Patients with end stage renal disease (ESRD) suffer from high rates of sudden cardiac death, often attributed to the development of reactive fibrosis and arrhythmias. The contraindication of ESRD patients to gadolinium-based contrast agents precludes the accurate detection of cardiac fibrosis via late gadolinium enhancement cardiac MRI (CMR). In a prior study we demonstrated a strong association between ΔS/S_o values measured using magnetization transfer weighted 2-point bSSFP and signal enhancement on late gadolinium enhancement CMR¹. In this study we utilize 2-point bSSFP to assess cardiac fibrotic burden and examine its association with measurements of contractile function and cardiac structure.

In this ongoing study, 11 patients on routine hemodialysis for ESRD and 9 healthy controls completed CMR on a 1.5T Siemens Aera scanner (Erlangen, Germany). Pairs of prospectively gated cine bSSFP images were acquired at flip angles of 5º and 45° from left ventricular base to apex [TR/TE= 3.2/1.36 ms, FOV= 260x260 mm², Matrix= 256x256, Thickness= 8mm, phases set to fill the cardiac interval]. Maps of ΔS/S_o were generated as ΔS/S_o=(S₄₅-S₅)/S₅*100 (%), where S_i represents signal intensity per voxel at flip angle i (see Figure 1A,B). The distribution of ΔS/S_o values across all healthy controls was used to define a reference standard healthy cumulative distribution function (Figure 1C). To account for variations in heart size, this distribution was dynamically resized to generate a cumulative distribution function of ΔS/S_o values matched to each individual heart by size. Each subject’s observed ΔS/S_o distribution was compared to the appropriately-sized standard using a one-sided Kolmogorov-Smirnov (KS) test, yielding a divergence value that represents the degree of rightward shift (elevated signal) from the standard (example in Figure 1C). Global circumferential and longitudinal strains and strain rates were analyzed from 45º flip angle images using a custom feature tracking algorithm². Isolation of strain vs. time curves from the septum and free wall were used to calculate septo-lateral dyssynchrony as the absolute difference in time to peak (as a percent of cardiac cycle) contraction between the two regions. Accounting for heart rate variability, global time to peak contraction was calculated as a percentage of the cardiac cycle.

ESRD patients displayed greater divergence from the standard ΔS/S_o distribution (15.5 ± 12.9%) than controls (5.4 ± 4.6%, p=0.03), as shown in Figure 2A. Circumferential peak strain (Figure 2B), systolic and diastolic strain rates (Figure 2C), and longitudinal strain and strain rates were all similar between groups. Time to peak circumferential contraction was significantly delayed in ESRD patients (52.2 ± 7.9% of cardiac cycle) compared to controls (42.1 ± 6.1%, p=0.01). Septo-lateral dyssynchrony trended higher in ESRD patients (3.9 ± 1.6%) compared to controls (2.8 ± 0.7%, p=0.06). QRS complex duration measured by 12-lead EKG was within normal range for all but 1 ESRD patient and did not correlate with divergence values. While ESRD patients displayed greater left ventricular mass indexed to body surface area (69.3 ± 21.8 g/m²) and septal to ventricular internal radius ratio (0.5 ± 0.1) compared to controls (49.7 ± 9.2 g/m², p=.03 and 0.4±.1, p<.01, respectively), neither demonstrated an association with divergence values. Linear regression modeling revealed a moderate correlation between time to peak and divergence (R²=0.44, p<.01) as well as dyssynchrony and divergence (R²=0.39, p<.01) across groups (Figure 3).

In this ongoing study, ESRD patients demonstrate heightened ΔS/S_o values compared to healthy controls, which is consistent with a greater fibrotic burden. In order to assess global fibrotic burden we calculate the divergence of ΔS/S_o values in each individual against a healthy standard and demonstrate that divergence is increased in ESRD patients even in the absence of changes to global contractile function. We did, however, observe an association between divergence and septo-lateral dyssynchrony, suggesting that combined assessment could be useful for the evaluation of cardiac risk in ESRD patients or the investigation of future anti-fibrotic therapies.

The varied and often diffuse pattern of cardiac fibrosis coupled with the inability to perform contrast-based CMR poses an ongoing challenge for the accurate identification of fibrosis and heart failure risk in ESRD patients. Combining non-contrast 2-pt bSSFP CMR for tissue characterization with analysis of regional contractile function offers a promising approach to identify potential MRI biomarkers of cardiac risk in ESRD.