Comprehensive evaluation of cardiac function and detection of myocardial infarction based on a semi-automated analysis using full-volume real time three-dimensional echocardiography

Cole Streiff, Meihua Zhu, Jill Panosian, David Sahn, Muhammad Ashraf

Research output: Contribution to journalArticle

7 Scopus citations

Abstract

Objective Quantitative left ventricular mass (LVM) as well as regional strain values may be obtained from full-volume real time 3D echocardiography data via semi-automated feature tracking and represent indices of heart function, both in health and disease. Methods Fresh adult porcine and ovine hearts were passively pumped to simulate normal cardiac motion at stroke volumes (SVs) varying from 30 to 70 mL. A 3V-D Matrix probe, interfaced with a GE Vivid E9 ultrasound system, was used to image each heart at baseline conditions and after simulated myocardial infarction (MI). The 4D LV quantification function of EchoPAC PC was used to quantify the LVM and longitudinal and circumferential strain (LS & CS) of LV segments at each SV prior and subsequent to simulated MI. LVM was validated by volumetric displacement, while LS and CS values were compared to sonomicrometry-based strain. Results Linear regression analyses show excellent correlations in LVM, LS, and CS between the 4D echo and volumetric/sonomicrometric displacement with R2 values of 0.99, 0.88, and 0.67, respectively. Bland-Altman analyses for all variables validate the compatibility of both methods. It was also determined that EchoPAC PC was able to detect a decrease in LS and CS in the relevant segments between pre- and post-MI at all SVs (P <0.05). Conclusions EchoPAC PC is a robust utility with the ability to accurately obtain quantitative LVM, LS, and CS values from 4D echo volumes and has the potential to improve the yield of clinical studies in cases of suspected MI.

Original languageEnglish (US)
Pages (from-to)332-338
Number of pages7
JournalEchocardiography
Volume32
Issue number2
DOIs
Publication statusPublished - Feb 1 2015

    Fingerprint

Cite this