Reproducibility study on myocardial strain assessment using fast-SENC cardiac magnetic resonance imaging

Abstract

Myocardial strain is a well validated parameter for estimating left ventricular (LV) performance. The aim of our study was to evaluate the inter-study as well as intra- and interobserver reproducibility of fast-SENC derived myocardial strain. Eighteen subjects (11 healthy individuals and 7 patients with heart failure) underwent a cardiac MRI examination including fast-SENC acquisition for evaluating left ventricular global longitudinal (GLS) and circumferential strain (GCS) as well as left ventricular ejection fraction (LVEF). The examination was repeated after 63 [range 49‒87] days and analyzed by two experienced observers. Ten datasets were repeatedly assessed after 1 month by the same observer to test intraobserver variability. The reproducibility was measured using the intraclass correlation coefficient (ICC) and Bland-Altman analysis. Patients with heart failure demonstrated reduced GLS and GCS compared to healthy controls (-15.7 ± 3.7 vs. -20.1 ± 1.4; p = 0.002 for GLS and -15.3 ± 3.7 vs. -21.4 ± 1.1; p = 0.001 for GCS). The test-retest analysis showed excellent ICC for LVEF (0.92), GLS (0.94) and GCS (0.95). GLS exhibited excellent ICC (0.99) in both intra- and interobserver variability analysis with very narrow limits of agreement (-0.6 to 0.5 for intraobserver and -1.3 to 0.96 for interobserver agreement). Similarly, GCS showed excellent ICC (0.99) in both variability analyses with narrow limits of agreement (-1.1 to 1.2 for intraobserver and -1.7 to 1.3 for interobserver agreement), whereas LVEF showed larger limits of agreement (-14.4 to 10.1). The analysis of fast-SENC derived myocardial strain using cardiac MRI provides a highly reproducible method for assessing LV functional performance.

Figure 1

The upper two rows demonstrate fast-SENC images acquired in LV two- (A,D), three- (B,E) and four-chamber (C,F) views of a healthy volunteer during baseline and follow-up CMR studies. Acquisition was performed at the same level of the heart. Lower two rows images demonstrate three short-axis views at LV basal (G,J), mid-ventricular (H,K) and apical (I,L) level. LV = left ventricular; CMR = cardiac magnetic resonance.

Figure 2

Demonstrating identical CMR images as in Fig. 1 acquired in a patient with heart failure with reduced LV ejection fraction (HFrEF). CMR = cardiac magnetic resonance; LV = left ventricular.

Figure 3

The images of the upper row demonstrate LV two- (A), three- (B) and four-chamber (C) views at end-systolic phase used to calculate LV volumes, mass-, and ejection fraction as well as LV global and segmental circumferential strain. The lower row images show short-axis views at LV basal (D), mid-ventricular (E) and apical (F) level used to derive global and segmental LV longitudinal strain. LV = left ventricular.

Figure 4

Bland-Altman plots with limits of agreement (1.96 SD) for the intersession (A), interobserver (B) and intraobserver (C) variability as shown in the measurements of LV global longitudinal strain. Similar plots with limits of agreement (1.96 SD) for the intersession (D), interobserver (F) and intraobserver (G) variability as shown in the measurements of LV global circumferential strain. LV = left ventricular.

Figure 5

ROC analysis for determining the optimal value for global longitudinal strain (A) and global circumferential strain (B) in identifying patients with a reduced LV ejection fraction. ROC = received operating characteristic; LV = left ventricular.