Quantification of regional myocardial perfusion using semiautomated translation-free analysis of contrast-enhanced power modulation images

Quantitative analysis of myocardial perfusion is currently based on manual tracing and frame-by-frame realignment of regions of interest. We developed a technique for semiautomated identification of myocardial regions from power modulation images as a potential tool for quantification of myocardial contrast enhancement. This approach was tested in 13 anesthetized pigs during continuous intravenous infusion of contrast at baseline, left anterior descending coronary artery occlusion, and reperfusion. Regional pixel intensity was calculated for each consecutive end-systolic frame after a high-energy ultrasound impulse, and fitted with an exponential function. Perfusion defects caused by occlusion of left anterior descending coronary artery were confirmed by a significant decrease in both postimpulse steady-state intensity and the initial rate of contrast replenishment (P .05), which were reversed with reperfusion. Automated measurements of myocardial intensity correlated highly with conventional manual tracing (r = 0.90 to 0.97), and resulted in improved signal-to-noise ratios. This technique allows translation-free quantification of regional myocardial perfusion, without the need for manual tracing.