Quantitative Gadolinium-Free Cardiac Fibrosis Imaging in End Stage Renal Disease Patients Reveals A Longitudinal Correlation with Structural and Functional Decline

Figures

• 

  Figure 1. Representative ΔS/So maps from an ESRD patient. (A–H) ΔS/So maps from the left ventricular base towards the apex in a patient with ESRD demonstrate a gradient of elevated ΔS/So across slice positions. Elevated values consistent with dense scar appear in the septal mid-wall at the base of the left ventricle (A–C). At the mid-ventricle (D,E) ΔS/So values are consistent with diffuse fibrosis in the septum and normal tissue in the free wall. Towards the apex (G,H) ΔS/So values are consistent with healthy tissue throughout the myocardium. (I) The corresponding bulls-eye plot illustrates patterns of elevated mean ΔS/So, particularly in septal segments. Global mean ΔS/So was 170% and divergence was 37.0 AU.

• 

  Figure 2. MT-weighted CMR for Tissue Characterization. End diastolic cine bSSFP images acquired with a 45° excitation flip angle in a (A) representative healthy control and (E) patient with ESRD are paired with images acquired at excitation flip angle of 5° (B and F, respectively). (C,G) Maps of ΔS/So are generated as ΔS/So = (S45 − S5)/S5 * 100 (%). (C) ΔS/So maps from healthy controls demonstrate low and uniform ΔS/So values. (G) Elevations in ΔS/So observed in patients with ESRD are consistent with those seen in fibrotic tissue (see12). (D,H) Segmental average ΔS/So values from LV base (outer ring) to apex (inner ring) are represented using a bullseye plot. Compared to uniformly low ΔS/So values of a healthy control (D), ΔS/So demonstrates regional heterogeneity in a patient with ESRD, with elevated values primarily in the septum (H). (I) Using all LV myocardial ΔS/So values from the control group, a standard cumulative distribution function was generated for this study (black curve). The simulated ΔS/So distribution was dynamically resized to match the number of voxels per individual LV and compared to the distribution of ΔS/So values for that individual. Cumulative distributions from…

• 

  Figure 3. Focal fibrosis development. (A–C) Images acquired in a 42 year old female patient on hemodialysis for approximately two years with plural effusion demonstrate minor signal elevations but (D) low average ΔS/So across the left ventricle. (E–H) While global divergence increased by only 8.7 AU at follow-up, regional elevations are evident at the basal and mid ventricular septum. The elevated values are consistent with scar tissue as observed in Stromp et al.12. Imaging of the same basal/mid-ventricular slice (B,E) highlights the ability to detect the development of mid-wall enhancement between visits.

• 

  Figure 4. Patients with ESRD demonstrated preserved diastolic circumferential strain rate and reduced global longitudinal strain compared to controls. (A) Circumferential diastolic strain rate was similar between ESRD (162.8 ± 41.7%/sec) and control groups (174.1 ± 33.9%/sec, p = 0.247). (B) Peak global longitudinal strain (ELL) was attenuated in ESRD (−20.5 ± 4.1%) compared to controls (−22.9 ± 3.5%, p = 0.035). Gray bars indicate group means, *p < 0.05.

• 

  Figure 5. Association of divergence with structural and functional measurements. (A) Divergence was moderately correlated with LVMI (ρ = 0.31, p = 0.014), however, among severely hypertrophic individuals (LVMI >100 g/cm2) divergence was highly variable. (B) Neither diastolic strain rate (ρ = 0.06, p = 0.646) nor (C) global longitudinal strain (ρ = 0.02, p = 0.910) correlated with divergence.

• 

  Figure 6. Various spatial patterns of elevated ΔS/So were observed among hypertrophic patients with ESRD. (A–D) Representative end diastolic images acquired using a 45° excitation flip angle at the mid-ventricle are shown for 4 hypertrophic patients with ESRD. (E–H) Corresponding ΔS/So maps reveal different patterns and magnitudes of ΔS/So elevation despite the common structural feature of hypertrophy. An expanded version of this figure including 12 patients can be found in the Supplemental Data.

• 

  Figure 7. Baseline levels of fibrosis and global longitudinal strain correlated with corresponding parametric changes over one year. Patients with higher baseline divergence values demonstrated greater loss of (A) global longitudinal strain (ΔGLS) and (B) global longitudinal diastolic strain rate (ΔGLDSR) over one year. (C) Interestingly, there was also a correlation between attenuated global longitudinal strain at baseline and subsequent increases cardiac fibrosis over one year. Pearson’s correlation coefficients and statistical values can be found in Table 3.

• 

  Figure 8. Hypertrophy correlated with attenuation in cardiac mechanics. (A) Increases in LVMI over one year strongly correlated with decreases in circumferential diastolic strain rate (r2 = 0.58, p = 0.007) and (B) with attenuation in global longitudinal strain (r2 = 0.38, p = 0.043).