Comparison of Three-Dimensional and Two-Dimensional Speckle-Tracking Longitudinal Strain with Late Gadolinium Enhancement by Cardiac Magnetic Resonance for Left Ventricular Myocardial Fibrosis in Patients with End-Stage Heart Failure

Aims

Left ventricular (LV) strain derived by three-dimensional (3D) speckle-tracking echocardiography (STE) has been demonstrated to be correlated with myocardial fibrosis (MF). However, whether 3D-STE parameters provided a similar estimation of LV MF compared with two-dimensional (2D) STE indices and late gadolinium enhancement (LGE) of cardiac magnetic resonance (CMR) imaging against histological MF remains unknown. The aim of this study was to determine whether 3D-STE parameters were similar to 2D-STE indices and LGE CMR in assessing LV MF in patients with end-stage heart failure (HF) against histological MF.

Methods

Speckle-tracking echocardiography and CMR were performed in 109 patients with end-stage HF who underwent heart transplantation. Left ventricular global longitudinal strain (GLS), global circumferential strain (GCS), and global radial strain (GRS) were obtained from 2D-STE and 3D-STE. The degree of MF was assessed using LGE CMR and Masson’s staining of the LV myocardial samples. One hundred nine patients were divided into 3 groups based on the tertiles of histologic MF.

Results

Myocardial fibrosis was strongly correlated with 3D-GLS (r = 0.70, P < .001), moderately correlated with LGE extent and 2D-GLS (r = 0.64, r = 0.58; P < .001 for each), and weakly correlated with 3D-GRS, 3D-GCS, 2D-GRS, and 2D-GCS (r = −0.40, r = 0.29, r = −0.38, r = 0.25; P < .01 for each). The correlation of 3D-GLS with LV MF was similar to that of LGE with LV MF (0.70 vs 0.64; P = .427). The correlations of segmental 3D longitudinal strain with regional MF were similar to that of segmental LGE with regional MF for both the ventricular septum and LV lateral wall (r = 0.68 vs 0.57, r = 0.71 vs 0.62, P > .05 for both). The positive and negative predictive values, as well as the area under the curve for identifying severe MF, were comparable among 3D-GLS, 2D-GLS, and LGE. The model with 3D-GLS (R2 = 0.53, P < .001; Akaike information criterion = 413) was similar to that with LGE (R2 = 0.50, P < .001; Akaike information criterion = 418) for reflecting the degree of LV MF.

Conclusions

Three-dimensional GLS can be considered a novel functional parameter that may correlate with the extent of LV MF in patients with end-stage HF, demonstrating accuracy comparable to that of 2D-GLS and LGE CMR in identifying severe MF.

Comments (0)

No login
gif