Speckle analysis of 3D echocardiography improves information on left ventricle (LV) segmental and global deformation by avoiding loss of speckles as it is the case in monoplane 2D analysis. Our goal was to evaluate the accuracy of 3D deformation parameters to detect myocardial delayed enhancement (MDE) transmural extent by cardiac magnetic resonance imaging (CMR) in myocardial infarction (MI).

Patients .– We included 72 patients (57.3 ± 12.4 yo) with first acute MI who underwent within 3 days following revascularization both CMR (GE 3T) and echocardiography (GE Vivid E9) including a 3D acquisition of full LV volume. Furthermore, 114 normal subjects (54.9 ± 12.0 yo) underwent a complete echocardiography. Automated analysis of 3D allowed the calculation of 3D global LV area (3DGAS), longitudinal (3DGLS), circumferential (3DGCS) and radial (3DGRS) strains (S%). Peak systolic 2D and 3D S values from the 17 LV myocardial segments were recorded. For each segment MDE was defined as transmural (MDE > 66%), intermediate (33–66%) and subendocardial (< 33%). Pearson was used to study correlation between 2D, 3DTTE and CMR measurements. ROC analysis identified strain cut-off value predicting scar extent.

Results .– The 72 MI pts show a slightly decreased CMR-LVEF (46.3 ± 9.5%) with a small infarct size (global scar extent 20 ± 13%). CMR identified 864 non-infarcted segments (75.2%) and 188 segments with transmural (16.2%), 80 with intermediate (7.0%) and 18 with subendocardial MDE (1.6%). A good tracking quality was obtained respectively in 87% and 93% of the segments in control and MI pts with good inter observer reproducibility (ICC 0.824 for 3DGLS and 0.945 for 3DGAS). All S values were significantly higher in control than in MI pts (3DGAS: −34.8 ± 3.2 vs. −26.0 ± 5.8; 3DGLS:−20.3 ± 2.7 vs. −14.1 ± 3.7; 3DGCS:−18.7 ± 2.9 vs. −14.8% ± 3.8; 3DGRS: 54.6 ± 9.9 vs. 39.3 ± 11.8, P < .0001). All 3DGS values were correlated with CMR-LVEF (3DGLAS r = −0.715; 3DGLS r = 0.602; 3DGCS r = −0.64; 3DGRS r = 0.66; 2DGLS r = −0.652; all P < 0.0001). All 3D S values were significantly different between non-infarcted, subendocardial, intermediate and transmurally infarcted segments ( P < 0.0001) and were significantly lower in non-infarcted segments of MI patients than in segments of control pts. The optimal cut-off value for segmental 3DAS to predict a transmural extent vs. control was −26% with a sensitivity of 86.7% and a specificity of 85.6% (AUC:0.92).

3D speckle imaging is an interesting tool in the acute phase of MI and 3D area strain seems the most valuable parameter, both as a global marker of LV dysfunction and as a regional marker of transmural scar.