Lung water density is increased in patients at risk of heart failure and is largely independent of conventional CMR measures

Abstract

Aims: Non-invasive methods to quantify pulmonary congestion are lacking in clinical practice. Cardiovascular magnetic resonance (CMR) lung water density (LWD) mapping is accurate, reproducible, and has prognostic value. However, it is not known whether LWD is associated with routinely acquired CMR parameters. Methods and Results: This was a prospective, observational cohort including healthy controls and patients at risk of heart failure. Lung water density was measured using CMR with a free-breathing short echo time 3D cartesian gradient-echo sequence with a respiratory navigator at 1.5T. Associations were assessed between LWD, lung water volume (LWV) and cardiac volumes, left ventricular (LV) mass and function, myocardial native T1 and extracellular volume fraction (ECV). In patients at risk for heart failure (n=155), LWD was greater than in healthy controls (n=15) (30.4{+/-}5.0 vs 27.2{+/-}4.3%, p=0.02). Using receiver operating characteristic analysis, the optimal cut-off for LWD was 27.6% to detect at-risk patients (sensitivity 72%, specificity 73%, positive likelihood ratio 2.7, inverse negative likelihood ratio 2.6). LWD was univariably associated with body mass index (BMI), hypertension, right atrial (RA) area and LV mass (p<0.05 for all). In multivariable linear regression, only BMI remained associated with LWD (R2=0.32, p<0.001). In a separate multivariable analysis including only CMR markers, only RA area remained associated with LWD (R2=0.05, p=0.006). Conclusion: LWD is increased in patients at risk for heart failure compared to controls, and is only weakly explained by conventional CMR measures. LWD provides diagnostic information which is largely independent of conventional CMR measures.