We have shown that: (i) The QA method using a high temporal resolution sequence through the right pulmonary artery using a calculation that accounts for wave reflections yields consistently better within scan, inter-scan, intra-observer and inter-observer reproducibility; (ii) Age related arterial stiffening as seen in the systemic circulation does not occur in the pulmonary vessels; (iii) Pulmonary PWV is stable and consistent at both rest and exercise.

This is the most comprehensive analysis of the techniques for measuring pulmonary PWV to date.  Prior studies have either provided within scan reproducibility in n=10 using just the TT technique,[25] within and between scan reproducibility in n=17 using just the QA technique,[26] or inter-observer comparison but without interscan assessment in n=33.[24]  Thus our study of both measurement techniques as well as interrogating the effects of different acquisition sequences and different post-processing methods provides the first in depth and rigorous assessment of pulmonary PWV assessment. 

It is perhaps not surprising that the RPA QA PWV measurements were the most precise as the RPA suffers from significantly less through plane motion during the cardiac cycle than the MPA as well as suffering from less respiratory variation in location. However justification of focusing on one of the branch pulmonary arteries rather than the main pulmonary artery relies on changes affecting all three arteries equally.  Whilst it may be reasonable to assume this in diffuse pulmonary disease states such as COPD or idiopathic pulmonary arterial hypertension, prior work in chronic thromboembolic pulmonary hypertension (CTEPH) has demonstrated that the capacitance varies between the left and right pulmonary arteries.[30]  Thus while the RPA may provide the best results in terms of reproducibility, care may have to be taken when using this in disease states, particularly CTEPH where despite its greater inter measure variability the MPA may be better suited.  Further work is required in other disease states to evaluate the degree of variability in measures of pulmonary stiffness.  Further consideration must be given to the fact that there appears to be a change in stiffness between the main pulmonary trunk and the branch pulmonary arteries observed in our study with both the right and left pulmonary arteries exhibiting lower stiffness than the main pulmonary artery.  Current studies demonstrating prognostic significance of proximal pulmonary arterial stiffness have all focussed on changes in the main pulmonary artery, thus these findings will have to replicated in the right pulmonary artery in order to validate its usefulness as a marker of arterial stiffness of prognostic significance.

Our finding of improved reproducibility with a lower spatial resolution and higher temporal resolution is slightly counter intuitive as it has previously been thought that high spatial resolution was the more important factor in QA assessment due to the need for accurate area measurement.[24,27]  However it may be that increasing the temporal resolution increases the datapoints for deriving the flow-area gradient thus improving the reproducibility. Indeed the QAInv, which uses all the points in early systole, was seen to be more precise than the QA3 technique which uses only the first three.  The good agreement of QAInv and QA3 compared with QATrad despite their different sampling windows and different calculations also shows the importance of correcting for the reflected wave in early systole as they both produced significantly lower results than the QATrad using both the high spatial resolution and high temporal resolution sequences.  Previous animal models have shown an early expansion wave arriving during systole in the MPA,[31,32] and this becomes even more important in disease when a backward compression wave starts to arrive in systole.[33]  Recently a phase contrast acquisition technique utilising a golden angle radial acquisition has been described in the pulmonary circulation which maintains high edge sharpness whilst maintaining good temporal resolution which may further improve the reproducibility of this technique.[28]  The improved inter-scan reproducibility of the QA technique is an important finding as previous studies have shown QA to have a lower intra- and interobserver reproducibility compared with the TT technique, in addition to being significantly slower to perform.[24,27] 

In our study we found no change in pulmonary arterial PWV with exercise.  This is in contradistinction to the recent findings of Forouzan et al.[34] who demonstrated an increase in PWV on exercise in a study of n=15 using the QA technique.  This previous study obtained a larger increase in cardiac output compared to our own, thus it may be that in our study the participants were insufficiently stressed to illicit a change in the proximal pulmonary arterial stiffness.  However there are several reasons to support the accuracy of the findings of the current study. Firstly, our measures were obtained during rather than after exercise, and we have used both the TT and QA technique with a similar lack of change seen with both.  Secondly, the previous study used the QATrad method for calculating the data, which we have shown provides a higher measurement of PWV than methods accounting for wave reflections.  Wave reflections are known to increase substantially during exercise states thus the change in their PWV may have been due to a change in the magnitude of wave reflections rather than a change in arterial stiffness.[32]  Finally our observations are in agreement with two prior invasive studies: the first of these by Laskey et al.[35] using invasive pressure and velocity wires to calculate the arterial input impedance spectrum demonstrated no change in PWV with exercise in healthy controls; and a second by Domingo et al. using right heart catheterisation and intravascular ultrasound demonstrating an increased pulsatility, but with a fall in elastic modulus and no significant change in capacitance on exercise.[36]

We have shown for the first time a lack of change in the pulmonary arterial PWV with age in comparison with the aorta where the expected increase in PWV was observed. Burman et al.[37] have recently demonstrated decreasing pulmonary relative area change with age, however as relative area change falls with falling stroke volume which is also known to fall with age,[38] these previous findings do not necessarily indicate arterial stiffening in the pulmonary circulation with age. Ex vivo studies have suggested increased stiffness of the pulmonary vessels with age, however these have used far greater forces than found in the pulmonary circulation, even in advanced pulmonary hypertension, to exhibit this.[39,40]  While invasive catheter studies have shown an effect of age on the resistance-compliance time constant, this effect was small and predominantly driven by the pulmonary capillary wedge pressure.[41]

Several limitations must be mentioned regarding this study.  First, no gold standard was used to compare the techniques against.  The main reason for this is that the gold standard is invasive right heart catheterisation, which in a healthy population with no comorbidities would be difficult to justify.  However use of the technique in patients due to undergo a clinically indicated right heart catheterisation would be a useful avenue for future work.  In addition, while none were compared to the gold standard, reproducibility and reduced interscan variability is as, if not more, important than accuracy as it allows detection of smaller changes within the study group of patient cohort of interest.  Secondly a high spatial resolution sequence was not applied through either the RPA or LPA, thus whilst it is unlikely given the trend of results seen in the MPA, it cannot be excluded that higher spatial resolution sequences through these regions would not result in improved reproducibility compared with the high temporal resolution sequences.  Thirdly, while the number undergoing repeat measurement whilst on table was reasonably sized, those undergoing repeat measurement at 6 months was substantially smaller, however the results of this largely mirrored the results of the on-table repeat measure thus aiding in the validation of their findings.

In conclusion use of the QA technique combined with a high temporal resolution acquisition and a post processing technique to account for wave reflections yields the most reproducible measurements of pulmonary PWV.  Stability of pulmonary PWV with age indicates that any observed changes in pulmonary PWV are indicative of pulmonary vascular pathophysiological changes rather than arterial aging.

Normal 0 false false false EN-US JA X-NONE