Download Magnetic Resonance–Derived 3-Dimensional Blood Flow Patterns

Magnetic Resonance–Derived 3-Dimensional Blood Flow
Patterns in the Main Pulmonary Artery as a Marker of
Pulmonary Hypertension and a Measure of Elevated Mean
Pulmonary Arterial Pressure
Gert Reiter, PhD*; Ursula Reiter, PhD*; Gabor Kovacs, MD; Bernhard Kainz, MSc;
Karin Schmidt, MD; Robert Maier, MD; Horst Olschewski, MD; Rainer Rienmueller, MD
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
Background—Pulmonary hypertension is a disease characterized by an elevation in pulmonary arterial pressure that is
diagnosed invasively via right heart catheterization. Such pathological altered pressures in the pulmonary vascular
system should lead to changes in blood flow patterns in the main pulmonary artery.
Methods and Results—Forty-eight subjects (22 with manifest pulmonary hypertension, 13 with latent pulmonary
hypertension, and 13 normal control subjects) underwent time-resolved 3D magnetic resonance phase-contrast imaging
of the main pulmonary artery. Velocity fields that resulted from measurements were calculated, visualized, and analyzed
with dedicated software. Main findings were as follows: (1) Manifest pulmonary hypertension coincides with the
appearance of a vortex of blood flow in the main pulmonary artery (sensitivity and specificity of 1.00, 95% confidence
intervals of 0.84 to 1.00 and 0.87 to 1.00, respectively), and (2) the relative period of existence of the vortex correlates
significantly with mean pulmonary arterial pressure at rest (correlation coefficient of 0.94). To test the diagnostic
performance of the vortex criterion, we furthermore investigated 55 patients in a blinded prospective study (22 with
manifest pulmonary hypertension, 32 with latent pulmonary hypertension, and 1 healthy subject), which resulted in a
sensitivity of 1.00 and specificity of 0.91 (95% confidence intervals of 0.84 to 1.00 and 0.76 to 0.98, respectively).
Comparison of catheter-derived mean pulmonary artery pressure measurements and calculated mean pulmonary artery
pressure values resulted in a standard deviation of differences of 3.6 mm Hg.
Conclusions—Vortices of blood flow in the main pulmonary artery enable the identification of manifest pulmonary
hypertension. Elevated mean pulmonary arterial pressures can be measured from the period of vortex existence. (Circ
Cardiovasc Imaging. 2008;1:23-30.)
Key Words: hypertension, pulmonary 䡲 magnetic resonance imaging 䡲 blood flow 䡲 blood pressure 䡲 catheterization
P
ulmonary hypertension (PH) is a disease characterized by
an elevation in pulmonary arterial pressure that manifests
in its early stages during exercise. The diagnosis of PH is
established when mean pulmonary artery pressure (mPAP)
measured invasively by right heart catheterization exceeds
25 mm Hg at rest (manifest PH) or 30 mm Hg with exercise
(latent PH).1
which is related to mPAP,6 the quantity used to define PH.
Parameters that directly correlate with mPAP or pulmonary
vascular resistance have been related to the shortened acceleration times of blood flow in the right ventricular outflow
tract.2,3,7–11 However, the overall applicability and accuracy
of the above-described methods are controversial.12–16
Additionally, various authors10,11,17,18 have observed spatially highly inhomogeneous cross-sectional flow profiles and
fractions of retrograde flow during systole and diastole in the
main pulmonary artery in patients with PH. This was not
observed in normal subjects. Even though no correlation
between retrograde flow and the degree of PH was found,10,11
these observations suggest PH-caused flow topologies in the
Clinical Perspective see p 30
Noninvasive imaging modalities such as echocardiography
and MRI have been used to predict PH. Evaluation of the
regurgitant tricuspid jet velocity2,3 and septal bowing4,5 have
been used to estimate systolic pulmonary artery pressure,
Received January 3, 2008; accepted May 14, 2008.
From Siemens Medical Solutions (G.R.), Graz, Austria; University Clinic of Radiology (U.R., K.S., R.R.), University Clinic of Internal Medicine,
Department of Pulmonology (G.K., H.O.) and Department of Cardiology (R.M.) of the Medical University of Graz, Graz, Austria; and the Institute for
Computer Graphics and Vision (B.K.), Technical University of Graz, Graz, Austria.
Clinical trial registration information—URL: http://www.clinicaltrials.gov. Unique identifier: NCT00575692.
The online-only Data Supplement is available with this article at http://circimaging.ahajournals.org/cgi/content/full/1/1/23/DC1.
Correspondence to Gert Reiter, Siemens Medical Solutions, Strassgangerstrasse 315, A-8054 Graz, Austria. E-mail [email protected]
*The first 2 authors contributed equally to this work.
© 2008 American Heart Association, Inc.
Circ Cardiovasc Imaging is available at http://circimaging.ahajournals.org
23
DOI: 10.1161/CIRCIMAGING.108.780247
24
Circ Cardiovasc Imaging
Table 1.
July 2008
Study Population: Exploratory Study
Manifest PH
(n⫽22)
Characteristic
Latent PH
(n⫽13)
Patients
Without PH
(n⫽3)
Healthy
Volunteers
(n⫽10)
Sex (female/male), n
10/12
11/2
2/1
6/4
Age, y
61⫾10
66⫾6
55⫾9
26⫾7
mPAP at rest, mm Hg
41⫾13
16⫾4
12⫾2
䡠䡠䡠
39⫾6
23⫾5
䡠䡠䡠
mPAP during exercise, mm Hg
䡠䡠䡠
Values for age and mPAP are given as mean⫾SD.
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
right ventricular outflow tract. Neither echocardiography nor
conventional magnetic resonance (MR) through-plane cine
phase-contrast imaging appears to be an appropriate technique for a detailed investigation of this assumption, but the
MR phase-contrast method can be used to assess timeresolved 3D velocity information of blood flow.19 Complex
cardiovascular flow topologies have been studied with this
technique and various adequate visualization tools.20,21 The
purpose of the present study was to investigate characteristic
differences in 3D blood flow patterns in the main pulmonary
artery of patients with manifest PH, patients with latent PH,
and subjects without PH to determine the blood flow pattern–
related measures for mPAP and PH.
Methods
important change in drug treatment or disease state occurred between
the 2 examinations.
Right Heart Catheterization and
Patient Classification
Right heart catheter examinations were performed with a 7F
quadruple-lumen, balloon-tipped, flow-directed Swan-Ganz catheter
by the transjugular approach (Baxter Healthcare Corp, Irvine, Calif).
Parameters that were obtained included mPAP. Patients with no
manifest PH underwent a symptom-limited exercise test on a cycle
ergometer in a semisupine position with a right heart catheter in
place. A standard protocol with a 25-W increase every 2 minutes was
used, and mPAP was measured at each stage of exercise.
On the basis of the results of the catheterization, patients were
assigned to 3 groups1: manifest PH (clinical classification; see online
Data Supplement), latent PH, and subjects without PH. Patients without
PH and healthy volunteers were assigned to the control group.
Subjects
MR Imaging
The present study was approved by the local ethics review board, and all
subjects gave written informed consent. Subjects with contraindications
to MR were not enrolled. Patients with suspected or known PH were
sent consecutively to MRI after right heart catheterization
was completed.
The study was subdivided into exploratory and blinded prospective studies. A total of 42 patients were included in the exploratory
study. Thirty-eight patients (Table 1) were investigated by right heart
catheterization and MR phase-contrast imaging of the pulmonary
artery. Four patients were excluded from evaluation because of air
trapping during ergometry (1 patient), missing catheter data during
exercise (1 patient), and incomplete MR investigation (2 patients).
Patients underwent catheter and MR investigation within 2.8⫾4.1
days.
In addition, 10 healthy volunteers with no history of cardiovascular or pulmonary disease (Table 1) underwent MR phase-contrast
imaging of the pulmonary artery. Conventional ECG-gated cine MRI
was used to exclude valvular regurgitations and to confirm normal
left and right ventricular function.22
Sixty patients were included in the blinded prospective study. Five
patients were excluded from evaluation because of missing catheter
data during exercise (3 patients) and incomplete MR investigation (2
patients). The remaining 55 patients (Table 2) underwent catheter
and MR investigation within 6.4⫾6.8 days. Generally, no clinically
ECG-gated MR imaging was performed on a 1.5-T field-strength
scanner (Magnetom Sonata; Siemens, Erlangen, Germany) with a
6-channel cardiac-array coil. All subjects were investigated in the
supine position. Velocity field measurements were part of a comprehensive protocol that included retrospectively ECG-gated cine
gradient echo imaging with steady-state free precession in the
standard cardiac views (3- and 4-chamber views, right and left
ventricular 2-chamber views, right and left ventricular outflow tract,
and short-axis planes) to verify in particular the normal cardiac
function of healthy volunteers.
To acquire velocity field data, the main pulmonary artery was
covered with 3 to 5 gapless slices in the right ventricular outflow tract
orientation by 2D, retrospectively ECG-gated, spoiled gradient echo–
based phase-contrast sequences with a 4-point velocity encoding
scheme.19 Velocity encoding was set to 90 cm/s in all directions and
adapted in the case of aliasing in the main pulmonary artery. Further
parameters of the protocol used in the present study were a 234 to
276⫻340-mm2 field of view, 96 to 114⫻192 matrix (interpolated to 192
to 228⫻384 matrix), 15° flip angle, 89-ms temporal resolution (3
segments, 7.5-ms repetition time), reconstruction of 20 cardiac phases,
4.1-ms echo time, and 451 Hz/pixel bandwidth. GRAPPA (generalized
auto-calibrating partially parallel acquisition) with a parallel acquisition
factor of 2 and an average of 24 reference lines was used to keep the
average imaging time per slice at 22 heartbeats, which allowed the
performance of measurements during breath-holds in inspiration. For
patients who were not able to hold their breath well, the investigation
was performed under free-breathing conditions with 3-fold averaging to
suppress breathing artifacts.
Table 2.
Study Population: Blinded Prospective Study
Characteristic
Manifest PH
(n⫽22)
Latent PH
(n⫽32)
Patients
Without PH
(n⫽1)
Sex (female/male), n
12/10
27/5
1/0
Age, y
67⫾13
56⫾11
51
mPAP at rest, mm Hg
39⫾10
17⫾3
14
40⫾6
25
mPAP during exercise, mm Hg
䡠䡠䡠
Values for age and mPAP are given as mean⫾SD.
Image Processing and Analysis in the
Exploratory Study
Calculation of velocity fields from the acquired phase images, their
visualization, and analysis thereof were performed with dedicated
software (4D Flow, Siemens).20 –23 This software provides visualization
of each velocity vector as a color-encoded vector in 3D space: The
length and color of the vector indicate the magnitude of velocity,
whereas the direction of the vector represents the direction of the
Reiter et al
Blood Flow Patterns in Pulmonary Hypertension
25
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
Figure 1. Typical flow patterns in the right
ventricular outflow tract at different cardiac
phases for a patient with manifest PH (A,
D, and G), a patient with latent PH (B, E,
and H), and a normal subject (C, F, and I).
PA indicates main pulmonary artery; PV,
pulmonary valve; and RV, right ventricle.
At maximum outflow (A through C), flow
profiles were distributed homogenously
across the cross sections of the main pulmonary artery in the manifest PH group
(A), the latent PH group (B), and the normal group (C). In later systole (D through
F), a vortex was formed in the manifest
PH group (D). No such vortex could be
found in the latent PH group (E) or the
normal group (F). After pulmonary valve
closure (G through I), the vortex in the PH
group persisted for some time. In all
cases, continuous diastolic blood flow
upward along the anterior wall of the main
pulmonary artery could be observed.
Although this phenomenon disappeared
quickly in controls (I), it was observed significantly longer in latent PH (H) and manifest PH (G).
velocity. Velocity vectors are projected onto (opaque) corresponding
anatomic phase-contrast images. The suppression of noisy pixels (by a
signal-to-noise threshold in the anatomic phase-contrast images) and an
adjustable thinning out of the velocity vector fields enable interpretation
of the blood flow patterns within the anatomic context (Figure 1).
In addition to determination of the systolic acceleration time (time
difference between maximum and onset of pulmonary outflow),
systolic ejection time (time difference between onset and end of
pulmonary outflow), maximum blood flow velocity above the
pulmonary valve at maximum pulmonary outflow, and maximum
velocity above the pulmonary valve at its closure, 3 features of the
flow field in the main pulmonary artery were investigated in more
detail. First, the existence of vortices in the primary flow direction in
the main pulmonary artery was investigated. In general, a vortex is
defined as a ring- or spiral-shaped motion of fluid or gas. In terms of
the vector field representation of blood flow, this means the
formation of concentric ring- or spiral-shaped curves in the main
pulmonary artery to which velocity vectors are tangential. A vortex
in the primary flow direction in the main pulmonary artery means
that there is forward and backward flow through a cross section of
the main pulmonary artery through the center of the vortex (Figure
2A). Both the existence of such a vortex and the relative period of
existence (tvortex) of a nonvalvular vortex in the main pulmonary
artery (number of cardiac phases with vortex divided by the total
number of imaged cardiac phases) were determined visually and
independently by 2 observers.
The second specific feature investigated was the diastolic streamline of blood flow upward along the anterior wall of the main
pulmonary artery. In terms of the vector field representation of blood
flow, this means that blood moves continuously up the anterior wall
of the main pulmonary artery during diastole (Figure 2B). The
relative period of existence of diastolic streamlines (tstreamlines; the
number of cardiac phases after pulmonary valve closure with
streamlines along the main pulmonary artery divided by the total
number of imaged cardiac phases) was determined visually and
independently by 2 observers.
Third, a location index was introduced to characterize the velocity
profile of blood flow above the pulmonary valve in the anteriorposterior direction. Depending on whether the maximum velocity
through the cross section of the main pulmonary artery appeared in
the anterior, middle, or posterior third of the vessel, the location
index was set to ⫹1, 0, or ⫺1, respectively (Figure 2C). In case of
equal maximum velocities in multiple sections, the location index
was set to the average value of the corresponding thirds (maximum
velocities in the section thirds were interpreted as being different if
their difference exceeded a typical pixel-by-pixel variation of velocities). The location index was determined in the cardiac phase with
maximum pulmonary outflow, as well as in the cardiac phase of
pulmonary valve closure.
Image Processing and Analysis in the Blinded
Prospective Study
After analogous postprocessing of MR data, the existence of a vortex
and its relative period of existence (tvortex) were determined by a
single observer blinded to the catheter results.
Statistical Analysis
Statistical analysis was performed with NCSS (J. Hintze, NCSS,
LLC, Kaysville, Utah). Mean values are given together with standard
deviations (SDs). Group means of the manifest PH, latent PH, and
control groups were compared via 1-way analysis of variance and
Tukey-Kramer multiple-comparison test. We generally set the significance level to 0.05. For tstreamlines, we additionally performed an
analysis of covariance with age as the covariate.
A ␬-coefficient was calculated to specify the interrater agreement
with respect to the existence of a vortex. To characterize the
diagnostic accuracy of predicting manifest PH via vortex formation,
26
Circ Cardiovasc Imaging
July 2008
Results
Description of Flow Patterns
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
Figure 2. Specific flow features in the main pulmonary artery.
PA indicates main pulmonary artery; PV, pulmonary valve; RV,
right ventricle; a, anterior; and p, posterior. A, Vortex in the primary flow direction in the main pulmonary artery. B, Streamlines
of blood flow up the anterior wall of the main pulmonary artery.
C, Maximum velocities vmax,a, vmax,m, and vmax,p through a cross
section of the main pulmonary artery above the pulmonary valve
were determined in the anterior (a), middle (m), and posterior (p)
third of the vessel. Depending on which of these velocities was
maximal, the location index for the velocity profile was set to
⫹1 (vmax,a maximal), 0 (vmax,m maximal), or ⫺1 (vmax,p maximal). It
is 0 in the above example.
sensitivity and specificity were determined together with their 95%
confidence intervals.
Interobserver variability in the determination of relative periods of
the existence of vortices and streamlines was specified by withinsubject SDs (calculated by variance component analysis) and intraclass correlation coefficients. Mean values of the data of the 2
observers were used to study differences between group means and
to analyze the relationship between periods of existence of vortices
and mPAP by correlation and regression analysis. For the blinded
prospective study, comparison of mPAP values measured by catheter
and calculated by the regression equation obtained in the exploratory
study was performed by Bland-Altman analysis.
The authors had full access to the data and take full responsibility
for the integrity of the data. All authors have read and agree to the
manuscript as written.
During systolic acceleration of blood flow through the pulmonary valve, the flow patterns in the manifest PH, latent PH,
and control groups were similar. Blood flow velocities were
uniformly distributed across the cross section of the main
pulmonary artery. Their primary direction was parallel to the
vessel wall. Velocities increased from one cardiac phase to
the next until a maximum was reached (Figure 1A through
1C). Although mean maximal peak velocities did not differ
significantly between the groups (79⫾21 cm/s in manifest
PH, 74⫾11 cm/s in latent PH, and 77⫾11 cm/s in the
controls), the mean acceleration time was significantly
shorter in the manifest PH group and was not significantly
different between the latent PH and control groups (mean
acceleration times of 107⫾25 ms in manifest PH, 133⫾24 ms
in latent PH, and 141⫾26 ms in the control group; P⫽0.0006
for F test).
After maximum outflow, blood flow velocities gradually
decreased until the pulmonary valve closed. No significant
differences in total ejection times were observed (374⫾54 ms
in manifest PH, 365⫾64 ms in latent PH, and 392⫾60 ms in
controls). During the deceleration period, the flow profiles
changed from rather homogeneous velocity distributions
across the cross section of the main pulmonary artery to
profiles with higher velocities at the anterior wall (Figure 1D
through 1F): Although the position indexes for velocity
profiles above the pulmonary valve were 0.1⫾0.5, 0.2⫾0.3,
and 0.1⫾0.3 at maximal outflow, they were 1.0⫾0.1,
1.0⫾0.0, and 1.0⫾0.0 at valvular closure in the manifest PH,
latent PH, and control groups, respectively.
There was, however, a structure that occurred only in
patients with manifest PH: A vortex in the primary flow
direction was formed, typically below the right pulmonary
artery at the posterior wall of the main pulmonary artery
(Figure 1D). In 2 patients with high mPAP, vortex formation
had already begun in the acceleration phase. The rotational
direction of the vortex was always more distal from the right
ventricle from anterior to posterior and more proximal to the
right ventricle from posterior to anterior (as drawn in Figure
2A and visible in Figure 1D and 1G). During deceleration of
blood flow through the pulmonary valve, the region covered
by the vortical motion typically increased, and the center of
the vortex moved in the main pulmonary artery.
After closure of the pulmonary valve, blood flow in the
main pulmonary artery did not stop immediately. Mean
maximum velocities above the pulmonary valve were 28⫾9
cm/s in manifest PH, 21⫾6 cm/s in latent PH, and 18⫾5 cm/s
in controls, and the differences between manifest PH and both
latent PH and controls were significant (P⫽0.0012 for F test).
In all subjects, blood flow was observed spatially continuously up the anterior wall of the main pulmonary artery after
pulmonary valve closure; in the case of manifest PH, this
flow was accompanied by the rotational flow of the vortex
(Figure 1G through 1I). The mean relative periods of the
existence of these streamlines (tstreamlines) were 0.42⫾0.15 in
manifest PH, 0.29⫾0.07 in latent PH, and 0.15⫾0.05 in
controls (Figure 3) and differed significantly between all
Reiter et al
Blood Flow Patterns in Pulmonary Hypertension
27
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
Figure 3. Comparison of tstreamlines for the manifest PH, latent PH,
and control groups.
groups (P⬍0.0001 for F test). Age dependence was not
significant (P⫽0.10 in analysis of covariance). Interobserver
variability of the measurement of tstreamlines was 0.04, and the
intraclass correlation coefficient was 0.93.
Vortices and mPAP
For all 22 patients in the manifest PH group, the described
vortex was found, whereas this flow pattern was not detected
in any of the subjects without manifest PH (␬-index for
evaluating twice the existence of vortices was 1.00). The use
of the existence of vortices in the primary flow direction in
the main pulmonary artery as a diagnostic criterion for the
presence of manifest PH resulted in a sensitivity of 1.00 and
a specificity of 1.00, with 95% confidence intervals of 0.84 to
1.00 and 0.87 to 1.00, respectively.
The correlation coefficient between tvortex and mPAP of
patients with manifest PH was 0.94 (95% confidence interval
0.85 to 0.97). The corresponding linear regression equation
was mPAP (in millimeters mercury) ⫽ 16.7 ⫹ 58.0 䡠 tvortex,
with an SD of 4.8 mm Hg from the regression line (Figure
4A). Interobserver variability of the measurement of the
relative period of existence of the vortex (tvortex) in the primary
flow direction was 0.03, and the intraclass correlation coefficient was 0.97.
In the blinded prospective study, all 22 patients with
manifest PH had a vortex of blood flow in the main
pulmonary artery, whereas 30 patients without manifest PH
had none. In 3 patients with latent PH, a vortex was detected
(mPAP at rest of 21, 21, and 23 mm Hg with tvortex of 0.05,
0.05, and 0.25, respectively). Consequently, the vortex criterion for the presence of manifest PH led to the same
sensitivity as the exploratory study. Specificity obtained was
0.91 (95% confidence interval 0.76 to 0.98). The comparison
of mPAP measured by catheter and mPAP calculated by the
regression equation resulted in a nonsignificant bias of
⫺0.2 mm Hg, an SD of differences of 3.6 mm Hg, and 95%
limits of agreement of ⫺7.3 and 6.9 mm Hg, respectively
(Figure 4B).
Discussion
The main findings of the exploratory study were as follows:
(1) Manifest PH coincides with the appearance of a vortex of
Figure 4. A, Correlation and linear regression line (together with
95% confidence bounds) of mPAP of patients with manifest PH
and relative period of the existence of a vortex (tvortex) in the primary flow direction in the main pulmonary artery. B, BlandAltman plot of mPAP values measured by catheter and calculated by the regression equation mPAP⫽16.7⫹58.0䡠tvortex (with
95% confidence bounds for bias and 95% limits of agreement)
for the blinded prospective study.
blood flow in the main pulmonary artery; (2) the relative
period of existence of the vortex (tvortex) correlates with mPAP
at rest, with a high correlation coefficient of 0.94 in patients
with manifest PH; and (3) the relative period (tstreamlines) of
continuous diastolic flow upward along the anterior wall of
the main pulmonary artery differed significantly between
patients with manifest PH, those with latent PH, and controls.
Vortex Formation
Elevated mPAP in manifest PH is accompanied by an
anomalous time course of pulse pressure,2,24 and the pressure
gradient between the right ventricle and the pulmonary artery
becomes zero and positive earlier during systole.25,26 This fact
explains the reduced acceleration time in patients with manifest PH, which is the rationale for attempts to assess mPAP
and PH directly by noninvasive methods.2,3,7–11 We also
observed reduced acceleration times in patients with manifest
PH.
Other features of the anomalous time course of the pulse
pressure of patients with manifest PH are consistent with the
observed vortex formation: The results for the position index
28
Circ Cardiovasc Imaging
July 2008
Figure 5. Schematic illustration of flow separation at a boundary
(eg, a vessel wall). When the pressure gradient becomes positive, the boundary layer thickens, and a vortex may develop
within the boundary layer.
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
of the velocity profiles above the pulmonary valve indicate
that in later systole, generally spatial peak velocities move
from a central position toward the anterior wall of the main
pulmonary artery. This in turn may be interpreted as an
increase in the thickness of the boundary layer at the posterior
wall of the main pulmonary artery; in other words, flow
separates from the posterior wall of the main pulmonary
artery. Flow separation at boundaries is a common physical
phenomenon if the driving pressure gradient becomes positive.27 In particular, thickened boundary layers become susceptible to backward flow and the formation of vortices
(Figure 5). During the midsystolic period, when the pressure
gradient between the right ventricle and pulmonary artery is
already close to zero, a short positive wave occurs in the time
course of the pressure gradient for patients with manifest
PH.25,26 We conjecture that this wave, supported by the
present high velocities, triggers vortex formation in the main
pulmonary artery.
Relation Between the Existence of Vortices and
Elevated mPAP
A priori, it is not obvious that the relative period of existence
of the vortex (tvortex) correlates with mPAP; however, PH is
accompanied by increased pulmonary vascular resistance and
decreased compliance of the pulmonary vasculature.28 During
the ejection period of the right ventricle, higher resistance
means that less blood can be released to the peripheral
pulmonary vasculature. Less compliance means that the
released blood cannot be stored adequately in the pulmonary
vasculature by a volume change of the vessels. The vortex
appears to be a natural mechanism to preserve a part of the
kinetic energy of blood, to release the blood to the periphery
during diastole. Because diastolic blood release is again
determined by pulmonary vascular resistance and compliance, it is plausible that both the vortex and the upward flow
persist longer when mPAP is higher.
The correlation coefficient of 0.94 between tvortex and mPAP in
the case of manifest PH means that the quantities almost
exclusively determine one other. The tvortex enables the noninvasive estimation of mPAP via the linear regression equation (in
millimeters mercury: mPAP ⫽ 16.7 ⫹ 58.0 䡠 tvortex) with an SD
of 4.8 mm Hg. This SD is small compared with other noninva-
sive estimates of pulmonary pressure.29 In addition to measurement errors that resulted from the MR sequence technique and
vortex determination, which will be discussed below, deviations
between mPAP determined by catheter and values that were
calculated may have 2 additional causes: First, catheter and MR
measurements were not performed simultaneously, and it has
been shown30 that the variation within long-term mPAP measurements is 8%. Consequently, an SD of ⬎3 mm Hg (resulting
from 8% of our average mPAP value at rest of 41 mm Hg; Table
1) appears to be explainable simply by the physiological change
of mPAP and the variability of its catheter measurements.
Furthermore, the linearity assumption for the relationship between mPAP and the relative period of existence (tvortex) of the
vortex in the main pulmonary artery is quite likely to be violated,
at least for higher mPAPs, especially because the maximal
predictable mPAP via the linear regression equation is
74.8 mm Hg. Other model equations with a steeper slope for
higher mPAPs could be motivated by a general change of the
relation of quantities such as compliance28 or acceleration time7
and mPAP, but a larger group of patients with manifest PH
would be needed to compare different model equations
statistically.
Diastolic Streamlines
Apart from the vortex pattern in the main pulmonary artery,
the study showed significant differences in diastolic streamlines of blood flow in the manifest PH, latent PH, and control
groups. Given the interpretation of the vortex as a “reservoir
of kinetic energy,” the difference between manifest PH and
the other groups is not surprising. It is, however, surprising
that the relative period (tstreamlines) of continuous diastolic flow
upward along the anterior wall of the main pulmonary artery
differed significantly in an investigation at rest between
patients with latent PH and controls and therefore allows
identifying latent PH. In accordance with the explanation of
the phenomenon for manifest PH, the prolonged tstreamlines
reflects changes in pulmonary vascular resistance and compliance of the pulmonary vasculature that are too small to
lead to vortex formation, as in manifest PH. A possible
limitation of the study is that a portion of the tstreamlines
difference between patients with latent PH and control
subjects may be attributed to the different age distributions in
these groups, although age was not significant as a covariate:
Although normal mPAP does not depend significantly on age,
normal pulmonary vascular resistance31 and compliance28 do.
Retrograde Flow
In contrast to previous studies that used MR phase-contrast
measurements to study PH,9 –11 we acquired time-resolved 3D
flow information in the main pulmonary artery to analyze
flow topology–related quantities. For patients with manifest
PH, increased retrograde flow at the posterior wall of the
main pulmonary artery10,11,17,18 can be explained immediately
by the existence of a vortex. We could not define a general
rule for the motion of the vortex throughout the cardiac cycles
which explains why no relation between the retrograde
through-plane velocities and mPAP was found in previous
studies.10,11
Reiter et al
Error Estimate and Verification by the Blinded
Prospective Study
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
As mentioned, part of the SD of 4.8 mm Hg in the linear
regression determination of mPAP via tvortex resulted from the
MR sequence technique and the method of vortex determination: We used a moderate time resolution to keep the
imaging time short and interpolated the retrospectively acquired data to images of 20 cardiac phases. Consequently, a
difference of 1 cardiac phase in the determination of tvortex or
tstreamlines corresponded to at least 5% of the RR interval, and
the minimal difference in mPAP that could be resolved with
this time resolution was 2.9 mm Hg.
The color-encoded vector field representation of measured
velocity fields was appropriate for the interpretation of blood
flow patterns. The existence of vortices and streamlines was
determined visually with excellent interobserver agreement,
which might decrease if catheter results are unknown to the
observers. However, the variability of 3% of the RR interval
for the determination of tvortex contributes a further variability
of 1.7 mm Hg to the mPAP calculation. Summing up the
given error estimates of the physiological change of mPAP,
the variability of its catheter measurements, and the variability of the tvortex determination, the SD of 4.8 mm Hg between
invasively determined and calculated mPAP values is explained fully.
To validate the findings with regard to vortices of blood
flow, a blinded prospective study was performed. Within
confidence limits, this study reproduced the results of the
exploratory study. Nevertheless, the appearance of vortices in
3 patients with latent PH should be discussed. Whereas the
large deviation of measured and calculated mPAP for 1
patient was probably due to physiological variability
(⫺8.2 mm Hg, Figure 4B), 2 patients with a tvortex of 0.05
illustrate a problem associated with the moderate time resolution of MR measurements: Short-living vortices might be
missed, and these vortices might correspond to mPAP values
slightly below 25 mm Hg according to the linear regression
equation (calculated for mPAP values above 25 mm Hg;
Figure 4A).
In summary, 3D visualization of blood flow patterns in the
main pulmonary artery via MR phase-contrast measurements
is a promising tool for noninvasive detection and understanding of PH. It allows the identification of manifest and latent
PH. Elevated mPAP can be determined with physical accuracy by measuring the period of existence of the vortex
associated with PH.
Disclosures
Dr Gert Reiter is employed by Siemens Medical Solutions. The
remaining authors report no conflicts.
References
1. Galiè N, Torbicki A, Barst R, Dartevelle P, Haworth S, Higenbottam T,
Olschewski H, Peacock A, Pietra G, Rubin LJ, Simonneau G, Priori SG,
Garcia MA, Blanc JJ, Budaj A, Cowie M, Dean V, Deckers J, Burgos EF,
Lekakis J, Lindahl B, Mazzotta G, McGregor K, Morais J, Oto A,
Smiseth OA, Barbera JA, Gibbs S, Hoeper M, Humbert M, Naeije R,
Pepke-Zaba J; Task Force on Diagnosis and Treatment of Pulmonary
Arterial Hypertension of the European Society of Cardiology. Guidelines
on diagnosis and treatment of pulmonary arterial hypertension. Eur
Heart J. 2004;25:2243–2278.
Blood Flow Patterns in Pulmonary Hypertension
29
2. Chemla D, Castelain V, Herve P, Lecarpentier Y, Brimioulle S. Haemodynamic evaluation of pulmonary hypertension. Eur Respir J. 2002;20:
1314 –1331.
3. Bossone E, Bodini BD, Mazza A, Allegra L. Pulmonary arterial hypertension: the key role of echocardiography. Chest. 2005;127:1836 –1843.
4. King ME, Braun H, Goldblatt A, Liberthson R, Weyman AE. Intraventricular septal configuration as a predictor of right ventricular systolic
hypertension in children: a cross-sectional echocardiographic study. Circulation. 1983;68:68 –75.
5. Roeleveld RJ, Marcus JT, Faes TJ, Gan TJ, Boonstra A, Postmus TE,
Vonk-Nordegraaf A. Intraventricular septal configuration at MR imaging
and pulmonary arterial pressure in pulmonary hypertension. Radiology.
2005;234:710 –717.
6. Chemla D, Castelain V, Humbert M, Hébert JL, Simonneau G, Lecarpentier Y, Hervé P. New formula for predicting mean pulmonary artery
pressure using systolic pulmonary artery pressure. Chest. 2004;126:
1313–1317.
7. Kitabatake A, Inoue M, Asao M, Masuyama T, Tanouchi J, Morita T,
Mishima M, Uematsu M, Shimazu T, Hori M, Abe H. Noninvasive
evaluation of pulmonary hypertension by a pulsed Doppler technique.
Circulation. 1983;68:302–309.
8. Tardivon AA, Mousseaux E, Brenot F, Bittoun J, Jolivet O, Bourroul E,
Duroux P. Quantification of hemodynamics in primary pulmonary hypertension with magnetic resonance imaging. Am J Respir Crit Care Med.
1994;150:1075–1080.
9. Laffon E, Vallet C, Bernard V, Montaudon M, Ducassou D, Laurent F,
Marthan R. A computed method for noninvasive MRI assessment of
pulmonary arterial hypertension. J Appl Physiol. 2004;96:463– 468.
10. Kondo C, Caputo GR, Masui T, Foster E, O’Sullivan M, Stulbarg MS,
Golden J, Catterjee K, Higgins CB. Pulmonary hypertension: pulmonary
flow quantification and flow profile analysis with velocity-encoded cine
MR imaging. Radiology. 1992;183:751–758.
11. Mousseaux E, Tasu JP, Jolivet O, Simonneau G, Bittoun J, Gaux JC.
Pulmonary arterial resistance: noninvasive measurement with indexes of
pulmonary flow estimated at velocity-encoded MR imaging: preliminary
experience. Radiology. 1999;212:896 –902.
12. Denton CP, Cailes JB, Phillips GD, Wells AU, Black CM, Bois RM.
Comparison of Doppler echocardiography and right heart catheterization
to assess pulmonary hypertension in systemic sclerosis. Br J Rheumatol.
1997;36:239 –243.
13. Laaban JP, Diebold B, Zelinski R, Lafay M, Raffoul H, Rochemaure J.
Non-invasive estimation of systolic pulmonary artery pressure using
Doppler echocardiography in patients with chronic obstructive pulmonary
disease. Chest. 1989;96:1258 –1262.
14. Hinderliter AL, Willis PW, Barst RJ, Rich S, Rubin LJ, Badesch DB,
Groves BM, McGoon MD, Tapson VF, Bourge RC, Brundage BH,
Koerner SK, Langleben D, Keller CA, Murali S, Uretsky BF, Koch G, Li
S, Clayton LM, Jöbsis MM, Blackburn SD, Crow JW, Long WA. Effects
of long-term infusion of prostacyclin (Epoprostenol) on echocardiographic measures of right ventricular structure and function in primary pulmonary hypertension. Circulation. 1997;95:1479 –1486.
15. Nanna M, Lin SL, Tak T, McKay C, Meltzer RS, Rahimtoola SH,
Chandraratna PA. Inaccuracy of Doppler estimates of pulmonary artery
pressure using pulmonary flow acceleration time. Can J Cardiol. 1990;
6:19 –23.
16. Torbicki A, Kurzyna M, Ciurzynski M, Pruszczyk P, Pacho R, KuchWocial A, Szulc M. Proximal pulmonary emboli modify right ventricular
ejection pattern. Eur Respir J. 1999;13:616 – 621.
17. Bogren HG, Klipstein RH, Mohiaddin RH, Firmin DN, Underwood SR,
Rees RS, Longmore DB. 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;118:990 –999.
18. Mohiaddin RH, Yang GZ, Kilner PJ. Visualization of flow by vector
analysis of multidirectional cine MR velocity mapping. J Comput Assist
Tomogr. 1994;18:383–392.
19. Pelc NJ, Herfkens RJ, Shimakawa A, Enzmann DR. Phase contrast cine
magnetic resonance imaging. Magn Reson Q. 1991;7:229 –254.
20. Kim WY, Walker PG, Pedersen EM, Poulsen JK, Oyre S, Houlind K,
Yoganathan AP. Left ventricular blood flow patterns in normal subjects:
a quantitative analysis by three-dimensional magnetic resonance velocity
mapping. J Am Coll Cardiol. 1995;26:224 –238.
21. Kilner PJ, Yang GZ, Wilkes AJ, Mohiaddin RH, Firmin DN, Yacoub
MH. Asymmetric redirection of flow through the heart. Nature. 2000;
404:759 –761.
30
Circ Cardiovasc Imaging
July 2008
22. Hudsmith LE, Petersen SE, Francis JM, Robson MD, Neubauer S.
Normal human left and right ventricular and left atrial dimensions using
steady state free precession magnetic resonance imaging. J Cardiovasc
Magn Reson. 2005;7:775–782.
23. Reiter G, Reiter U, Kainz B, Greiser A, Bischof H, Rienmüller R. MR
vector field measurement and visualization of normal and pathological
time-resolved three-dimensional cardiovascular blood flow patterns.
J Cardiovasc Magn Reson. 2007;9:237–238. Abstract.
24. Castelain V, Hervé P, Lecarpentier Y, Duroux P, Simonneau G, Chemla
D. Pulmonary artery pulse pressure and wave reflection in chronic pulmonary thromboembolism and primary pulmonary hypertension. J Am
Coll Cardiol. 2001;37:1085–1092.
25. Tahara M, Tanaka H, Nakao S, Yoshimura H, Sakurai S, Tei C, Kashima
T. Hemodynamic determinants of pulmonary valve motion during systole
in experimental pulmonary hypertension. Circulation. 1981;64:
1249 –1255.
26. Turkevich D, Groves BM, Micco A, Trapp JA, Reeves JT. Early partial
systolic closure of the pulmonic valve relates to severity of pulmonary
hypertension. Am Heart J. 1988;115:409 – 418.
27. Oertel H. Dynamics of fluid flow. In: Oertel H, ed. Prandtl’s Essentials
of Fluid Mechanics. 2nd ed. New York, NY: Springer; 2004:63–216.
28. Reuben SR. Compliance of the human pulmonary arterial system in
disease. Circ Res. 1971;29:40 –50.
29. Naeije R, Torbicki A. More on the noninvasive diagnosis of pulmonary
hypertension: Doppler echocardiography revisited. Eur Respir J. 1995;8:
1445–1449.
30. Rich S, D’Alonzo GE, Dantzker DR, Levy PS. Magnitude and implications of spontaneous hemodynamic variability in primary pulmonary
hypertension. Am J Cardiol. 1985;55:159 –163.
31. Ehrsam RE, Perruchoud A, Oberholzer M, Burkart F, Herzog H.
Influence of age on pulmonary haemodynamics at rest and during supine
exercise. Clin Sci (Lond). 1983;65:653– 660.
CLINICAL PERSPECTIVE
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
Pulmonary hypertension is associated with a poor prognosis. Currently, the only established method to diagnose pulmonary
hypertension is to directly measure pulmonary arterial pressure on right heart catheterization. We present a new method
to measure elevated mean pulmonary arterial pressure noninvasively by magnetic resonance velocity imaging. Our findings
suggest that (1) manifest pulmonary hypertension coincides with the appearance of a vortex of blood flow in the main
pulmonary artery; (2) the relative period of existence of this vortex correlates in a highly significantly manner with mean
pulmonary arterial pressure; and (3) the relative period of continuous antegrade diastolic flow along the anterior wall of
the main pulmonary artery is significantly prolonged in patients with latent pulmonary hypertension compared with
controls. These findings suggest that flow patterns in the main pulmonary artery allow not only the identification of
manifest and latent pulmonary hypertension but also the measurement of elevated mean pulmonary arterial pressures.
Magnetic Resonance−Derived 3-Dimensional Blood Flow Patterns in the Main Pulmonary
Artery as a Marker of Pulmonary Hypertension and a Measure of Elevated Mean
Pulmonary Arterial Pressure
Gert Reiter, Ursula Reiter, Gabor Kovacs, Bernhard Kainz, Karin Schmidt, Robert Maier, Horst
Olschewski and Rainer Rienmueller
Downloaded from http://circimaging.ahajournals.org/ by guest on May 2, 2017
Circ Cardiovasc Imaging. 2008;1:23-30
doi: 10.1161/CIRCIMAGING.108.780247
Circulation: Cardiovascular Imaging is published by the American Heart Association, 7272 Greenville Avenue,
Dallas, TX 75231
Copyright © 2008 American Heart Association, Inc. All rights reserved.
Print ISSN: 1941-9651. Online ISSN: 1942-0080
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circimaging.ahajournals.org/content/1/1/23
Data Supplement (unedited) at:
http://circimaging.ahajournals.org/content/suppl/2008/07/21/1.1.23.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation: Cardiovascular Imaging can be obtained via RightsLink, a service of the Copyright Clearance
Center, not the Editorial Office. Once the online version of the published article for which permission is being
requested is located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation: Cardiovascular Imaging is online at:
http://circimaging.ahajournals.org//subscriptions/