Mechanisms of left ventricular systolic dysfunction in light chain amyloidosis: a multiparametric cardiac MRI study

Abstract

Background: Cardiac systolic dysfunction is a poor prognostic marker in light-chain (AL) cardiomyopathy, a primary interstitial disorder; however, its pathogenesis is poorly understood.

Purpose: This study aims to analyze the effects of extracellular volume (ECV) expansion, a surrogate marker of amyloid burden on myocardial blood flow (MBF), myocardial work efficiency (MWE), and left ventricular (LV) systolic dysfunction in AL amyloidosis.

Methods: Subjects with biopsy-proven AL amyloidosis were prospectively enrolled (April 2016-June 2021; Clinicaltrials.gov ID NCT02641145) and underwent cardiac magnetic resonance imaging (MRI) to quantify rest MBF by perfusion imaging, LV ejection fraction (LVEF) by cine MRI, and ECV by pre- and post-contrast T1 mapping. The MWE was estimated as external cardiac work from the stroke volume and mean arterial pressure normalized to the LV myocardial mass.

Results: Rest MBF in 92 subjects (62 ± 8 years, 52 men) with AL amyloidosis averaged 0.87 ± 0.21 ml/min/g and correlated with MWE (r = 0.42; p < 0.001). Rest MBF was similarly low in subjects with sustained hematologic remission after successful AL amyloidosis therapy (n = 21), as in those with recently diagnosed AL amyloidosis. Both MBF and MWE decreased by ECV tertile (p < 0.01 for linear trends). The association of ECV with MWE comprised a direct effect (84% of the total effect; p < 0.001) on MWE from adverse interstitial remodeling assessed by ECV and an indirect effect (16% of the total effect; p < 0.001) mediated by MBF. There was a significant base-to-apex gradient of rest MBF in subjects with higher amyloid burden.

Conclusions: In AL amyloidosis, both MBF and MWE decrease as cardiac amyloid burden and ECV expansion increase. Both structural and vascular changes from ECV expansion and myocardial amyloid burden appear to contribute to lower MWE.

Keywords: cardiac MRI (CMR); cardiac amyloidosis; extracellular volume; myocardial blood flow; myocardial work efficiency.

Figure 1

Example of the CMR assessment of perfusion and T1 mapping in a patient with AL-CMP. (A) For perfusion imaging at rest, three short-axis slices were imaged at every heartbeat during the first pass of a gadolinium contrast agent. Perfusion images are shown for the mid-LV level slice. T1 mapping was performed before contrast injection (“native T1”) and at 10 and 20 min after injection of the contrast agent at slice locations matching the perfusion assessment. Cine imaging of the LV indicated mild LV hypertrophy (mid-slice LV wall thickness = 15 mm). (B) Endo- and epicardial contours were drawn in motion-corrected first-pass perfusion images (red and green contours in the perfusion images above) to generate signal–intensity vs. time curves for a region in the blood pool and myocardial segments. The signal intensity curves were converted into curves showing the change of R1 with time. (C) Myocardial R1 vs. time curves were fit with the arterial input measured at the basal level by model-independent deconvolution. The solid curve shows the myocardial response calculated from the estimated impulse response and the measured arterial input. Based on Ziegler's central volume principle, the amplitude of the impulse response estimates the myocardial blood flow.

Figure 2

(A) Rest myocardial blood flow, (B) rest left ventricular ejection fraction, and (C) myocardial work efficiency stratified by ECV tertiles. (A) Myocardial blood flow decreased by tertiles of increasing ECV (p = 0.004 for linear trend). (B) Similarly, LVEF decreased with ECV (p < 0.001 for linear trend) and (C) for myocardial work efficiency defined as external work (stroke volume×mean arterial pressure) divided by left ventricular mass. The p-values in brackets are from the t-tests and adjusted for multiple comparisons. ECV, extracellular volume; LVEF, left ventricular ejection fraction.

Figure 3

(A) Correlation of myocardial work efficiency (MWE) to rest myocardial blood flow (MBF) and (B) NYHA classification. (A) MBF correlated positively with the MWE, calculated as external work (mean arterial pressure×stroke volume×heart rate) divided by myocardial mass (r = 0.42; p < 0.001). The correlation in the AL-CMP group was R = 0.39, p < 0.001 with N = 75 patients and R = 0.18, p = 0.5 in the AL non-CMP group with N = 17 patients. (B) MWE decreased with the NYHA class. NYHA, New York Heart Association Functional Classification of Heart Failure.

Figure 4

(A) Coefficient estimates from a multivariable linear regression model for myocardial work efficiency (MWE) with rest MBF and ECV (as tertiles) as predictors. MWE was positively associated with rest MBF, reflecting the demand–supply relationship between external cardiac work normalized by LV mass and MBF and negatively associated with ECV, a surrogate marker of amyloid burden. *p < 0.05; ***p < 0.001. (B) Estimates with and without normalization, with a mediation model of direct and indirect effects of extracellular volume (ECV) on myocardial work efficiency (MWE), with rest MBF (MBF) as the mediator.

Figure 5

Correlation of segmental rest myocardial blood flow (MBF) with segmental ECV by myocardial segment based on the standard American Heart Association (AHA) segmentation model for LV. Correlation coefficients were calculated by Pearson's method.

Figure 6

Segmental rest myocardial blood flow (MBF) at the apex, middle, and basal regions of the left ventricle. Each data point represents the average of rest MBF at the basal, middle, and apical slice levels, respectively, with data points from the same patient connected by dotted lines. Rest MBF increased significantly from the base to the apex in the two upper tertiles of ECV. The p-values for paired comparisons (t-test), shown above the brackets, were adjusted by Holm's method. A linear mixed-effects model for segmental rest MBF predicted a negative fixed effect of ECV (p = 0.015) and significantly higher MBFs in the middle (+0.66 ml/min/g; p < 0.001) and apical slices (+0.11 ml/min/g; p < 0.001) compared to the basal slice (0.82 ml/min/g).

Figure 7

The presence of LGE was associated with the extracellular volume. (A) The presence of LGE in the LV was associated with global ECV. In a logistical regression model for LV LGE, each 1% change of ECV increased the odds of LV LGE by 1.55 (95% CI: 1.3–2.1; p < 0.001). The solid line represents the logistical regression curve estimate. (B) The proportion of patients with LV LGE increased by ECV tertile (p < 0.001; chi-squared test).