Comparison of real-time three-dimensional speckle tracking to magnetic resonance imaging in patients with coronary heart disease

Abstract

This study compared strain values from 2-dimensional (2D) and real-time 3-dimensional (3D) speckle tracking with hyperenhancement transmural extent by magnetic resonance imaging (MRI). The study included 18 control subjects (mean age 51 ± 10 years) and 25 patients (20 men, mean age 62 ± 16 years) with ischemic left ventricular (LV) dysfunction (mean LV ejection fraction 41 ± 9%) referred for viability assessment using MRI. Longitudinal, radial, and circumferential strain values were computed using 2D speckle tracking. From analysis of 3D speckle tracking, conventional strain markers (longitudinal, radial, and circumferential) and 2 new 3D strain indexes (area and 3D strains) were obtained from apical view 3D datasets. A hyperenhancement transmural extent segment (16-segment model) was defined as delayed contrast enhancement >50%. Overall, 661 of 688 segments (96%) were analyzable by MRI and 3D speckle tracking. All 3D strain components in hyperenhancement transmural extent segments (n = 154) were lower than in nontransmural necrosis (n = 219) and control (n = 288) segments. Longitudinal strain by 3D, but not by 2D, differentiated nontransmural segments with scar <25%. All 3D global strain indexes correlated with LV ejection fraction (r(2) = 0.67 to 0.26, p <0.05 for all comparisons), whereas only area, longitudinal, and circumferential 3D strains correlated with global scar extent. The best reproducibility was provided by 3D longitudinal (6%) and area (8%) strains. In conclusion, longitudinal and area strains by 3D speckle tracking provide an accurate and reproducible measurement of myocardial deformation that correlate with infarct size in patients with ischemic LV dysfunction.