RESUMO
Introduction: Coronary angiography (CAG) is invasive and expensive, while numbers of patients suspected of coronary artery disease (CAD) undergoing CAG results have no coronary lesions. Aim: To develop machine learning algorithms using symptoms and clinical variables to predict CAD. Material and methods: This study was conducted as a cross-sectional study of patients undergoing CAG. We randomly chose 2082 patients from 2602 patients suspected of CAD as the training set, and 520 patients as the test set. We utilized LASSO regression to do feature selection. The area under the receiver operating characteristic curve (AUC), confusion matrix of different thresholds, positive predictive value (PPV) and negative predictive value (NPV) were shown. Support vector machine algorithm performances in 10 folds were conducted in the training set for detecting severe CAD, while XGBoost algorithm performances were conducted in the test set for detecting severe CAD. Results: The algorithm of logistic regression achieved an average AUC of 0.77 in the training set during 10-fold validation and an AUC of 0.75 in the test set. When probability predicted by the model was less than 0.1, 11 patients in the test set (520 patients) were screened out, and NPV reached 90.9%. When probability predicted by the model was less than 0.2, 110 patients in the test set were screened out, and reached 83.6%. Meanwhile, when threshold was set to 0.9, PPV reached 97.4%. When the threshold was set to 0.8, PPV reached 91.5%. Conclusions: Machine learning algorithm using data from hospital information systems could assist in severe CAD exclusion and confirmation, and thus help patients avoid unnecessary CAG.
RESUMO
BACKGROUND: Inferior wall ST-segment elevation myocardial infarction (STEMI) is mostly caused by acute occlusion of right coronary artery (RCA) and left circumflex artery (LCX). Several methods and algorithms using 12-lead ECG were developed to localize the lesion in inferior wall STEMI. However, the diagnostic properties of these methods remain under-recognized. AIMS: The aim of this meta-analysis is to compare the diagnostic properties among the methods of identifying culprit artery in inferior wall STEMI using 12-lead ECG. METHODS: We performed a meta-analysis to calculate the pooled sensitive, specificity, area under the curve (AUC) and diagnostic odds ratio (DOR) of each method. RESULTS: Thirty-three studies with 4414 participants were included in the analysis. Methods using double leads had better diagnostic properties, especially ST-segment elevation (STE) in III > II [with pooled sensitivity 0.89 (0.84-0.93), specificity 0.68 (0.57-0.79), DOR 17 (9-32), AUC 0.88 (0.85-0.91)], ST-segment depression (STD) in aVL > I [with pooled sensitivity 0.82 (0.72-0.90), specificity 0.69 (0.48-0.86), DOR 11 (4-29), AUC 0.85 (0.81-0.88)], and STD V3/STE III ≤1.2 [with pooled sensitivity 0.88 (0.78-0.95), specificity 0.59 (0.42-0.75), DOR 12 (5-27), AUC 0.82 (0.78-0.85)]. Diagnostic algorithms, including Jim score[pooled sensitivity 0.70 (0.55-0.85), specificity 0.88 (0.75-0.96)], Fiol's algorithm [pooled sensitivity 0.54 (0.44-0.62), specificity 0.92 (0.88-0.96)] and Tierala's algorithm [pooled sensitivity 0.60 (0.49-0.71), specificity 0.91 (0.86-0.96)], were not superior to these simple methods. CONCLUSIONS: Our meta-analysis indicated that diagnostic methods using double leads had better properties. STE in III > II together with STD in aVL > I may be the most ideal method, for its accuracy and convenience.
