Abstract

Analysis of left ventricular (LV) regional wall motion (RWM) is subjective and may be challenging in patients with suboptimal images, even with contrast enhancement. It was hypothesized that the amplitude and timing of RWM obtained from contrast-enhanced echocardiograms can be accurately represented in still-frame parametric images. This study was designed to (1) test this hypothesis, (2) establish the diagnostic value of these images as an aid for inexperienced readers, and (3) test the feasibility of automated quantitative analysis of RWM. Contrast-enhanced apical 4-chamber, 2-chamber, and 3-chamber LV views were acquired in 45 patients with poor acoustic windows. The interpretation of dynamic images by an experienced reader who classified RWM as normal or abnormal was used as a reference for comparisons against (1) visual interpretation of parametric images, (2) interpretation of dynamic images by two inexperienced readers (American Society of Echocardiography level I) without and subsequently with parametric images, and (3) automated quantification of RWM. Expert readers detected abnormal RWM in 30 patients (437 of 945 segments). Visual interpretation of parametric images showed good agreement with the reference (sensitivity, 85%; specificity, 82%; accuracy, 84%). The interpretations by inexperienced readers improved with the addition of parametric images, with increases in specificity (from 58% to 79%) and accuracy (from 74% to 84%), despite a slight decrease in sensitivity (from 92% to 91%). Automated classification was feasible and accurate (sensitivity, 82%; specificity, 78%; accuracy, 80%). Parametric images derived from contrast-enhanced echocardiograms of patients with poor acoustic windows accurately depicted RWM, improved the diagnostic accuracy of inexperienced readers, and allowed the objective detection of RWM abnormalities.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call