Clinically Stable Non-Ischemic Heart Failure with Recovered Ejection Fraction Exhibit Myocardial Disorganization via Diffusion Tensor Imaging

Background: Advances in therapy have enabled many patients with heart failure with reduced ejection fraction (HFrEF) to achieve normalization of left ventricular ejection fraction (LVEF), termed heart failure with recovered ejection fraction (HFrecEF)^1,2. However, whether these patients have achieved true myocardial recovery or are merely in a state of remission remains unclear. Cardiac diffusion tensor imaging (cDTI) provides an assessment of myocardial microstructure and may distinguish remission from recovery. As such, we sought to characterize myocardial microstructural features in clinically stable non-ischemic HFrecEF. We hypothesized that HFrecEF patients would have equivalent cDTI-derived fractional anisotropy (FA) and mean diffusivity (MD) to healthy controls.

Methods: In this cross-sectional study, we enrolled 10 adult patients with clinically stable non-ischemic HFrecEF, defined as baseline LVEF ≤40% followed by ≥2 consecutive LVEF ≥50% over ≥6 months, normalized sex-specific left ventricular internal diameter at end-diastole, and N-terminal pro-B-type natriuretic peptide (NT-proBNP) ≤300 pg/mL (a blood biomarker of cardiac overload). Exclusion criteria included ischemic cardiomyopathy, hypertrophic cardiomyopathy, congenital heart disease, cardiac transplant, and chronic kidney disease. Ten age- and sex-matched healthy adults served as controls. MRI was performed to quantify structure and function, including T1, T2 and cDTI. cDTI was used to derive MD, reflecting extracellular expansion, and FA, reflecting myocardial tissue alignment.

Results: Median age was 58 years (IQR, 55-73) in the HFrecEF group and 59 years (IQR, 56-62) in the control group, with 6 males and 4 females in each group. HFrecEF patients had median LVEF = 57% (IQR, 53-59%) and NT-proBNP = 78 pg/mL (IQR, 37-122). Controls had median LVEF = 59% (IQR, 56-64%). Compared with controls, patients with HFrecEF exhibited significantly higher MD (1.66×10⁻³ vs. 1.51×10⁻³ mm²/s; IQR, 1.60-1.70×10⁻³ vs. 1.45-1.57×10⁻³ mm²/s; p=0.0036) (Figure 1A) and significantly lower FA (0.26 vs. 0.30; IQR, 0.25-0.28 vs. 0.28-0.31; p=0.0022) (Figure 1B). Importantly, overlap existed between groups, with some HFrecEF patients having MD and FA values within the healthy range. There were no significant differences between the HFrecEF and healthy groups with regard to T1 (1257 vs. 1242 ms; IQR, 1257 vs. 1205 ms; p=0.43) and T2 (43 vs. 42 ms; IQR, 41-45 ms; p=0.41) (Figure 2).

Conclusion: Despite normalized LVEF, NT-proBNP, and T1/T2, many patients with clinically stable HFrecEF have abnormal myocardial microstructure. However, a subset demonstrated normal cDTI measures, suggesting potential true myocardial recovery. No significant differences in T1 or T2 were found, so this suggests that neither edema nor fibrosis is present, but rather abnormal microstructure. This is the first study to evaluate MD and FA in clinically stable HFrecEF, highlighting the potential promise of cDTI as a novel biomarker to distinguish remission from recovery.

Figure 1: A) Mean diffusivity in the heart failure with recovered ejection fraction (HFrecEF) patients compared to the healthy group (median, 1.66x10−3 vs 1.51x10−3 mm2/s; IQR, 1.60-1.70 x10−3 vs 1.45-1.57 x10−3; p=0.0036). (B) Fractional anisotropy in the HFrecEF vs healthy group (median, 0.26 vs 0.30; IQR, 0.25-0.28 vs 0.28-0.31; p=0.0022). ** = p<0.001 after Wilcoxon signed-rank test.
Figure 2: A) There was no significant differences in the heart failure with recovered ejection fraction (HFrecEF) patients and healthy group’s T1 (1257 vs 1242ms; IQR 1257 vs. 1205ms, p=0.43). B) There was no significant differences in the HFrecEF patients and healthy group’s T2 (43 vs. 42ms; IQR 41-45ms, p=0.41). n.s. = p>0.05 after Wilcoxon signed-rank test.