Two separate study populations were recruited:

Young healthy volunteers (YHV): Ten healthy volunteers under 40 (3 Males, 7 females, Mean age 31.5 ± 2.4 years) with no history of cardiovascular or lung disease were recruited to the study. All individuals underwent high temporal resolution phase contrast scans of their main pulmonary artery and branch pulmonary arteries at baseline, during exercise, and at 6 months follow-up.

Older healthy volunteers (OHV): 20 healthy volunteers over the age of 55 (9 male, 11 female, mean age 60.2 ± 1.1 years) with no history of cardiovascular or lung disease were recruited to the study.  All individuals underwent high temporal resolution phase contrast scans of their main pulmonary artery and branch pulmonary arteries followed by a high spatial resolution phase contrast scan of their main pulmonary artery. All 4 sequences were repeated during the same scanning session.

Ethical approval was granted by the East of Scotland Ethics committee 1.  All participants gave written informed written consent for the study.

MRI

Images were acquired on a 32 RF cardiac receiver channel, 3 Tesla MRI scanner (Magnetom Trio, Siemens, Erlangen, Germany).  A three-plane localiser was first obtained, following which 4 chamber, 2 chamber and short axis localisers of the heart were obtained. An axial half-Fourier acquisition turbo spin echo (HASTE) stack was acquired of the chest.   From these a balanced steady state free precession (bSSFP) of the right ventricular outflow tract was planned following which an orthogonal plane was acquired to optimally visualise the main pulmonary artery and valve.  Localisers along the length of the right and left pulmonary artery were then obtained.  From these, phase contrast imaging was acquired in three planes through the main pulmonary artery (MPA), right pulmonary artery (RPA) and left pulmonary artery (LPA)  to provide a true cross section through each of the three arteries.  The MPA slice was located as close to the valve as possible in order to maximise the distance for the TT technique while also avoiding the valve throughout the cardiac cycle.  The RPA and LPA were placed as close to the hila as possible while remaining proximal to the origins of the first visualised branch.

For the high temporal resolution scan the acquisition parameters of the phase contrast sequence were as follows: Slice thickness=8 mm, TR/TE=7/4 ms, no. averages=1, phases=128, velocity encoding=150 cm/s, bandwidth/pixel=340 Hz, flip angle=15°, field of view (FOV) =320 × 320mm2, matrix=256 × 256.  For the high spatial resolution scan, the acquisition parameters were: Slice thickness=8 mm, TR/TE=12/4 ms, no. averages=1, phases=80, velocity encoding=150 cm/s, bandwidth/pixel=340 Hz, flip angle=15°, field of view (FOV) = 512 × 512mm2, matrix=256 × 256. Both sequences were free breathing, with an acquisition time of approximately 4 minutes depending on heart rate.

In addition, both YHVs and OHVs also underwent aortic PWV assessment of their aortic arch.  The high temporal resolution acquisition sequence was performed at the level of the RPA with the distance between the ascending and descending aorta measured on a candycane view of the thoracic aorta as previously described.[27]

Exercise

Isometric exercise was performed by the volunteer by crossing their feet and then forcefully plantarflexing the superior foot while dorsiflexing the inferior foot against each other.  The participant was instructed that if any discomfort developed they should swap the positions of the feet and to continue the exercise.  Image acquisition began after five minutes of this exercise with the exercise continued throughout the duration of the image acquisition.  Gentle encouragement was given to ensure sustained effort throughout the process.

Image analysis

The images were exported with image analysis performed using CVI 42 (Circle Cardiovascular Imaging Inc.,Calgary,Alberta,Canada).

For the TT method both distance and time data need to be measured.  For the distance, the HASTE axial images were used to measure the distances between the imaging planes.  Where the right or left pulmonary artery lay on different slices from the main pulmonary artery a vertical height was calculated from the slice thickness and number of slices.  Using this vertical height and horizontal height measured on the axial slices the final distance was calculated using Pythagoras theorem.  For the time component the phase and magnitude images were pulled up side by side.  A contour was manually drawn around the perimeter of the vessel on the magnitude image.  This was then automatically propagated throughout the remainder of the images, and manually corrected where malposition occurred.  The program then automatically calculated area, flow and velocity data which was exported to Excel 2010 (Microsoft, US).  The flow curves from the MPA, RPA and LPA were plotted, and the time to foot of the systolic waves then calculated.  This was identified as the intersection between the up stepping edge and baseline flow.  The up stepping edge was calculated as the line through the data points that lay between the 20% and 80% levels of the maximum flow rate.  The baseline was the horizontal line at minimum velocity before the up-stepping edge.  Pulse wave was calculated for both RPA and LPA using equation 1:

TT PWV=  ∆d/∆t

Where Δd is the distance between the planes and Δt is the difference in the time to foot between the MPA and RPA/LPA. Fig. 1

For the QA method the phase and magnitude images were pulled up side by side.  A contour was manually drawn around the perimeter of the vessel on the magnitude image on each image for the first 200ms of the cardiac cycle.  From this the programme calculated the total area and flow within the cross section of the pulmonary vessel.  The area and flow were then plotted against one another during early systole.  Early systole was defined as the time period in systole during which both the vessel area and flow were simultaneously increasing.  Three techniques have been described for the calculation of QA PWV with all being variations on the basic premise that:

QA PWV=  ∆Q/∆A

Where Q is the flow and A is the area through the pulmonary artery. The first is as described by Peng et al.[26] whereby the gradient of the line is fitted through these points using a minimum squared difference technique hereby known as QATrad. Fig. 2 The technique proposed by Quail et al.[28] follows the same principle but restricts analysis to the first 3 data points of the systolic upstroke in order to avoid the influence of reflected waves (QA3). Finally, Davies et al.[29] have proposed a technique that accounts for effects of reflected waves thereby allowing usage of more datapoints than the Quail et al. technique whilst maintaining accuracy.  This was originally described for pressure and velocity data derived from invasive catheter measurements, however has been adapted as follows (QAInv):

PWV=√((∑∆Q^2)/(∑∆A^2 ))

This used all datapoints in early systole, similar to the QATrad  technique.

Statistics

Descriptive statistics were used for the analysis of the demographic and clinical features of the cohorts with data expressed as mean ± SEM.  A dependant sample t-test was used to compare the difference between the first scan and the repeated measure, inter-scan measure and the exercise measure. An independent t-test was used to compare PWV between the two cohorts.  Bland-Altman plots were used to further investigate the inter-scan and inter-observer reproducibility.  All data were analysed using SPSS statistical package (version 21.0, SPSS Inc. Chicago, Illinois). Significance was assumed when p < 0.05.