Non-invasive diagnosis of coronary artery disease by quantitative stress echocardiography: optimal diagnostic models using off-line tissue Doppler in the MYDISE study

Abstract

Aims: To develop optimal methods for the objective non-invasive diagnosis of coronary artery disease, using myocardial Doppler velocities during dobutamine stress echocardiography.

Methods and results: We acquired tissue Doppler digital data during dobutamine stress in 289 subjects, and measured myocardial responses by off-line analysis of 11 left ventricular segments. Diagnostic criteria developed by comparing 92 normal subjects with 48 patients with coronary disease were refined in a prospective series of 149 patients referred with chest pain. Optimal diagnostic accuracy was achieved by logistic regression models, using systolic velocities at maximal stress in 7 myocardial segments, adjusting for independent correlations directly with heart rate and inversely with age and female gender (all p<0.001). Best cut-points from receiver-operator curves diagnosed left anterior descending, circumflex and right coronary disease with sensitivities and specificities of 80% and 80%, 91% and 80%, and 93% and 82%, respectively. All models performed better than velocity cut-offs alone (p<0.001).

Conclusion: Non-invasive diagnosis of coronary artery disease by quantitative stress echocardiography is best performed using diagnostic models based on segmental velocities at peak stress and adjusting for heart rate, and gender or age.