[How to Apply Bayesian Theorem to the Evaluation of Myocardial Injury by Measuring High Sensitive Cardiac Troponins in the Patients with Suspected Acute Myocardial Infarction].

118 consecutive patients of suspected acute myocardial infarction with acute chest pain and shortness of breath visiting our emergency room were subjected for this clinical study. Based on final diagnosis of acute myocardial infarction (AMI) comprehensively determined by medical record, physical examination, ECG, echocardiography, cardiac catheterization, etc., except for cardiac biomarkers, the patients were classified into two groups, with AMI group (1) and without AMI group (0) and then ROC curve analysis was performed between without AMI group (1) and with AMI group (0). As a result of ROC curve analysis, AUC, cutoff value, sensitivity, specificity and likelihood ratio (LR) were calculated as shown in Fig. 4 (1-7) and Table 2 (1-7). Based on calculating equation led from Bayesian rules, post-test odds were calculated as product of pre-test odds and LR at the cutoff value in each biomarker such as hsCTnT, hsCTnI, h-FABP CK, CKMB activity and CKMB mass. As a result, post-test probability was improved from predictive pre-test probability 30% to post-test probability 89% and 86% in hsCTnT and hsTnI, respectively but less improved from 30% to 68% in h-FABP and unexpectedly improved from 30% to 82% in CKMB mass compared with hsCTnT and hsTnI. Based on Bayesian rule, it is very valuable to predict post-test probability from predictive pre-test probability 30% by calculation in particular, when post-test probability is over 85-90%. In conclusion, we believe that prediction of post-test probability by Bayesian rule can be surely used to evaluate clinical quality of biomarkers which are not depend on at least, specialty and experience of physicians.