validate Nomogram Model For Predicting New-onset Atrial Fibrillation In Patients With Embolic Stroke Of Undetermined Source

Purpose:To investigate the clinical significance of a complex model based on hematology and imaging indicators in predicting new atrial fibrillation(AF)in patients with embolic stroke of undetermined source(ESUS).Method:In the study,151 patients with ESUS were included and divided into 29 patients in the NOAF group and 122 patients in the non-NOAF group according to the occurrence of atrial fibrillation(AF)during follow-up.Basic information,clinical features,laboratory examinations and imaging examinations of all subjects were collected in hospital or outpatient.Logistic regression equation was established to analyze the independent risk factors of NOAF in ESUS patients.A nomogram model was constructed based on the results of the Logistic regression risk model,and the consistency index(C-index)was calculated.The predictive ability of this model was evaluated by receiver operating characteristic(ROC)curve and calibration curve.Results:After the measurement data is truncated with the median,there were statistically significant differences in age,BMI,Lp(a),ALB,PLR,NTpro BNP,history of coronary artery disease,left atrial inner diameter(LAD)and left ventricular end-diastolic inner diameter(LVEDD)between NOAF group and non-NOAF group(P<0.05).Binary logistic regression analysis showed that age > 65 years old,Lp(a)>57 mg/dl,ALB≦3.16g/dl,NTpro BNP > 590ng/L,LAD > 38 mm,LVEDD > 50 mm were independent risk factors for AF in ESUS patients(P<0.05).Based on the above six independent risk factors,a line graph model was established for the risk of AF in ESUS patients.The C-index was as high as 0.873,95%CI: 0.791-0.955.The ROC curve was used to evaluate the effect of the curve,the area under curve(AUC)was0.873,and the calibration curve was closed to an ideal curve with a slope of 1.Conclusion:Multivariate Logistic regression was used to construct the nomogram risk model for new AF in ESUS patients,and the C-index was 0.873,indicating that this model has strong predictive power for new AF in ESUS patients.The robustness of the model was tested by the method of internal data set validation,which proved that the robustness of the model was good and could be applied to clinical practice.