Imaging contrast agent concentration and extracellular volume fraction in the right ventricle

Background

Globally increased myocardial extracellular volume fraction (ECVF) has been associated with diffuse myocardial fibrosis. ECVF can be estimated using blood and tissue concentrations of gadolinium contrast agent, [Gd], which are calculated using baseline and post-contrast T₁ values [1]. To date, T₁ quantification has been limited to the left ventricle (LV) with moderate spatial resolution (~2 mm) and long imaging windows (>200 ms) to accommodate breath-hold acquisitions. These methods have insufficient spatial resolution to image the relatively thin-walled right ventricle (RV). A new cine-imaging approach for the measurement of contrast agent concentration and ECVF using saturation-recovery preparation is evaluated in the LV and RV.

Methods

A saturation-recovery gated-segmented cine SSFP sequence, similar to the multi-contrast late enhancement imaging method [2], provides a short acquisition window (< 50 ms) enabling end-systolic imaging and higher spatial resolution (~1 mm). Bloch equation simulations of the sequence were used to generate a look-up table to relate the measured ratio of post- to pre-contrast image intensity to the tissue concentration of contrast agent (CLAIR - Contrast Level Assessment using Intensity Ratios). Short axis images were acquired in 9 subjects from an ongoing study of heart failure (Alberta HEART), with contrast-enhanced images at 15 min post 0.15mmol/kg Gadovist. Typical CLAIR pulse sequence parameters: FOV=300mm, 256 matrix, 8 mm slice, flip angle=73°, TE=1.66ms, TR=3.32ms, VPS=14, TI=300ms. Average LV [Gd] in subjects was compared to values obtained using a saturation-recovery SSFP T₁-mapping sequence [3] calculated using [Gd] = ΔR1/r (ΔR1 = change in 1/T₁ with contrast, r = relaxivity). For both methods ECVF = (1-Hct)*[Gd]Tissue/[Gd]Blood, with [Gd]Blood obtained via the T₁ mapping sequence and an assumed Hct of 0.4. Data are presented as mean±SD and differences compared with the two-tailed paired Student’s t-test.

Results

Subject age was 60.7±14.3yrs, with 5 males. Images from an individual using CLAIR (end-systole) and conventional T₁-mapping (end-diastole) are shown in Fig. 1. LV [Gd] is not statistically different between CLAIR and T₁ mapping (0.188±0.042 vs. 0.198±0.029 mM, p=0.151) and is significantly correlated between methods (p<0.01) (Fig. 2 left). LV ECVF is not statistically different between CLAIR and T₁ mapping (0.216±0.027 vs. 0.231±0.027, p=0.120) with negligible bias (Fig. 2 right) between the two methods. CLAIR RV [Gd] (0.212±0.066 mM) and ECVF (0.245±0.054) were not statistically different from LV values (p=0.134 and p=0.093).

Figure 1

CLAIR image (left) at end-systole (1.17 mm resolution, 46 ms temporal resolution) and conventional T₁-mapping image at end-diastole (right) (1.88 mm resolution, 225 ms temporal resolution) from the same subject (15 minutes post contrast).

Figure 2

Comparison of CLAIR and conventional T₁-mapping methods: LV contrast concentration (left) and Bland-Altman plot of extracellular volume fraction (right).

Conclusions

CLAIR yields similar LV myocardial contrast concentration and ECVF in the LV to T₁-mapping and provides sufficient temporal and spatial resolution for end-systolic RV imaging.

Funding

CIHR.

Alberta HEART Team Grant.