Key words
hemodynamics/flow dynamics - MR-imaging - aorta
Abbreviations
2 D flow MRI:
time-resolved 2-dimensional phase contrast flow
MRI:
magnetic resonance imaging
4 D flow MRI:
time-resolved 3-dimensional phase contrast flow
kt-GRAPPA:
GeneRalized Autocalibrating Partially Parallel Acquisition with linear interpolation
of missing data in the k-space
FFV:
forward flow volume
BFV:
backward flow volume
NVF:
net flow volume
VAX:
axial velocity
VABS:
absolute velocity
Introduction
Time resolved 2-dimensional phase contrast flow measurements (2 D flow MRI) play a
major role in the assessment of cardiovascular pathologies in magnetic resonance imaging
(MRI) [1]. Although singular 2 D flow MRI measurements can be performed rather fast, it is
necessary to perform an individual measurement of each vessel of interest. Furthermore,
it is mandatory to acquire each 2 D flow MRI measurement exactly perpendicular to
the main flow vector. In the setting of congenital heart disease, e. g., in aortic
isthmic stenosis (coarctation) with a possible elongation of the aorta or otherwise
altered anatomic conditions, it can be a quite time-consuming process, if you need
one in-plane and maybe three through-plane measurements pre-, intra- and post-stenotic.
Misaligned measurements lead to inaccurate results and might need to be repeated [2].
Similar to conventional 2 D flow MRI sequences, time-resolved 3-dimensional phase
contrast flow measurements (4 D flow MRI) enable absolute quantification of flow parameters,
such as forward and backward flow volumes, flow velocities, and shunt volumes [3]
[4]. In addition, 4 D flow MRI enables quantification of advanced flow parameters, such
as helical flow, wall shear stress, and analysis of flow displacement with full coverage
of the complete vascular systems, such as the great mediastinal vessels [5]. In contrast to 2 D flow MRI, isotropic data in all spatial directions can be obtained
in 4 D flow MRI, making it possible to create 3 D reconstructions of every vessel
within a given field of view after data acquisition. This enables off-line placement
of measuring planes during post-processing for retrospective evaluation of blood flow-like
flow volumes and velocities in multiple vessels [6]. With these reconstructions from 4 D flow MRI data sets, results are independent
of plane angulation or flow directions, enabling fast forward planning during image
acquisition. Numerous studies demonstrated good agreement between 2 D and 4 D flow
MRI measurements in healthy volunteers and patients so that 4 D flow MRI could replace
conventional 2 D flow MRI [7]
[8]
[9]. However, 4 D flow MRI is still neither an established part of the evaluation of
patients, nor part of the decision-making process. To overcome this issue and to achieve
a broader use of 4 D flow MRI in daily practice, it could be helpful to define normal
values of “2 D flow MRI parameters” derived from 4 D flow MRI, like forward and backward
flow volumes and flow velocities. These normal ranges may differ substantially, because
not only one flow direction as in 2 D flow MRI is incorporated but rather flow components
in all spatial directions are used. This study is an extension of previous work, in
which we evaluated the used 4 D flow MRI sequence against a flow phantom using pulsatile
and non-pulsatile flow and the standard of care 2 D flow MRI and found no significant
differences between 2 D and 4 D flow MRI measurements [8]. The aim of this study was to utilize 4 D flow MRI in the thoracic aorta of healthy
volunteers to generate normal values of the “2 D flow MRI parameters” derived from
4 D flow MRI in order to possibly replace multiple 2 D flow MRI measurements (e. g.,
in the setting of aortic isthmic stenosis (coarctation)) by one 4 D flow MRI measurement.
Materials and methods
Study cohort
116 healthy volunteers with no history of cardiovascular disease (58 females, mean
age 43 ± 13 years) were included. To investigate the influence of age, we included
participants in various age groups: 19–30 (n = 24); 31–40 (n = 28); 41–50 (n = 26);
51–60 (n = 22), and ≥ 61 years (n = 16). The local ethics board approved the study
and written informed consent was obtained from all participants.
Magnetic resonance image acquisition
The 4 D flow MRI datasets were acquired at 3 Tesla using a 16-channel surface coil
in combination with a 12-element spine coil (Magnetom Verio Dot, Siemens Healthcare
GmbH, Erlangen, Germany). The used 4 D flow MRI kt-GRAPPA5 (GeneRalized Autocalibrating Partially Parallel Acquisition) sequence is commercially available and was validated before ex vivo using a flow phantom with pulsatile and non-pulsatile flows and in vivo in healthy volunteers against the standard of care (2 D flow MRI) [8]
[9]. The imaging parameters were: TR = 4.6 ms, TE = 2.8 ms, flip angle 10°, FOV 320x240 mm
with a mean temporal resolution of 39.2 ms (35.3–39.9 ms), and a spatial resolution
of 2.5 × 2.5 × 2.5 mm3, VENC 150 cm/s, phase encoding direction anterior to posterior, slice number 24.
Respiratory gating was performed using a navigator on the diaphragm/liver interface.
As standard of care, 2 D flow MRI measurements were performed: TR = 8.5 ms, TE = 2.9 ms,
flip angle 15°, with a mean temporal resolution of 20.5 ms, spatial resolution of
1.9 × 1.9 mm2, slice thickness 5 mm, VENC 150 cm/s.
Data analysis
Vessel segmentation, blood flow visualization, and pre-processing
All processing and measurement steps were carried out using the software Bloodline
(University of Magdeburg, Germany) [10]. The segmentation of the aorta and the placement of the centerline were performed
automatically as described by Köhler et al. [10]. The ascending aorta was defined as the volume of the aorta between the aortic valve
and the origin of the brachiocephalic trunk, the aortic arch was defined as the volume
between the origin of the brachiocephalic trunk and the left subclavian artery, and
the thoracic descending aorta was defined as the volume between the origin of the
left subclavian artery and the diaphragm. A centerline was semi-automatically drawn
through the whole thoracic aorta. Aortic blood flow was visualized using time-resolved
pathlines. We corrected for eddy currents and background noise as described previously
[11]
[12].
Measurements and flow quantifications
The software tool “Bloodline” enables positioning of multiple measuring planes for
the assessment of flow volumes and flow velocities as mentioned elsewhere [10]. Measuring planes were positioned at specific landmarks ([Fig. 1]) as follows: In the middle of the ascending aorta (plane A), behind the origin of
the brachiocephalic artery (plane B), behind the origin of the left common carotid
artery (plane C), behind the origin of the left subclavian artery (plane D), and at
the level of the diaphragm (plane E). All measuring planes were oriented perpendicular
to the centerline of the thoracic aorta ([Fig. 1]). 2 D flow MRI measurements were performed on planes A and D. Planes A and D in
2 D and 4 D flow MRI were matched manually by the investigator.
Fig. 1 Visualization of intraaortic blood flow with time-resolved 3 D pathlines. Measuring
planes in the mid-ascending aorta (green, Plane A), in the aortic arch behind the origin of the brachiocephalic trunk (blue, Plane
B), behind the origin of the left carotid artery (red, Plane C), behind the origin of the left subclavian artery (white, Plane D), and in the descending aorta on the level of the diaphragm (purple, Plane E).
Abb. 1 Visualisierung des intraaortalen Blutflusses mit zeitlich aufgelösten 3 D pathlines.
Messebenen in der mittleren Aorta ascendens (grün, Ebene A), im Aortenbogen hinter dem Ursprung des Truncus brachiocephalicus (blau, Ebene B), hinter dem Ursprung der linken Arteria carotis communis (rot, Ebene C), hinter dem Ursprung der linken Arteria subclavia (weiß, Ebene D) und in der Aorta descendens auf Höhe des Zwerchfells (lila, Ebene E).
Flow volumes were assessed by measuring the forward flow volume (FFV), the backward flow volume (BFV), and the net flow volume (NFV) in [ml/heartbeat] that passes through the above-mentioned measuring planes. Net
flow was calculated by subtracting BFV from FFV. The regurgitant fraction (RF) was calculated by dividing BFV by FFV in [%]. As quality control, comparisons of
NFV in the main pulmonary artery and the ascending aorta has been performed in 10
datasets (Qp/Qs) as suggested in a consensus statement by Dyverfeldt et al. [12].
Flow velocities were assessed with different parameters:
The parameter maximum axial velocity (VAX
) describes the maximum velocity in [m/s] of blood flow that passes strictly perpendicularly
(axial) through a measuring plane, corresponding to the “classic” 2 D flow MRI velocity
acquired from 2 D flow MRI sequences. This parameter describes the through-plane component
of the velocity vector. The maximum velocity was defined as the local velocity maximum
on a measurement plane at the timepoint of maximum blood flow.
The 4 D flow MRI parameter maximum absolute velocity (VABS
) is defined as the maximum velocity in [m/s] of blood flow through a measuring plane
independent of the flow orientation. This parameter also contains the radial and the
circumferential component of the blood flow (Supplemental Fig. 1).
Intra‑ and interobserver variability
For the assessment of the inter-observer variability, a second investigator with > 9
years of cardiac MRI experience analyzed a subgroup of 10 randomly selected datasets,
according to the above-described methodology (including all import and segmentation
steps). For the assessment of the intra-observer variability, this investigator repeated
all measurements 10 days after the first assessment. These assessments have been performed
without further differentiation regarding age or sex.
Statistical analysis
All results were given as their mean values and standard deviation (SD). Statistical
analysis was performed using the statistical software package SAS 9.4 (SAS Institute
Inc., Cary, NC, USA). In a first step, normal distribution was confirmed using the
Shapiro-Wilk test. Differences between male and female volunteers and between 2 D
and 4 D flow MRI were assessed using a paired t-test. A p-value < 0.05 was considered
statistically significant. Dependencies between variables and age or sex were assessed
using Pearson’s correlations. Since all measured values were distributed normally,
normal ranges were given as mean ± 2SD. Intra- and interobserver variability was assessed
using interclass correlation (ICC) and were given as correlation coefficient R.
Results
Volunteer characteristics:
116 volunteers were included: 58 females, mean age 43 ± 13 years. The mean body mass
index was 24.6 ± 5.2 kg/m2. The mean resting heart rate during the examination was 69 ± 12/ min., the mean cardiac
output was 6.45 ± 1.59 l/min as measured by 4 D flow MRI.
Data acquisition
The mean scan time was 8.4 ± 4.2 minutes for the 4 D flow MRI sequence. All datasets
were completely assessable.
Flow volumes
The female and male normal ranges of all measuring planes are given in [Table 1], [2], respectively. At the level of the ascending aorta (plane A), the overall mean FFV,
BFV, and NFV were generally slightly, but not significantly higher with 93.5 ± 14.8,
3.6 ± 2.8, and 89.9 ± 15.8 ml/heartbeat using the 4 D flow MRI sequence as compared
to 90.1 ± 13.4, 3.2 ± 3.1, and 85.7 ± 16.2 ml/heartbeat using the 2 D flow MRI sequence.
The overall mean regurgitant fraction (RF) was 3.9 ± 2.9 % (4 D flow MRI) and 3.6 ± 3.4 %
(2 D flow MRI). The mean Qp/Qs in a subset of 10 datasets was 1.2 ± 0.1.
Table 1
Mean 2 D flow MRI volumes derived from 4 D flow MRI including normal ranges of female
healthy volunteers.
Tab. 1 Aus der 4D-Fluss MRT abgeleitete 2D-Fluss-MRT-Volumina und Normbereiche von weiblichen
gesunden Probanden.
|
Parameter in female volunteers
|
Measuring plane
|
Mean (95 % CI)
|
SD
|
Lower limit
|
Upper limit
|
|
Forward flow volume [ml/heartbeat]
|
Plane A
|
84.8 (80.1–89.6)
|
13.1
|
58.6
|
111.1
|
|
Plane B
|
68.9 (64.4–73.4)
|
12.4
|
44.2
|
93.7
|
|
Plane C
|
64.4 (62.1–76.7)
|
13.1
|
38.2
|
84.6
|
|
Plane D
|
56.8 (51.9–61.7)
|
11
|
34.9
|
78.7
|
|
Plane E
|
55.1 (49.9–60.3)
|
10.7
|
33.7
|
76.6
|
|
Backward flow volume [ml/heartbeat]
|
Plane A
|
2.8 (2.3–3.3)
|
1.1
|
0.6
|
5.0
|
|
Plane B
|
2.2 (1.8–2.6)
|
1.0
|
0.1
|
4.3
|
|
Plane C
|
3.6 (3.2–4.0)
|
0.9
|
1.9
|
5.3
|
|
Plane D
|
3.8 (3.6–4.0)
|
1.8
|
0.5
|
7.1
|
|
Plane E
|
3.6 (3.2–4.0)
|
0.9
|
1.8
|
5.4
|
|
Regurgitant fraction [%]
|
Plane A
|
3.3 (2.9–3.7)
|
1.3
|
0.7
|
5.9
|
|
Plane B
|
3.2 (2.8–3.6)
|
1.5
|
0.2
|
3.2
|
|
Plane C
|
5.6 (4.6–6.7)
|
1.6
|
2
|
8.8
|
|
Plane D
|
6.7 (5.9–7.5)
|
3.2
|
0.3
|
13.1
|
|
Plane E
|
6.6 (5.8–7.3)
|
1.8
|
3.0
|
10.2
|
|
Net flow volume [ml / heartbeat]
|
Plane A
|
82.1 (78.4–85.8)
|
14.6
|
53.0
|
111.2
|
|
Plane B
|
66.7 (63.9–70.5)
|
13
|
40.8
|
92.6
|
|
Plane C
|
60.8 (57.2–64.4)
|
14.5
|
31.9
|
89.8
|
|
Plane D
|
55.1 (53.2–57.0)
|
11.9
|
31.3
|
77.9
|
|
Plane E
|
53.5 (50.4–56.6)
|
12.6
|
28.3
|
75.7
|
Table 2
Mean 2 D flow MRI volumes derived from 4 D flow MRI including normal ranges of male
healthy volunteers.
Tab. 2 Aus der 4D-Fluss MRT abgeleitete 2D-Fluss-MRT-Volumina und Normbereiche von männlichen
gesunden Probanden.
|
Parameter in male volunteers
|
Measuring plane
|
Mean (95 % CI)
|
SD
|
Lower limit
|
Upper limit
|
|
Forward flow volume [ml/heartbeat]
|
Plane A
|
102.2 (100.2–104.3)
|
16.4
|
69.4
|
135.1
|
|
Plane B
|
91.9 (88.3–94.7)
|
20.3
|
51.2
|
112.4
|
|
Plane C
|
71.1 (69.6–73.6)
|
13.9
|
43.4
|
98.7
|
|
Plane D
|
65.8 (63.7–67.3)
|
13.4
|
39.0
|
92.4
|
|
Plane E
|
63.8 (60.2–67.3)
|
13.2
|
37.5
|
90.2
|
|
Backward flow volume [ml/heartbeat]
|
Plane A
|
4.2 (4.0–4.4)
|
1.9
|
0.4
|
8.0
|
|
Plane B
|
5.3 (5.0–5.6)
|
1.5
|
2.3
|
8.3
|
|
Plane C
|
4.3 (3.9–4.8)
|
0.8
|
2.7
|
5.9
|
|
Plane D
|
4.2 (3.5–4.9)
|
1.7
|
0.8
|
7.6
|
|
Plane E
|
3.8 (3.2–4.4)
|
1.4
|
1.0
|
5.6
|
|
Regurgitant fraction [%]
|
Plane A
|
4.1 (3.9–4.3)
|
1.9
|
0.3
|
7.9
|
|
Plane B
|
5.8 (5.6–6.0)
|
1.6
|
2.2
|
9.4
|
|
Plane C
|
6.0 (5.7–6.3)
|
1.1
|
3.8
|
8.2
|
|
Plane D
|
6.4 (6.0–6.8)
|
2.6
|
1.2
|
11.6
|
|
Plane E
|
5.9 (5.5–6.3)
|
2.2
|
1.5
|
10.3
|
|
Net flow volume [ml/heartbeat]
|
Plane A
|
98.1 (96.2–100.1)
|
16.9
|
64.4
|
131.7
|
|
Plane B
|
86.6 (84.4–88.8)
|
14.3
|
58.0
|
115.2
|
|
Plane C
|
66.9 (61.5–72.3)
|
14
|
38.9
|
94.9
|
|
Plane D
|
61.6 (58.2–64.9)
|
12.1
|
37.4
|
85.8
|
|
Plane E
|
59.9 (56.3–63.5)
|
12.5
|
34.9
|
84.9
|
These values were significantly higher in males compared to female volunteers (p < 0.05).
The FFV, BFV, and NFV were 102.2 ± 16.4, 4.2 ± 1.9, and 98.1 ± 16.9 ml/heartbeat in
males and 84.8 ± 13.1, 2.8 ± 1.1, and 82.1 ± 14.6 ml/heartbeat in females using 4 D
flow MRI and 100.9 ± 15.9, 3.8 ± 1.6, and 97.1 ± 14.3 ml/heartbeat in males and 82.0 ± 14.5,
2.2 ± 1.6, and 78.8 ± 13.3 ml/heartbeat in females using 2 D flow MRI. The RF in 4 D
and 2 D flow MRI was 4.1 ± 1.9 and 3.8 ± 1.6 % in males and 3.3 ± 1.3 and 2.7 ± 2.0 %
in females ([Table 3]). There were no relevant correlations between age and FFV, BFV, or NFV.
Table 3
Comparison of flow volumes using 2 D flow and 4 D flow at measuring planes A and D.
Tab. 3 Vergleich der Flussvolumina von 2D-Fluss und 4D-Fluss in den Messebenen A und D.
|
Parameter
|
Measuring plane
|
male
|
female
|
|
4 D flow mean ± SD
|
2 D flow mean ± SD
|
Comparison 4 D vs. 2 D flow p-value
|
4 D flow mean ± SD
|
2 D flow mean ± SD
|
Comparison 4 D vs. 2 D flow p-value
|
|
Forward flow volume [ml/heartbeat]
|
Plane A
|
102.2 ± 16.4
|
100.9 ± 15.9
|
> 0.05
|
84.8 ± 13.1
|
82.0 ± 14.5
|
> 0.05
|
|
Plane D
|
65.8 ± 13.4
|
66.2 ± 15.1
|
> 0.05
|
56.8 ± 11.0
|
52.9 ± 13.3
|
> 0.05
|
|
Backward flow volume [ml/heartbeat]
|
Plane A
|
4.2 ± 1.9
|
3.8 ± 1.6
|
> 0.05
|
2.8 ± 1.1
|
2.2 ± 1.6
|
> 0.05
|
|
Plane D
|
4.2 ± 1.7
|
3.8 ± 1.1
|
> 0.05
|
3.8 ± 1.8
|
3.5 ± 2.1
|
> 0.05
|
|
Regurgitant fraction [%]
|
Plane A
|
4.1 ± 1.9
|
3.8 ± 1.6
|
> 0.05
|
3.3 ± 1.3
|
2.7 ± 2.0
|
> 0.05
|
|
Plane D
|
6.4 ± 2.6
|
5.7 ± 1.7
|
> 0.05
|
6.7 ± 3.2
|
6.6 ± 4.0
|
> 0.05
|
|
Net forward flow volume [ml/heartbeat]
|
Plane A
|
98.1 ± 16.9
|
97.1 ± 14.3
|
> 0.05
|
82.1 ± 14.6
|
78.8 ± 13.3
|
> 0.05
|
|
Plane D
|
61.6 ± 12.1
|
62.4 ± 14.7
|
> 0.05
|
55.1 ± 11.9
|
49.4 ± 14.1
|
> 0.05
|
Flow velocities
The female and male normal ranges of all measuring planes are given in [Table 4], [5], respectively. At the level of the ascending aorta (plane A), the overall VAX
and VABS
were 1.01 ± 0.31 and 1.23 ± 0.35 m/s using the 4 D flow MRI sequence and the overall
VAX
using the 2 D sequence was 1.11 ± 0.38 m/s. There were no significant differences
regarding sex and age (p < 0.05, respectively) and no significant differences between
4 D flow MRI and 2 D flow MRI (p < 0.05). There were no datasets with phase wraps.
Therefore, no aliasing correction had to be performed.
Table 4
Flow velocities including normal ranges of male and female healthy volunteers.
Tab. 4 Flussgeschwindigkeiten und Normbereiche von männlichen und weiblichen gesunden Probanden.
|
Parameter in male and female volunteers
|
Measuring plane
|
Mean (95 % CI)
|
SD
|
Lower limit
|
Upper limit
|
|
Maximum axial velocity [m/s]
|
Plane A
|
1.01 (0.97–1.05)
|
0.31
|
0.39
|
1.63
|
|
Plane B
|
1.11 (1.07–1.15)
|
0.36
|
0.40
|
1.70
|
|
Plane C
|
0.98 (0.93–1.03)
|
0.14
|
0.41
|
1.89
|
|
Plane D
|
0.95 (0.91–0.99)
|
0.3
|
0.35
|
1.67
|
|
Plane E
|
0.90 (0.84–0.96)
|
0.27
|
0.36
|
1.68
|
|
Maximum absolute velocity [m/s]
|
Plane A
|
1.23 (1.17–1.29)
|
0.35
|
0.53
|
1.93
|
|
Plane B
|
1.19 (1.11–1.27)
|
0.36
|
0.48
|
1.98
|
|
Plane C
|
1.05 (1.00–1.10)
|
0.3
|
0.45
|
1.97
|
|
Plane D
|
1.06 (1.00–1.11)
|
0.26
|
0.55
|
1.88
|
|
Plane E
|
1.05 (1.00–1.10)
|
0.26
|
0.54
|
1.98
|
Table 5
Comparison of flow velocities using 2 D flow and 4 D flow at measuring planes A and
D.
Tab. 5 Vergleich der Durchflussgeschwindigkeiten bei 2D- und 4D-Fluss in den Messebenen
A und D.
|
Parameter
|
Measuring plane
|
Overall volunteers
|
|
4 D flow mean ± SD
|
2 D flow mean ± SD
|
|
Maximum axial velocity [m/s]
|
Plane A
|
1.01 ± 0.31
|
1.11 ± 0.38
|
|
Plane D
|
0.95 ± 0.27
|
0.97 ± 0.27
|
|
Maximum absolute velocity [m/s]
|
Plane A
|
1.23 ± 0.35
|
|
|
Plane D
|
1.06 ± 0.26
|
|
Inter- and intraobserver variability
The interobserver variabilities ranged from R = 0.91 to 0.99 (p < 0.05) for the forward
and backward flow volumes and from R = 0.89 to 0.99 (p < 0.05) for the flow velocities.
The intraobserver variabilities ranged from R = 0.94 to 0.99 (p < 0.05) for the forward
and backward flow volumes and from R = 0.93 to 0.99 (p < 0.05) for the flow velocities
on the different measuring planes ([Table 6]).
Table 6
Inter- and intraobserver correlations for 4 D flow MRI measurements of flow volumes
and flow velocities.
Tab. 6 Inter- und Intrauntersucher-Varianz der 4D-MRT-Messungen von Flussvolumina und -geschwindigkeiten.
|
|
Measuring plane
|
Interobserver correlation, correlation coefficient R, and p-value
|
Intra-observer correlation, correlation coefficient R, and p-value
|
|
Forward flow volume [ml/heartbeat]
|
Plane A
|
R = 0.95; p < 0.05
|
R = 0.98; p < 0.05
|
|
Plane B
|
R = 0.94, p < 0.05
|
R = 0.97; p < 0.05
|
|
Plane C
|
R = 0.91, p < 0.05
|
R = 0.96; p < 0.05
|
|
Plane D
|
R = 0.95; p < 0.05
|
R = 0.96; p < 0.05
|
|
Plane E
|
R = 0.99; p < 0.05
|
R = 0.99; p < 0.05
|
|
Backward flow volume [ml/heartbeat]
|
Plane A
|
R = 0.94; p < 0.05
|
R = 0.95; p < 0.05
|
|
Plane B
|
R = 0.95, p < 0.05
|
R = 0.94; p < 0.05
|
|
Plane C
|
R = 0.91, p < 0.05
|
R = 0.97; p < 0.05
|
|
Plane D
|
R = 0.95; p < 0.05
|
R = 0.94; p < 0.05
|
|
Plane E
|
R = 0.99; p < 0.05
|
R = 0.99; p < 0.05
|
|
Maximum axial velocity [m/s]
|
Plane A
|
R = 0.92; p < 0.05
|
R = 0.94; p < 0.05
|
|
Plane B
|
R = 0.91, p < 0.05
|
R = 0.97; p < 0.05
|
|
Plane C
|
R = 0.89, p < 0.05
|
R = 0.93; p < 0.05
|
|
Plane D
|
R = 0.97; p < 0.05
|
R = 0.95; p < 0.05
|
|
Plane E
|
R = 0.99; p < 0.05
|
R = 0.99; p < 0.05
|
|
Maximum absolute velocity [m/s]
|
Plane A
|
R = 0.93; p < 0.05
|
R = 0.97; p < 0.05
|
|
Plane B
|
R = 0.90, p < 0.05
|
R = 0.97; p < 0.05
|
|
Plane C
|
R = 0.91, p < 0.05
|
R = 0.94; p < 0.05
|
|
Plane D
|
R = 0.92; p < 0.05
|
R = 0.95; p < 0.05
|
|
Plane E
|
R = 0.98; p < 0.05
|
R = 0.99; p < 0.05
|
Discussion
The aim of this study was to utilize 4 D flow MRI measurements to acquire normal ranges
of “conventional 2 D flow MRI parameters” regarding aortic blood flow volumes and velocities in healthy volunteers. The used
4 D flow MRI sequence has been evaluated before in a phantom study at pulsatile and
non-pulsatile flow, in healthy volunteers and patients with bicuspid aortic valves
including head-to-head comparisons between 4 D and 2 D flow MRI as the current standard
of care [8]
[9]
[13].
However, thorough validations of the MRI sequences being used are required for reporting
normal values. Recent studies suggest that there is a relevant variability in 4 D
flow MRI-derived stroke volume and flow velocity using different sequences and different
scanners [14]
[15]. There are numerous studies regarding the validation of 4 D flow MRI sequences against
2 D flow MRI in vivo showing no significant differences between 2 D flow MRI and 4 D flow MRI, and excellent
correlation between both techniques with a correlation coefficient of up to R = 0.98
has been found [13]
[16]
[17]. Contrary to those results, other groups reported that 4 D flow MRI significantly
underestimates systolic peak flow velocities, while 2 D flow MRI gives accurate results
[18]
[19]. Other groups found significant underestimation of aortic or pulmonary regurgitation
and intracardiac flow when using 4 D flow MRI measurements [20]
[21]
[22]. It is not clear if those differences occur when using sequences and scanners from
different vendors or if there are other explanations. However, those results underline
that thorough validation and quality control of the MRI sequences being used are mandatory.
In this current study 2 D flow MRI measurements have been performed at 2 specific
landmarks (plane A and D) for quality control without significant differences between
both techniques. As further quality control, Qp/Qs in 10 datasets has been measured
and we found no abnormalities. In a previous study, our group validated the used 4 D
flow MRI sequence against 2 D flow MRI sequences using a custom-made flow phantom
with pulsatile and non-pulsatile flow and found no significant differences between
both techniques and an excellent correlation between flow measurements and the reference
given by the flow phantom [8]. Therefore, the assumption that flow volumes and velocities derived from 4 D flow
MRI might differ substantially from 2 D flow MRI measurements because 4 D flow MRI
includes flow volumes and velocities in all spatial directions can be rejected for
most parameters. However, we were able to demonstrate an effect on the overall results
with a tendency towards generally higher flow volumes in the flow volumes derived
from 4 D flow MRI as compared to the conventional 2 D flow MRI volumes, but these
differences were not statistically significant and can therefore be neglected, except
for the parameter VABS
, which provided approximately 20 % higher velocities than VAX
. VABS
should therefore not be used interchangeably. Since the comparability of commercially
available 4 D flow MRI sequences with standard 2 D flow MRI sequences has been demonstrated
in many studies before, including by our own group and in this current study, there
is sufficient data to justify the use of 4 D flow MRI in the clinical routine and
to replace 2 D flow MRI. However, one major drawback of 4 D flow MRI measurements
is still the rather long acquisition time. Early 4 D flow MRI sequences took more
than 17 minutes for the assessment of the mediastinal vessels [9]. Newer sequences with more sophisticated acceleration techniques take significantly
shorter acquisition times and are therefore more suitable for clinical applications.
In this current study, the mean scan time was 8.4 minutes using a kt-GRAPPA 5 accelerated
sequence. In the clinical routine, for a complete assessment of blood flow of the
great mediastinal vessels, this is an acceptable duration. In addition, compared to
2 D flow MRI acquisitions, 4 D flow MRI acquisitions enable planning in a “straight forward” manner without the need for exact angulation of the measuring planes, therefore
saving time during the examination if multiple measurements have to be performed or
if altered anatomic conditions complicate the exact perpendicular planning of 2 D
measuring planes. In addition to that, a single 4 D flow MRI measurement includes
information about all other vessels within the field of view like, e. g., the pulmonary
artery and enables 3 D reconstructions of the vessel of interest ([Fig. 1]).
A recent study by Kroeger et al. elucidated stroke volumes in 44 healthy volunteers
using 4 D flow MRI and found a mean value of 72 ± 13.5 ml/heartbeat to be normal [23]. Kroeger did not differentiate between male and female participants, and in our
study, we found slightly higher sex-neutral normal values with 93.5 ± 14.8 ml/ heartbeat.
Nevertheless, the normal mean value of 72 ± 13.5 ml/heartbeat published by Kroeger’s
et al. lies within our normal range in both male and female volunteers at the aortic
plane level A (58.6–135.1 ml/heartbeat) – [Table 1], [2]. Another study reports flow volumes normalized to the body surface area of 126 healthy
individuals [24]. However, since in most guidelines, like the ACC/AHA Guideline for the Management
of Patients With Valvular Heart Disease, cut-off values for the grading of aortic
stenosis or regurgitation are given as absolute values in ml and m/s and are not normalized
to the body surface area, we waived the normalization [25].
We found a mean backward flow volume of 4.2 ± 1.9 ml/heartbeat and 2.8 ± 1.1 ml/heartbeat
to be normal in healthy males and females in the ascending aorta, which is equivalent
to a mean regurgitant fraction of 4.1 ± 1.9 % and 3.3 ± 1.3 %, respectively. The overall
sex-neutral mean regurgitant fraction was 3.9 ± 2.9 % ranging from 0–9.7 %. This is
in line with other studies highlighting that small aortic regurgitation should not
be taken as a pathologic finding and that regurgitant volumes of < 30 ml/heartbeat
and RF < 30 % should be considered mild [26]
[27]
[28].
Regarding flow velocities, we analyzed the axial flow velocity (VAX
), which describes blood flow that moves exactly perpendicular to the measurement
plane comparable to a standard 2 D flow MRI acquisition. We found mean values of 1.01 ± 0.31 m/s
to be normal, which is in line with many other sources, reporting comparable values
in mostly smaller cohorts [29]
[30]
[31]
[32]
[33]
[34]: Based on the analysis of 16 volunteers, Bollache et al. reported in 2016 0.96 ± 0.24 m/s
to be normal and in 2014 Schnell et al. found 1.06 ± 0.2 m/s to be normal [31]
[35]. Utilizing 2 D flow MRI, Lotz et al. described that misaligned measuring planes
lead to inaccurate findings [2]. Later, in a phantom study it was shown that 4 D flow MRI sequences deliver accurate
measurements even with misaligned acquisition planes [8]. Since 4 D flow MRI enables measurements of flow that does not move perpendicular
to the measuring plane, we introduced the parameter absolute flow velocity (VABS
), which includes both flow that moves perpendicular to the measuring plane and oblique
flow in all spatial directions. Our results indicate slightly higher values for VABS
compared to VAx
, indicating that flow that moves perpendicular to the measuring plane and oblique
flow add up. The authors hypothesize that both kinds of flow exist in vivo, flow that
moves perpendicular to the measuring plane and oblique flow and since VABS
includes any flow, it should depict in vivo flow more precisely compared to VAx
.
One limitation of this study is that it only provides 2 D flow MRI data for planes
A and D and not for all measuring planes, but since we found no significant differences
on these two planes, it is highly unlikely that the other planes would show any differences.
Additionally, head-to-head comparisons between the used kt-GRAPPA 4 D flow MRI sequence
and 2 D flow MRI has been published already in a phantom study and in vivo as well,
without significant differences between both techniques [8]
[9] in pulsatile and non-pulsatile flow. Another limitation is the rather long acquisition
time of the used 4 D flow MRI sequence, while there are other sequences with more
advanced acceleration techniques that allow for scanning of the whole aorta within
< 5 minutes [36]. This is because the aim of this study was to generate normal values using a commercially
available and widely accessible 4 D flow MRI sequence.
Conclusion
In conclusion, this current study provides sex-dependent quantitative 4 D flow MRI-derived
thoracic aortic normal ranges regarding flow volumes and flow velocities of aortic
blood flow in healthy individuals. The acquired normal values of forward, backward,
and net flow volumes as well as axial velocities did not differ significantly from
normal values of single 2 D flow MRI acquisitions and can therefore be used to replace
multiple single 2 D flow MRI acquisitions, which can lead to a broader use of 4 D
flow MRI in the clinical routine. The parameter VABS
should not be used interchangeably. This study shows that 4 D flow MRI can be integrated
in the clinical routine and due to its “straight forward planning”, it may replace
conventional 2 D flow MRI sequences in the future.
Funding
Deutsche Forschungsgemeinschaft (GR 4617/2–1 AOBJ 629 069)
