Measurement of pulmonary arterial pulse wave reflection from single-slice phase-contrast and steady-state free precession MRI

Summary

Pulmonary arterial hypertension (PAH) is associated with elevated pulmonary vascular resistance, resulting in increased reflection of pressure and flow waves from distal vessels¹. The gold standard for assessing PAH is right heart catheterization, an invasive procedure that carries a 5% risk of major complications². We validate a noninvasive method for quantifying pulmonary arterial reflection using phase-contrast (PC) and steady-state free precession (SSFP) sequences acquired in a single slice.

Background

An arterial segment approximates a hydraulic transmission line terminated distally by a reflection site that partially reflects forward-traveling pressure and flow (q) waves back toward the heart³. Due to finite pulse wave velocity (PWV), backward-traveling waves are minimal in early systole (Figure 1); since arterial cross-sectional area (a) increases roughly linearly with pressure, PWV = ∂q(t)/∂a(t). Combining this with the water hammer equation yields an expression for the backward flow wave⁴:

<center>q_b(t) = [q_meas(t) - PWV×a(t)]/2,</center>from which arterial reflection magnitude (R) can be computed in the frequency domain:

<center>R(ω) = Q_b(ω)/Q_f(ω).</center>

Figure 1

Illustration of single-slice wave separation. Before the onset of backward waves, flow and area are related by the constant factor PWV. Backward-traveling flow and area waves are equal in form but opposite in sign, and are proportional to the difference between PWV × area and flow.

Methods

The right pulmonary artery in three healthy adult volunteers was imaged on a 1.5 T MR system (Siemens, Germany) using retrospectively ECG-gated cine PC and SSFP sequences to quantify blood velocity and vessel cross-section, respectively. PC and SSFP images were co-registered in MATLAB (The MathWorks, USA). The arterial lumen was outlined semi-automatically using Segment (Medviso, Sweden), yielding flow and area time series that were resolved into forward and backward flow waves in MATLAB. The frequency-domain ratios of backward to forward flow waves yielded estimates of R which were then averaged over the fundamental heart frequency and the next two harmonics³ and compared to literature values using a two-tailed Student's t-test.

Results

The single-slice MRI method reliably resolved forward and backward flows in vivo (Figure 2), enabling noninvasive measurement of normal right pulmonary arterial reflection magnitudes, R (SD) = 0.34 (0.05), statistically equivalent (p = 0.74) to invasively measured literature values⁵ of R (SD) = 0.33 (0.13).

Figure 2

Flow wave separation in the right pulmonary artery of Subject 1 calculated from an initial MR study and a repeat scan 36 days later. Reflection magnitudes and standard deviations are tabulated for all subjects. Scan parameters were TR/TE = 9.8/2.2 ms, 192 × 192 matrix, 28 × 28 cm² FOV, 2 averages, flip angle = 10° (PC) and 45° (SSFP). The root-mean-variances of the forward, measured and backward flow waves between scans were 11, 16 and 8 ml/s, respectively.

Conclusions

The feasibility of single-slice MRI measurement of pulmonary arterial reflection in healthy adults motivates follow-up studies in adult and pediatric patient populations and lays the groundwork for noninvasive assessment of pulmonary hypertension.

Funding

This study was supported by the Canadian Institutes of Health Research (App #199854).