Non-invasive stroke volume assessment in patients with pulmonary arterial hypertension: left-sided data mandatory
© Mauritz et al; licensee BioMed Central Ltd. 2008
Received: 18 August 2008
Accepted: 05 November 2008
Published: 05 November 2008
Cardiovascular Magnetic Resonance (CMR) is an emerging modality in the diagnosis and follow-up of patients with Pulmonary Arterial Hypertension (PAH). Derivation of stroke volume (SV) from the pulmonary flow curves is considered as a standard in this respect. Our aim was to investigate the accuracy of pulmonary artery (PA) flow for measuring SV.
Thirty-four PAH patients underwent both CMR and right-sided heart catheterisation. CMR-derived SV was measured by PA flow, left (LV) and right ventricular (RV) volumes, and, in a subset of nine patients also by aortic flow. These SV values were compared to the SV obtained by invasive Fick method.
For SV by PA flow versus Fick, r = 0.71, mean difference was -4.2 ml with limits of agreement 26.8 and -18.3 ml. For SV by LV volumes versus Fick, r = 0.95, mean difference was -0.8 ml with limits of agreement of 8.7 and -10.4 ml. For SV by RV volumes versus Fick, r = 0.73, mean difference -0.75 ml with limits of agreement 21.8 and -23.3 ml. In the subset of nine patients, SV by aorta flow versus Fick yielded r = 0.95, while in this subset SV by pulmonary flow versus Fick yielded r = 0.76. For all regression analyses, p < 0.0001.
In conclusion, SV from PA flow has limited accuracy in PAH patients. LV volumes and aorta flow are to be preferred for the measurement of SV.
In pulmonary arterial hypertension (PAH), cardiovascular magnetic resonance (CMR) has been proposed as a standard for the assessment of right ventricular function and characteristics of the pulmonary vascular bed [1, 2]. Accurate assessment of stroke volume (SV) by CMR is critical in this respect, since earlier studies revealed that SV is closely related to prognosis and that a change in SV reflects treatment effects [3, 4]. Since most of the CMR protocols used in PAH [5, 6] measured pulmonary artery flow, SV can be assessed by measuring flow in the main pulmonary artery (PA).
Previous studies have shown that this method is accurate to measure SV from PA flow in healthy subjects [7–9]. Whether this also holds true in PAH is questionable, since the velocity profile in PAH is non-laminar, in contrast to the profile in healthy subjects [10–13].
For this test of accuracy, a clinical standard is required. This standard is provided by the measurement of SV by the direct Fick principle during right heart catheterisation (RHC) . However, this is an invasive procedure and thus not well suited for either screening or frequently repeated follow-up measurements.
Therefore the aim of the present study is to assess the accuracy of the PA flow by CMR for measuring SV in PAH patients, by comparing SV from this PA flow with the SV assessed by the Fick method. In addition, other CMR-derived SV measures from aorta flow and cine imaging in the same patients will also be compared versus the Fick-derived SV.
Materials and methods
This study was approved by the institutional Review Board on Research Involving Human Subjects of the VU University medical centre, and all participants gave written informed consent. Between January 2004 and April 2007, a total of 34 patients who were given a final diagnosis of PAH after a complete diagnostic workup including RHC, underwent CMR. RHC and CMR were performed within 12 hours. The study group consisted of 23 female (68%) and 11 male (32%) patients with a mean age of 45 years ± 17 (standard deviation), and an age range of 20–84 years.
CMR imaging protocol
CMR flow measurements
SV was measured using both phase contrast flow and volumetric methods as described below.
CMR was perforned with a Siemens 1.5 T 'Sonata' whole body scanner (Siemens Medical Solutions, Erlangen. Germany), equipped with a phased-array body coil.
Phase-contrast CMR was acquired during continuous breathing with a gradient echo MR sequence, with velocity encoding perpendicular to the imaging plane and a velocity sensitivity of 120 cm/sec. The flow sequence was run with the following parameters: orientation = orthogonal to the main PA, slice thickness = 6 mm, field of view = 240 × 320 mm2, matrix size = 140 × 256, echo time = 4.8 ms, repetition time = 11 ms, temporal resolution = 22 ms, flip angle = 25°.
To explore whether there is inaccurate determination of flow-derived SV due to inherent technical limitations of phase-contrast CMR, aortic flow was also measured in a subset of nine patients, approximately 2 – 4 cm above the aortic valve and distal to the coronary arterial ostia, and the aortic flow-derived SV was also compared to the Fick-derived SV.
After the flow images were acquired, a 7 litre bottle containing H2O, with per litre 1.25 g NiSO4.6H2O + 2.6 g NaCL ("phantom") was then imaged with identical imaging parameters, to serve as correction for the background phase error in the mean PA and Aorta.
RV and LV volumetric measurements
The short-axis slices needed to encompass the entire left and right ventricle volumes to measure right ventricular (RV) and left ventricular (LV) volumes-derived SV, were obtained by steady state free precession imaging: 11 phase encoding lines per heart beat (which means 11 lines per segment), flip angle = 60°, slice thickness = 6 mm, slice gap = 4 mm, temporal resolution 36.3 ms and retrospectively ECG gated. Acquisition was in breathhold, acquisition time was 14 heartbeats.
CMR flow measurements
CMR post-processing was performed using the 'FLOW' software package (Dept. of Radiology, Leiden University Medical Center, Leiden, The Netherlands). The heart rate during cardiac CMR was recorded from the images. No aliasing due to high peak systolic velocities was encountered. The contours of the mean pulmonary artery MPA were automatically traced, with manual correction when necessary, simultaneously on magnitude and velocity-map images of all reconstructed phases. The software then calculated the velocity in each of the pixels included within the contours. The flow in each pixel (velocity times pixel area) is calculated and the pixel flow within the contour is summed. This is done for every temporal phase, resulting in blood flow as a function of time through the main PA.
RV and LV volumetric measurements
Right Heart Catheterization (RHC)
Diagnostic right heart catheterization was performed with a balloon tipped, flow directed 7F Swan-Ganz catheter (131HF7; Baxter Healthcare Corp; Irvine, CA). The direct Fick Principle (DF) was used. The patient was in stable condition, lying supine and breathing room air. Right atrial, right ventricular, pulmonary artery and pulmonary capillary wedge pressures were measured. Heart rate was monitored continuously. Average steady state oxygen consumption was obtained at the same time during RHC using a timed collection of expired air to measure oxygen consumption by an on-line analyzer (Vmax 229, Sensormedics, Yorba Linda, USA) connected to a mouth piece. To ensure accuracy, the system was calibrated before each study and the values were time-averaged over at least 5 minutes. Simultaneous arterial and mixed venous blood samples were then drawn for measurement of arterial oxygen saturation (SaO2) and haemoglobin (Hb) concentration. Arterial blood samples were obtained through either a radial or femoral arterial puncture and the mixed venous blood samples were obtained from the distal port of the pulmonary artery catheter. The cardiac output (CO) was then calculated by dividing the average oxygen consumption (VO2) value with the difference between the concentration of arterial oxygen (CaO2) and concentration of mixed venous oxygen content (CvO2). Eventually, SV (ml) was determined by dividing CO (L/min) by HR (beats/min).
Intra-observer and inter-observer variability
Interobserver variability of the different CMR methods was assessed by a second investigator analyzing all of data sets. Intraobserver variability was analyzed by assessing SVs obtained by the different CMR methods twice by one observer. The two assessments were separated by a one month period, and the observer was blinded for his previous results.
The results are expressed as mean ± SD. For each measurement of SV, the values for SV obtained invasively direct Fick and by the CMR-based methods were compared by regression analysis. Bland-Altman analysis was used to compare the degree of agreement between the three CMR methods and the Fick method for SV measurement . Bias was defined as the mean value of the differences between CMR methods and the Fick principle. Precision was defined as 1 standard deviation (SD) of the differences and limits of agreement as the bias ± 2 SDs reported as millilitres. The percentage error was calculated as the ratio of 2 times the SD to mean SV. Tendency toward over- or underestimation of SV by the different CMR-based methods was assessed with the two-tailed Student's t-test for paired data.
The mean difference (bias) and the coefficient of variability (CoV = SD of repeated measures as % of their mean) were used to assess intra- and interobserver variability.
P values less than .05 were considered to indicate significant differences. All statistical analyses were performed using GraphPad PRISM (version 4.0, GraphPad software, San Diego, CA).
Baseline characteristics of the patients*
No of PAH patients
n = 34
n = 14
n = 6
PAH and Collagen vascular disease
n = 5
PAH and HIV infection
n = 6
PAH and Porto pulmonary syndrome
n = 3
54 ± 17
n = 18
n = 12
n = 4
6 min walk test, m
441 ± 109
7 ± 4.5
74 ± 21
29 ± 11
45 ± 10
6.9 ± 3.3
95 ± 2
63 ± 8
4.7 ± 1.3
2.5 ± 0.6
81.5 ± 12
58.8 ± 16
777 ± 402
Results of Bland-Altman analyses of CMR based measurements for stroke volume
Mean ± SD
95% Limits of Agreement
55.2 ± 13.1
-18.3 to 26.8
51.40 ± 13.3
-9.8 to 5.2
59.5 ± 15.7
-23.3 to 21.8
59.6 ± 16.2
-10.4 to 8.8
Bland-Altman analysis demonstrated a small degree of underestimation of SV by PA flow with mean difference of 4.2 ml, the underestimation was statistically significant (P = 0.039).
If those patients with echo-derived right-to-left shunting were excluded, than the correlation coefficient between PA Flow-derived SV and Fick-derived SV increased to r2 = 0.56 and the bias decreased to a non-significant level of 1.9 ± 3.7 ml.
The slight overestimation of SV by RV and LV volumes, 0.75 ml and 0.8 ml respectively, was not significant.
Intra-and interobserver variability of CMR based measurements for stroke volume.
Bias(95% Limits of agreement)*
Bias(95% Limits of agreement)
-0.8 ± 1.7
(-4.1 to 2.4)
-0.9 ± 2.2
(-5.3 to 3.5)
-0.7 ± 2.6
(-5.4 to 5.0)
-1.3 ± 2.1
(-4.1 to 5.2)
-2.1 ± 5.2
(-12.3 to 8.1)
-2.6 ± 7.9
(-18 to 13.1)
-1.8 ± 3.2
(-8.0 to 4.4)
-2.6 ± 3.8
(-10.1 tot 4.9)
To our knowledge this study is the first to assess the accuracy of SV derived non-invasively from PA flow, RV volumes, Aorta flow and LV volumes by comparing these values with stroke volume assessed by means of the invasive Fick method in a group of PAH patients.
Our results showed that SV derived non-invasively from PA flow and RV volumes were in poor agreement with the Fick-derived SV. By contrast SV measurement from LV volumes and aortic flow showed good agreement with Fick.
Although SV is highly variable in healthy subjects, an earlier study  revealed that SV is fixed in PAH and does not even change during exercise. Thus although the invasive and CMR measurements were not performed synchronously, we expect similar SV values for each patient. The direct Fick principle was chosen as the method of reference because it is considered as a standard for measuring SV in patients with PAH .
SV assessment with PA flow
SV determined from the pulmonary artery flow curve was poorly related to stroke volume assessed by means of Fick. There are several explanations for this discrepancy.
First, we observed non-laminar velocity profiles during systole (figure 1b) in the main PA of PAH patients, which can suggest turbulent or helical flow patterns. Phase-contrast measurements are optimized for laminar flow, whereas in turbulent flow patterns the precision of the CMR flow measurements declines [19, 20]. In case of helical flow, motion in the non-velocity encoding directions may also lead to phase-shifts which are not related to through-plane flow.
Second, differences may result from cardiac right-to-left shunting, which causes underestimation of SV determination by PA flow. This is supported by our observation that SV using PA flow showed statistically significant underestimation of 4.2 ml; this underestimation disappeared by excluding those patients with a right-to-left shunt.
Another potential explanation is an inherent inaccuracy of the flow measurement by means of phase-contrast CMR. However, this explanation is unlikely, since in the subset of nine patients where aorta flow was also measured, a much better correlation with the direct Fick is found (figure 6). The accuracy of the aorta flow measurements for SV has already been validated by comparison with invasive SV measurements by Hundley et al  in 23 subjects.
SV assessment with RV volumes
Several factors may explain the errors in the stroke volume measurements derived from the volumetric measures of the right ventricle. First, in the short axis view, in the most basal RV slice, with the RV outflow tract and the inflow region with the tricuspid valve, usually it is more complicated to define the RV contour than in the case of the left ventricle. Because of the large area of this basal slice, errors in choosing the exact RV contour can render inaccurate volume estimates and thus inaccurate SV measurements. Second, evaluation of right ventricular volumes in patients with PAH is difficult because of the complex anatomy of the right ventricle. Third RV endocardial boundary delineation is more complicated because of the considerably more trabeculation of the RV in comparison with the LV . These factors may be improved with further software development to automate boundary detection and also assist in the selection of the basal slice during analyses. Fourth, in patients with pulmonary hypertension tricuspid regurgitation (TR) is common. Thus, with considerable TR, the volumetric SV overestimates the actual SV , because it is impossible to differentiate between the volume that moves back through the tricuspid valves and forward though the pulmonary valves.
SV assessment with LV volumes
The correlation between the SV derived from LV end-diastolic and end-systolic volumes and the SV from Fick was very tight (Figure 5c). This is presumably so because for the LV it is rather easy to delineate the endocardial border and to select the basal slice.
Additionally there was no sign of systematic errors for low or high values indicating that the SV measure from LV volumes is valid for a wide range of SV Values. There were some high SV values in three patients with porto pulmonary syndrome which can be explained by the hyperdynamic circulation feature in these patients .
Intra-observer and inter-observer variability
The results of the current study indicated that the intra- and interobserver variability of SV measurements using PA flow, aortic flow and LV volumes measurements were low, showing small differences and low coefficients of variability. A larger interobserver and intraobserver variability was shown for the RV volumes. This could be explained by factors mentioned above. Together with the poor agreement of RV-derived SV versus Fick, this advises against the use of RV volumes for SV determination.
The accurate non-invasive SV assessment has clinical importance because monitoring of SV in PAH patients now becomes feasible as a routine method for monitoring the effects of medical treatment and the follow-up of patients non-invasively. The results of this study have important implications for PAH patients. The SV measurement using PA flow is less accurate in comparison to the SV derived from LV volumes and aorta flow. Therefore, LV volumes and aorta flow should be the methodology of choice in this patients group to measure SV accurately. Consequently, it is advisable to include in the CMR basic protocol a stack of contiguous short axis slices covering the full extend of the left ventricle, and an aortic flow measurement for accurate SV assessment.
We acknowledge some limitations of our study including the fact that catheter studies and CMR examinations could not be performed simultaneously. However, the results of this study demonstrate that SV derived from LV volumes correlate well with measurements obtained using the direct Fick method, supporting earlier findings that stroke volume does not vary over a short period of time in PAH, making this parameter of value to monitor PAH patients.
Furthermore we acknowledge the small sample size of the aorta flow measurements.
Finally, the present study does not provide conclusive evidence of the cause of the discrepancy between the PA flow and the direct Fick.
In patients with PAH, taking the direct Fick as a standard invasive method, non-invasive derivation of SV should preferably be carried out using aortic flow or LV volumes. SV assessment with PA flow or RV volumes shows limited accuracy.
coefficient of variability
end diastolic volume
end Systolic volume
cardiovascular magnetic resonance imaging
pulmonary arterial hypertension
right heart catheterisation
- Torbicki A: Cardiac magnetic resonance in pulmonary arterial hypertension: a step in the right direction. Eur Heart J. 2007, 28 (10): 1187-9. 10.1093/eurheartj/ehm074.View ArticlePubMedGoogle Scholar
- McLure LE, Peacock AJ: Imaging of the heart in pulmonary hypertension. Int J Clin Pract Suppl. 2007, 15-26. 10.1111/j.1742-1241.2007.01496.x. 156
- van Wolferen SA, Marcus JT, Boonstra A, Marques KM, Bronzwaer JG, Spreeuwenberg MD, et al: Prognostic value of right ventricular mass, volume, and function in idiopathic pulmonary arterial hypertension. Eur Heart J. 2007, 28 (10): 1250-7. 10.1093/eurheartj/ehl477.View ArticlePubMedGoogle Scholar
- Roeleveld RJ, Vonk-Noordegraaf A, Marcus JT, Bronzwaer JG, Marques KM, Postmus PE, et al: Effects of epoprostenol on right ventricular hypertrophy and dilatation in pulmonary hypertension. Chest. 2004, 125 (2): 572-9. 10.1378/chest.125.2.572.View ArticlePubMedGoogle Scholar
- Kreitner KF, Kunz RP, Ley S, Oberholzer K, Neeb D, Gast KK, et al: Chronic thromboembolic pulmonary hypertension – assessment by magnetic resonance imaging. Eur Radiol. 2007, 17 (1): 11-21. 10.1007/s00330-006-0327-x.View ArticlePubMedGoogle Scholar
- Marcus JT, Vonk NA, De Vries PM, van Rossum AC, Roseboom B, Heethaar RM, et al: MRI evaluation of right ventricular pressure overload in chronic obstructive pulmonary disease. J Magn Reson Imaging. 1998, 8 (5): 999-1005. 10.1002/jmri.1880080502.View ArticlePubMedGoogle Scholar
- Kondo C, Caputo GR, Semelka R, Foster E, Shimakawa A, Higgins CB: Right and left ventricular stroke volume measurements with velocity-encoded cine MR imaging: in vitro and in vivo validation. AJR Am J Roentgenol. 1991, 157 (1): 9-16.View ArticlePubMedGoogle Scholar
- Firmin DN, Nayler GL, Klipstein RH, Underwood SR, Rees RS, Longmore DB: In vivo validation of MR velocity imaging. J Comput Assist Tomogr. 1987, 11 (5): 751-6. 10.1097/00004728-198709000-00001.View ArticlePubMedGoogle Scholar
- Caputo GR, Kondo C, Masui T, Geraci SJ, Foster E, O'Sullivan MM, et al: Right and left lung perfusion: in vitro and in vivo validation with oblique-angle, velocity-encoded cine MR imaging. Radiology. 1991, 180 (3): 693-8.View ArticlePubMedGoogle Scholar
- Bogren HG, Buonocore MH: Complex flow patterns in the great vessels: a review. Int J Card Imaging. 1999, 15 (2): 105-13. 10.1023/A:1006281923372.View ArticlePubMedGoogle Scholar
- Bogren HG, Klipstein RH, Mohiaddin RH, Firmin DN, Underwood SR, Rees RS, et al: Pulmonary artery distensibility and blood flow patterns: a magnetic resonance study of normal subjects and of patients with pulmonary arterial hypertension. Am Heart J. 1989, 118 (5 Pt 1): 990-9. 10.1016/0002-8703(89)90235-4.View ArticlePubMedGoogle Scholar
- Okamoto M, Miyatake K, Kinoshita N, Sakakibara H, Nimura Y: Analysis of blood flow in pulmonary hypertension with the pulsed Doppler flowmeter combined with cross sectional echocardiography. Br Heart J. 1989, 51 (4): 407-15. 10.1136/hrt.51.4.407.View ArticleGoogle Scholar
- Reuben SR, Swadling JP, de Lee GJ: Velocity profiles in the main pulmonary artery of dogs and man, measured with a thin-film resistance anemometer. Circ Res. 1970, 27 (6): 955-1001.View ArticlePubMedGoogle Scholar
- Hoeper MM, Maier R, Tongers J, Niedermeyer J, Hohlfeld JM, Hamm M, et al: Determination of cardiac output by the Fick method, thermodilution, and acetylene rebreathing in pulmonary hypertension. Am J Respir Crit Care Med. 1999, 160 (2): 535-41.View ArticlePubMedGoogle Scholar
- Chernobelsky A, Shubayev O, Comeau CR, Wolff SD: Baseline correction of phase contrast images improves quantification of blood flow in the great vessels. J Cardiovasc Magn Reson. 2007, 9 (4): 681-5. 10.1080/10976640601187588.View ArticlePubMedGoogle Scholar
- Marcus JT, Gotte MJ, DeWaal LK, Stam MR, Geest Van der RJ, Heethaar RM, et al: The influence of through-plane motion on left ventricular volumes measured by magnetic resonance imaging: implications for image acquisition and analysis. J Cardiovasc Magn Reson. 1999, 1 (1): 1-6. 10.3109/10976649909080828.View ArticlePubMedGoogle Scholar
- Bland JM, Altman DG: Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1 (8476): 307-10.
- Holverda S, Gan CT, Marcus JT, Postmus PE, Boonstra A, Vonk-Noordegraaf A: Impaired stroke volume response to exercise in pulmonary arterial hypertension. J Am Coll Cardiol. 47 (8): 1732-3. 10.1016/j.jacc.2006.01.048.
- Bakker CJ, Hoogeveen RM, Viergever MA: Construction of a protocol for measuring blood flow by two-dimensional phase-contrast MRA. J Magn Reson Imaging. 1999, 9 (1): 119-27. 10.1002/(SICI)1522-2586(199901)9:1<119::AID-JMRI16>3.0.CO;2-F.View ArticlePubMedGoogle Scholar
- Evans AJ, Blinder RA, Herfkens RJ, Spritzer CE, Kuethe DO, Fram EK, et al: Effects of turbulence on signal intensity in gradient echo images. Invest Radiol. 1988, 23 (7): 512-8. 10.1097/00004424-198807000-00006.View ArticlePubMedGoogle Scholar
- Hundley WG, Li HF, Hillis LD, Meshack BM, Lange RA, Willard JE: Quantitation of cardiac output with velocity-encoded, phase-difference magnetic resonance imaging. Am J Cardiol. 75 (17): 1250-5.
- Grothues F, Moon JC, Bellenger NG, Smith GS, Klein HU, Pennell DJ: Interstudy reproducibility of right ventricular volumes, function, and mass with cardiovascular magnetic resonance. Am Heart J. 2004, 147 (2): 218-23. 10.1016/j.ahj.2003.10.005.View ArticlePubMedGoogle Scholar
- Hoeper MM, Tongers J, Leppert A, Baus S, Maier R, Lotz J: Evaluation of right ventricular performance with a right ventricular ejection fraction thermodilution catheter and MRI in patients with pulmonary hypertension. Chest. 2001, 120 (2): 502-7. 10.1378/chest.120.2.502.View ArticlePubMedGoogle Scholar
- Hoeper MM, Krowka MJ, Strassburg CP: Portopulmonary hypertension and hepatopulmonary syndrome. Lancet. 363 (9419): 1461-8. 10.1016/S0140-6736(04)16107-2.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.