Left and right ventricular strain using fast strain-encoded cardiovascular magnetic resonance for the diagnostic classification of patients with chronic non-ischemic heart failure due to dilated, hypertrophic cardiomyopathy or cardiac amyloidosis

Abstract

Aims: To compare the ability of left ventricular (LV) and right ventricular (RV) strain measured by fast-strain encoded cardiovascular magnetic resonance (CMR) (fast-SENC) with LV- and RV-ejection fraction for the diagnostic classification of patients with different stages of chronic heart failure (stages A-D based on American College of Cardiology/American Heart Association guidelines) due to non-ischemic cardiomyopathies.

Methods: Our study population consisted of 276 consecutive patients who underwent CMR for clinical reasons, and 19 healthy subjects. Wall motion score index and non-infarct related late gadolinium enhancement (LGE), LV ejection fraction (LVEF) and RV ejection fraction (RVEF) and global LV- and RV-longitudinal (GLS) and circumferential strain (GCS) based on fast-SENC acquisitions, were calculated in all subjects. The percentage of LV and RV myocardial segments with strain ≤ - 17% (%normal LV and RV myocardium) was determined in all subjects.

Results: LVEF and RVEF, LV-GLS, LV-GCS, RV-GLS, RV-GCS and %normal LV- and RV myocardium depressed with increasing heart failure stage (p < 0.001 for all by ANOVA). By multivariable analysis, %normal LV and RV myocardium exhibited closer associations to heart failure stages than LVEF and RVEF (r_partial = 0.79 versus r_partial = 0.21 for %normal LV myocardium versus LVEF and r_partial = 0.64 versus r_partial = 0.20 for %normal RV myocardium versus RVEF, respectively). Furthermore, %normal LV and RV myocardium exhibited incremental value for the identification of patients (i) with subclinical myocardial dysfunction and (ii) with symptomatic heart failure, surpassing that provided by LVEF and RVEF (ΔAUC = 0.22 for LVEF and ΔAUC = 0.19 for RVEF with subclinical dysfunction, and ΔAUC = 0.19 for LVEF and ΔAUC = 0.22 for RVEF with symptomatic heart failure, respectively, p < 0.001 for all). %normal LV myocardium reclassified 11 of 31 (35%) patients judged as having no structural heart disease by clinical and imaging data to stage B, i.e., subclinical LV-dysfunction.

Conclusions: In patients with non-ischemic cardiomyopathy, %normal LV and RV myocardium, by fast-SENC, enables improved identification of asymptomatic patients with subclinical LV-dysfunction. This technique may be useful for the early identification of such presumably healthy subjects at risk for heart failure and for monitoring LV and RV deformation during pharmacologic interventions in future studies.

Keywords: Cardiac magnetic resonance; Fast-strain-encoded MR (fast-SENC); Heart failure; Hypertrophy; Ischemic and non-ischemic cardiomyopathies; Late gadolinium enhancement; Myocarditis.

Fig. 1

Moderate correlations were found between left ventricular (LV) ejection fraction (LVEF) (%) with LV global longitudinal strain (GLS) and global circumferential strain (GCS) and with %normal myocardium for the LV (a–c) and right ventricle (RV) (d–f)

Fig. 2

Strain and non-infarct related late gadolinium enhancement (LGE) analysis by segments: Segments of controls showed the highest longitudinal and circumferential strain values, followed by segments of patients with non-ischemic cardiomyopathy without non-infarct related LGE and then by segments of patients with non-ischemic cardiomyopathy with non-infarct related LGE (a, b). For segmental analysis, the standard deviations of mean segmental strain values are provided. The corresponding polar maps of regional strain values in healthy subjects and in segments of patients with non-ischemic cardiomyopathy without and with LGE, as well as logistic regression analysis for the ability of mean longitudinal and circumferential segmental strain to predict the presence of non-infarct related LGE in a given myocardial segment, are provided in c, d

Fig. 3

LV-ejection fraction (LVEF) and RV-ejection fraction (RVEF), global circumferential strain (GCS), %normal LV and RV myocardium as well as mitral annular plane systolic excursion (MAPSE) and tricuspid annular plane systolic excursion (TAPSE) significantly decreased with increasing heart failure stage (a–h)

Fig. 4

%normal LV myocardium exhibited incremental value for the identification of patients with subclinical LV-dysfunction and for the differentiation of clinical versus subclinical heart failure, surpassing the value of LVEF and MAPSE. All 3 parameters were equal for the identification of patients with refractory stage D heart failure (a–c). %normal RV myocardium exhibited incremental value for the identification of patients with subclinical LV-dysfunction and for the differentiation of clinical versus subclinical heart failure, surpassing the value of RVEF and TAPSE. RVEF exhibited the highest accuracy for the identification of patients with refractory heart failure (d–f)

Fig. 5

Logistic regression analysis demonstrated the incremental predictive role of %normal LV (a) and RV (b) myocardium for the identification of patients with subclinical myocardial dysfunction independent of age, NYHA class and LVEF/RVEF

Fig. 6

Eleven of 31 (36%) patients at stage A, i.e., at risk for heart failure but with no structural or functional abnormalities where re-classified by %normal LV myocardium to stage B, i.e., subclinical LV-dysfunction. LV-dysfunction was detected by LVEF (threshold = 50%) and %normal LV myocardium (threshold = 80%) in 29/140(21%) and 121/140(86%) of patients with subclinical disease and in 55/105(52%) and 104/105(99%) of patients, respectively with overt heart failure

Fig. 7

Bland–Altman plots, exhibiting narrow limits of agreement for global longitudinal strain (GLS) and global circumferential strain (GCS), and for %normal myocardium