Study design
The exclusion criteria were considered as following age below 18 years, a left ventricular ejection fraction <50% based on the 2D echocardiography, a prior history of CAD, myocardial infarction, percutaneous coronary intervention (PCI) or coronary artery bypass graft, acute coronary syndrome (ACS), congestive heart failure, any wall motion abnormality, more than mild valvular heart disease, ventricular conduction disturbances, pathological Q waves in resting ECG, atrial fibrillation, poor image quality for assessing all the segments by speckle tracking, and having refused coronary angiography.
Standard echocardiography and speckle-tracking strain analysis
Upon admission, the patients’ complete medical history was taken, and they underwent a physical examination and 2D echocardiography for speckle-tracking evaluation and conventional transthoracic echocardiography at resting position using Vivid E9 ultrasound (USA, GE Ultrasound). Echocardiographic studies were performed and analyzed by a trained fellow of advanced echocardiography who was blinded to the coronary angiography results.
For longitudinal strain evaluation, three standards (apical two-, three-, and four-chamber) views were recorded from three consecutive beats (Fig. 1) and the offline longitudinal strain analysis was performed using the Automated Function Imaging (AFI) model (GE Ultrasound) by two echocardiologists separately who were expert in strain analysis.
In this method, the LV contour was traced in each view, and after some manual endocardial adjustment, the software automatically gave a longitudinal strain score to each myocardial segment (Fig. 2).
Finally, peak systolic longitudinal strain values were recorded for each segment in the form of 17-segment bull’s eye that was labeled as segmental longitudinal strain (SLS). Global longitudinal strain (GLS) was automatically calculated as the mean value of three apical projections (Figs. 3 and 4).
Territorial longitudinal strain (TLS) was calculated for three major coronary arteries (LAD, LCX, and RCA) by the mean value of SLS in segments perfused by each coronary artery. We applied 17-myocardial segment popular pattern which includes 7 segments related to LAD (basal anterior, basal anteroseptal, mid anterior, mid anteroseptal, anthropical, apical septal, apex cap), six segments related to LCX (basal lateral, basal posterior, mid posterior, mid lateral, apicolateral), and six RCA-related segments (basal septal, inferobasal, mid inferior, mid septal, inferoapical).
Coronary angiography
All the patients underwent coronary artery angiography within 1 week of the echocardiography by an interventional cardiologist, who was blinded to the echocardiography reports. Angiograms were performed via the femoral or radial artery approach, and at least two projections were made for each coronary vessel. Seventy percent and more stenosis in at least one coronary artery, including the left anterior descending (LAD) and its large branches (i.e., diagonals), the left circumflex coronary artery and its large branches (i.e., obtuse marginal [OM] branch), and the right coronary artery (RCA), and 50% and more stenosis in the left main (LM) coronary artery were taken as significant CAD. Stenosis between 50 and 70% in the coronary arteries (except LM) was taken as moderate CAD, and stenosis less than 50% was taken as mild CAD.
Statistical analysis
The results of the 2DSTE and coronary angiograms were statically analyzed in SPSS software (version 25). The T-test was used to evaluate the correlation between the strain scores and significance of CAD. A ROC curve analysis was performed to predict the cut-off point of GLS for the best sensitivity and specificity for predicting significant CAD.