Pulmonary arterial (PA) impedance takes in to account pulsatile blood flow through elastic pulmonary arteries as compared to static pulmonary vascular resistance. Increased PA impedance is an early physiological manifestation of PA remodeling. Currently PA impedance can only be detected invasively, is expensive and cumbersome to calculate and thus not done in routine clinical practice. The ability to assess PA impedance non-invasively can provide fundamental insights in evaluation of patients with normal PA pressures or mild pulmonary hypertension (PH) such as in patients with chronic obstructive lung disease. We propose a novel non-invasive parameter, the velocity transfer function (VTF), which is related to PA stiffness and impedance.PA pressure measurements are obtained by invasive techniques for quantification of PA impedance. Velocity and flow, however, can be measured non-invasively with phase-contrast cardiac MR (CMR) at arbitrary points in the PA tree. Figure 1 shows the pulsatile velocity profiles measured at the main PA (MPA) (input) and the proximal right PA (RPA) (output) in a normal human volunteer and a patient with PH. In a normal volunteer, with a compliant PA, the velocity profile not only is shifted in the transit time between the two sites, but there are also complex shape changes across the entire wave. In a patient with PH, the PA is stiffer and both the time shift and shape change are reduced. These time shifts and shape changes are frequency dependent and are related to the compliance and geometry of the artery between the two points. These frequency-dependent changes can be compactly described by the transfer function between the two velocity profiles measured non-invasively with phase-contrast MRI. A transfer function is a function that describes the relationship between the frequency spectra of any two functions that are linearly related. Impedance is a specific case of a transfer function when the two functions are voltage and current or pressure and flow. Here, the PA was modeled between two points as a VTF with the velocity measured at the MPA as the input and the velocity measured at the RPA as the output. Specifically, the VTF $$$[S_F(f)]$$$ was computed between the two measured velocity profiles by taking the Fourier transform of each velocity profile and dividing one by other as follows: $$$S_F(f) = V_{RPA}(f)/V_{MPA}(f)$$$. The VTF is similar to impedance because it describes predominantly the influence of vessel geometry and compliance/stiffness to cause frequency-dependent changes in the input velocity profile as it travels through the artery thereby producing the output velocity profile.The VTF was validated against invasive reference standard (invasive impedance) measured using pressure data from right heart catheterization and velocity data from Doppler echocardiography on 20 patients undergoing clinically indicated right heart catheterization. Each patient also underwent a comprehensive phase-contrast and cine CMR exam to calculate the VTF. Impedance and VTF were computed as a function of heart rate harmonic to compensate for heart rate variability.In Figure 2, impedance curves for patients with normal pulmonary vascular resistance (PVR) £2.5 Woods Units (WU) show a low modulus at zero harmonic, then rapidly descend down with first minimum modulus occurring at low harmonics (1 or 2). Impedance modulus curves of patients with high PVR (>2.5 WU) show high modulus at zero harmonic, and then slowly descends down with first minimum occurring at later harmonics (3 or 4 or higher). This is the expected behavior of impedance curves [1]. Corresponding VTF curves in Figure 2 demonstrate that VTF curves start at similar zero harmonic for patients with normal or high PVR but then shows differentiation at higher harmonics (5 or 6), when the VTF magnitude increases in patients with high PVR. For invasive impedance, the pressure and flow curves have different mean values, which reflects as high impedance moduli at zero and lower impedance harmonics. In contrast, for VTF, the input (MPA) and output (RPA) velocity curves have close to the same mean value, so the zero and lower VTF harmonics are close to 1 before separating at higher harmonics. Representative phase-contrast velocity curves and corresponding VTF for patients in the PVR£2.5 and PVR>2.5 groups are shown in Figure 3. Two patients in the study showed signs of early PA remodeling and were removed as outliers (upper left in Figure 4). The average of VTF magnitude harmonics 5 and 6 in the remaining patients were correlated with impedance harmonics 0 and 1 (Pearson r = 0.75, p=0.002). The VTF has the potential to assess PA impedance non-invasively and reliably using cMRI with potential to non-invasively detect early PA remodeling.