Characterizing extravascular lung water-A dual-contrast agent extracellular volume approach by cardiovascular magnetic resonance.

Background: Pathological extravascular lung water is a facet of decompensated congestive heart failure that current cardiovascular magnetic resonance (CMR) methods fail to quantify. CMR can measure total lung water density, but cannot distinguish between intravascular and extravascular fluid, and thus is not diagnostic. Therefore, we develop and evaluate a novel method to measure extravascular lung water by distinguishing intravascular from extracellular fluid compartments using two different contrast agents, extracellular gadolinium chelates and iron-based intravascular ferumoxytol.

Methods: We created two porcine models of pulmonary edema: reversible catheter-induced mitral regurgitation to induce extravascular lung water (n = 5); intravascular volume overload using rapid colloid infusion (n = 5); and compared to normal controls (n = 8). We sequentially acquired lung T1 maps and lung water density maps at 0.55T with native, gadolinium-based, and ferumoxytol contrast, from which we calculated the extracellular volume fraction (ECV) and blood plasma volume fraction in the pulmonary tissue, respectively. We computed extravascular ECV as the difference in ECV and plasma volume fractions. Extravascular lung water volumes were estimated.

Results: In the mitral regurgitation model, baseline vs mitral regurgitation ECV_{extravascular} increased from 27 ± 4.1% to 32 ± 1.9% (p = 0.006), and extravascular lung water volume increased from 105 ± 19 mL to 143 ± 15 mL (p = 0.048). Plasma volume fraction was similar at baseline vs mitral regurgitation (43 ± 4.2% vs 46 ± 5.4%, p = 0.26). Compared to naïve pigs, we measured higher plasma volume fractions in the intravascular volume-loaded model (42 ± 4.7% vs 51 ± 2.7%, p = 0.0054), but no differences in ECV_{extravascular} (21 ± 4.6% vs 21 ± 3.6%, p = 0.99) or extravascular lung water volume (67 ± 13 mL vs 89 ± 24 mL, p = 0.11). Assessing the regional distribution, the plasma volume was higher posteriorly, indicating gravitational dependency, whereas, the extravascular lung water was higher anteriorly.

Conclusion: Extravascular lung ECV measurements and derived lung water volumes corresponded well with predicted increases in extravascular and intravascular pulmonary fluid in animal models. This method may enable mechanistic studies of lung water in patients with dyspnea.