If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
CHIMERE UR 7516, Jules Verne University of Picardy, Amiens, FranceMedical Image Processing Department, Amiens Picardy University Hospital, Amiens, France
CHIMERE UR 7516, Jules Verne University of Picardy, Amiens, FranceMedical Image Processing Department, Amiens Picardy University Hospital, Amiens, FranceMRI Department, Jules Verne University of Picardy, Amiens, France
Compared to conventional phase-contrast MRI (Conv-PC), echo-planar phase-contrast MRI (EPI-PC) can provide flow rate in real-time.
•
EPI-PC can accurately quantify flow rate and flow waveforms.
•
Compared to Conv-PC, EPI-PC can adapt to lower spatial resolution.
•
Excessive velocity encoding has a greater impact on the accuracy of EPI-PC compared to Conv-PC.
Abstract
Purposes
To compare the accuracy of real-time phase-contrast echo-planar MRI (EPI-PC) and conventional cine phase-contrast MRI (Conv-PC) and to assess the influence of spatial resolutions (pixel size) and velocity encoding on flow measurements obtained with the two sequences.
Methods
Flow quantification was assessed using a pulsatile flow phantom (diameter: 9.5 mm; mean flow rate: 1150 mm3/s; mean flow velocity: 1.6 cm/s). Firstly, the accuracy of the EPI-PC was checked by comparing it with the flow rate in the calibrated phantom and the pulsation index from Conv-PC. Secondly, flow data from the two sequences were compared quantitatively as a function of the pixel size and the velocity encoding.
Results
The mean percentage difference between the EPI-PC flow rate and calibrated phantom flow rate was −2.9 ± 2.1% (Mean ± SD). The pulsatility indices for EPI-PC and Conv-PC were respectively 0.64 and 0.59. In order to keep the flow rate measurement error within 10%, the ROI in Conv-PC had to contain at least 13 pixels, while the ROI in EPI-PC had to contain at least 9 pixels. Furthermore, Conv-PC had a higher velocity-to-noise ratio and could use a higher velocity encoding than EPI-PC (20 cm/s and 15 cm/s, respectively).
Conclusions
The result of this in vitro study confirmed the accuracy of EPI-PC, and found that EPI-PC can adapt to lower spatial resolutions, but is more sensitive to velocity encoding than Conv-PC.
Conventional phase-contrast magnetic resonance imaging (Conv-PC) is a non-invasive technique that can be used to measure blood and cerebrospinal fluid (CSF) velocities. Conv-PC was used by Moran in 1982 to study flow velocities in humans.
Unfortunately, Conv-PC is limited by its relatively poor time resolution; it can only provide flow measurements for an averaged heartbeat cycle, which is reconstructed from all the acquired heartbeat cycles and uses gating.
Consequently, the flow velocities measured with Conv-PC may be breathing-dependent. Furthermore, Conv-PC is difficult to reveal the effects of breathing on the dynamics of blood or CSF flows.
To overcome this limitation, several research groups have developed a fast acquisition method based on echo-planar imaging (EPI) in which a complete k-space can be acquired using one or a small number of pulse excitations.
Eichenberger et al. in 1995 combined EPI with phase-contrast technique and thus introduced a novel sequence now commonly referred to as EPI-PC, using which they successfully quantified the blood flow of thoracic vessels at 20 mm diameter level in real-time with a spatial resolution of 5 × 5 mm2.
With improved hardware performance, higher spatial resolution and smaller velocity encoding (VENC) can recently be used in EPI-PC to quantify cerebrovascular blood flow with smaller cross-sectional areas and cerebrospinal fluid oscillations with slower flow rates in real-time. With a high time resolution (100 ms/image, or shorter), a freely determined acquisition time, and a simpler acquisition process (i.e. no need for synchronization), EPI-PC has clear advantages in the field of research but also opens new opportunities for clinical practice.
However, EPI-PC is more sensitive to eddy currents and has a longer readout window, resulting in a lower velocity-to-noise ratio (VNR) of the phase image than for Conv-PC.
Using a larger VENC can avoid the aliasing effects, but it will reduce the VNR and thus increase the segmentation difficulty. A better understanding of the effects of spatial resolution and VNR on EPI-PC allows us to further ensure quantification accuracy and improve imaging quality. However, to the best of our knowledge, there is a lack of specific literature on this field.
The objectives of this in vitro study were to quantitatively evaluate the accuracy of EPI-PC vs. Conv-PC and to assess the influence of pixel size and VENC on flow rate measurements.
2. Material and methods
2.1 Flow phantom
The phantom consisted of a series of four rigid, straight tubes (Tygon tubing, Saint-Gobain Performance Plastics, Akron, OH) with inner diameters of 9.5 mm (tube #1), 6.4 mm (tube #2), 4.4 mm (tube #3) and 2 mm (tube #4). We used water to simulate cerebrospinal fluid which has longer T1 and T2 relaxation times compared to blood.
The water flow was generated using a pulsatile flow pump. Six meters of tubing carried the water from the pump (located in the scanner control room) to the phantom's inlet. The phantom's outlet was connected to the tank that supplied water to the pump (Fig. 1a).
Fig. 1The flow phantom and the flow curve for the two sequences. a) A realistic pulse-based model of the craniospinal system (left). An amplitude image for four tubes and a static tube in the acquisition plane (right), b) the phase contrast images of Conv-PC are post-processed to obtain an average flow curve. c) The phase contrast images of EPI-PC are post-processed to obtain a continuous flow signal. The software automatically locates the minimum value of each pulse cycle (red points) in this continuos flow signal, which is used to segment the continuous signal into multiple independent pulse cycles. All these pulse cycles are used to reconstruct the average pulse cycle (reconstructed flow curve). d) Comparison of the flow curve of Conv-PC and the reconstructed pulse cycle of EPI-PC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
To validate our system's flow rates, the pump was calibrated to deliver a clinically relevant flow rate (pulsatile flow with 99 bpm), and the volume collected from the phantom's output was recorded as a function of time. The true (calibrated) phantom mean flow rate (1150 mm3/s) was obtained through repeated measurements and was used as the reference mean flow rate value for the present work. Both Conv-PC and EPI-PC were used to measure the flow rate of the first tube with 9.5 mm of diameter, its area of 70.8 mm2 was taken as reference for segmentation area. The flow waveform of Conv-PC was used as reference, the ratio of the amplitude flow rate (maximum flow rate minus minimum flow rate) and the mean flow rate was used as the reference pulsatility index.
The flow phantom was positioned in the center of a head coil. On the return tube, a gating-compliant balloon was used to capture the frequency of the oscillation and thus to synchronize the acquisition of the Conv-PC with the flow rate waveform. A water-filled tube (tube reference) was positioned beside the tubes, to define the static reference region.
2.2 Imaging procedure
All images were acquired on a 3T clinical scanner (Philips Achieva; maximum gradient: 80 mT/m; rate of gradient increase: 120 mT m−1 ms−1) using a 32-channel head coil.
The parameters of the imaging protocol for both sequences are shown in Table 1. The acquisition was repeated 10 times. The Conv-PC is a gradient echo based, blipped phase contrast sequence with a cartesian trajectory. During the acquisition of Conv-PC images, a finger pulse oximeter was posed on the gating balloon for retrospective cardiac gating. Typically, the velocity is encoded along the flow direction (A-P, in this study) by positioning a bipolar gradient (with opposite polarity) behind the slice selection gradient. The spins of flowing tissue are at different locations relative to the bipolar gradient's positive and negative lobes. These spins are then confronted with the magnetic field gradients and accumulate a residual phase difference, whereas the spins of stationary tissue do not produce phase differences. Phase data sets from before and after gradient reversal are subtracted to determine the “phase difference” of flowing spins, which is directly proportional to their underlying velocities. A description of the relationship between the measured phase and the velocities can be found in
Readout-segmented echo-planar imaging in diffusion-weighted mr imaging in breast cancer: comparison with single-shot echo-planar imaging in image quality.
EPI-factor is a parameter specific of EPI-PC that indicates the number of echoes acquired during a TR.
Each Conv-PC series contained 32 phase images after 23.6 s of acquisition, the interval between two images () during an average cycle was constant (∆t = 19 milliseconds). For EPI-PC series, the total number of acquired phase images was set to 150. The acquisition time was 9.3 s, and the ∆t was 62 milliseconds. The reconstructed images of the two sequences were obtained from the initial images by using interpolation algorithm (Table 1).
To assess the effects of pixel size and VENC on flow rate measurements for both sequences, the pixel size was set from 0.8 mm to 3.2 mm in increments of 0.4 mm, and VENC was set from 5 cm/s to 25 cm/s in increments of 5 cm/s. The acquisition was repeated four times for both sequences and for each parameter. Once the pixel size reached 2 × 2 mm2, the pixel size was increased by fixing the acquisition matrix and increasing the FOV.
2.3 Postprocessing procedure
To extract the flow rate curves, the MRI data were processed with in-house software (Flow
). To minimize the effects of eddy current on the measurements, the velocity was calibrated by measuring the mean velocity in the tube reference (Fig. 2.a, green circle). Furthermore, to compare the two sequences, the EPI-PC flow rate signal was reconstructed over an average pulse cycle of 32 points with the same model as Conv-PC (Fig. 1.b). The post-processing procedure is as follows:
Fig. 2An example of Conv-PC postprocessing. a A representative segmented image obtained with software, with ROI-Tube#1 on tube #1 (in red) and ROI-Reference on the static tube (in green), b the uneven signal obtained from the SD of the velocity within the ROI-Reference in each phase image, c the reference signal constituted by the mean velocity within the ROI-Reference in each phase image, d the original flow curve (in red) for a pulse cycle extracted from the ROI-Tube, and the corrected flow curve (in green) calculated from the original flow and reference signals. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
By using the software's segmentation function, a region of interest (ROI) within the tube #1 (ROI-Tube) can be automatically segmented on the phase image. The value of the segmented area can then be recorded. Likewise, a ROI within the static tube (ROI-Reference) was manually defined as the source of velocity noise (Fig. 2.a, green cycle). For each phase image, the mean (VRef) and standard deviation (SD) (σRef) velocity within ROI-Reference were calculated. The VRef and σRef values were used to define a reference signal and an uneven signal, respectively (Fig. 2.c & b).
2.3.2 Calibration of the measured velocities
The calibration compensated for the noise error in the measurement of velocity. In theory, the measured velocity does not represent the true velocity, and the velocity in the ROI-Reference is null. To calculate the corrected (true) velocity, the measured velocity was subtracted from the VRef (Fig. 2.d).
2.3.3 Calculation of the VNR
The mean uneven signal was used to calculate the VNR by dividing the mean velocity in the ROI-Tube by the mean uneven signal in ROI-Ref (Eq. (1)).
(1)
2.3.4 Reconstruction of the mean EPI-PC cycle
The steps in the segmentation and calibration of the EPI-PC data were the same as for the Conv-PC data (2.3.1, 2.3.2, 2.3.3). After the flow rate signal has been obtained from the EPI-PC data, the software's cropping tool can be used to extract all the single pulse cycles from the original signal (red points in Fig. 1.b). A spline interpolation algorithm was then used to increase the number of sampling points to 32 for each pulse cycle in the EPI-PC data. The sampling points at each position and each pulse cycles were then averaged to obtain the corresponding flow rate value for the reconstructed average pulse cycle. This average flow curve will be used for subsequent accuracy evaluations and analyses of the effects of pixel size and VENC.
2.3.5 The pulsatility index
Calculate the pulsatility index of flow curve of Conv-PC and the pulsatility index of average flow curve of EPI-PC using Eq. (2).
(2)
2.4 Accuracy assessment of EPI-PC
The accuracy of EPI-PC was evaluated by comparing the quantified results of EPI-PC with the reference flow rate and reference area of calibrated phantom, and with the pulsatility index of the flow curve of Conv-PC.
2.5 The influence of pixel size and VENC
The measured average flow, segmentation area and VNR of the two sequences at different pixel sizes and velocity encodings were compared to analyze the effect of these two parameters on Conv-PC and EPI-PC. Based on literature procedures for measuring the accuracy of phase contrast sequences,
we considered that the acceptable confidence intervals (CIs) for the segmentation error and the flow rate error were ± 10%.
2.6 Statistical analysis
The influence of pixel size and VENC on flow rate was evaluated using a regression analysis. Pearson's test was used to analyze the correlation between variables. Values are expressed as the mean ± standard deviation (SD). All statistical analyses were performed with R software (version 3.2.3, R Foundation for Statistical Computing, Vienna, Austria, www.r-project.org). The threshold for significance was set to p < 0.05.
3. Results
After several manual measurements, it was verified that the average flow rate of the phantom was 1150 mm3/s. The Reynolds number of the tube #1 is 1260 and so was <2100, the flow was considered to be laminar.
3.1 Comparison of EPI-PC and Conv-PC sequences
The EPI-PC and Conv-PC sequences were applied to tube #1 with the default parameters.
After 10 measurements, the distributions of the average flow rate, segmentation area and pulsatility index of EPI-PC and Conv-PC are shown in Fig. 3. The mean flow rates for EPI-PC and Conv-PC were 1116 ± 25 mm3/s and 1239 ± 26 mm3/s respectively, and the associated coefficient of variation was 2.2% in both cases (Fig. 3.b). The segmentation areas of EPI-PC and Conv-PC were 70.1 ± 1 mm2 and 70 ± 0.9 mm2 respectively, and their coefficient of variations were 1.5% and 1.3% respectively (Fig. 3.c). The pulsatility indices for EPI-PC and Conv-PC were respectively 0.64 ± 0.03 and 0.59 ± 0.01 respectively, and their coefficient of variations were 2.5% and 4% respectively (Fig. 3.d).
Fig. 3a) The average flow curves for Conv-PC (in blue) and EPI-PC (in red). The distributions of the average flow rate b), segmentation area c) and pulsatility index d) of EPI-PC and Conv-PC. The lines segment (dashed) indicated the reference flow rate b) and the reference area c) of calibrated phantom, and the lines segment (dotted) indicated the CIs for flow rate and segmentation area. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 2 shows the parameters and measurement results of Conv-PC and EPI-PC at different pixel sizes.
Table 2Parameters and quantification results of Conv-PC and EPI-PC at different pixel sizes.
Parameters
ACQ pixel size (mm2)
0.8 × 0.8
1.2 × 1.2
1.6 × 1.6
2.0 × 2.0
2.4 × 2.4
2.8 × 2.8
3.2 × 3.2
REC Pixel (mm2)
Conv-PC
0.31 × 0.31
0.31 × 0.31
0.39 × 0.39
0.52 × 0.52
0.58 × 0.58
0.75 × 0.75
0.86 × 0.86
EPI-PC
0.78 × 0.78
0.78 × 0.78
0.78 × 0.78
0.78 × 0.78
0.78 × 0.78
0.75 × 0.75
0.98 × 0.98
Average Flow rate (mm3/s)
Conv-PC
1007 ± 7
1035 ± 17
1101 ± 10
1128 ± 6
1189 ± 18
1278 ± 5
1446 ± 46
EPI-PC
971 ± 24
1064 ± 13
1117 ± 23
1140 ± 8
1143 ± 10
1177 ± 8
1394 ± 40
Area ROI (mm2)
Conv-PC
55.7 ± 1.2
58.8 ± 2.8
64.0 ± 1.1
66.5 ± 1.1
73.4 ± 2.2
73.2 ± 0.6
74.8 ± 3.3
EPI-PC
54.3 ± 2.7
56.8 ± 1.1
61.4 ± 3.0
67 ± 2.3
72.2 ± 4.1
79.3 ± 6.2
76.2 ± 4.6
Maximum flow rate (mm3/s)
Conv-PC
1333
1422
1467
1541
1622
1705
1854
EPI-PC
1200
1358
1467
1558
1614
1670
1865
Maximum flow velocity (mm/s)
Conv-PC
36
39
39
39
38
35
33
EPI-PC
37
37
38
40
40
33
32
ACQ time (second)
Conv-PC
33
22.6
17.1
20.2
17.1
15.9
15.3
EPI-PC
16.9
9.3
8.2
7.7
7.4
7.1
6.8
VNR
Conv-PC
9
16
13.9
15
14.3
15.9
18.8
EPI-PC
1.9
3.3
4.6
6.7
9.5
13.6
20.6
Flow with fixed ROI (mm3/s)
Conv-PC
1044
1098
1116
1140
1172
1252
1380
EPI-PC
1104
1114
1180
1157
1100
1125
1363
ACQ = acquisition; REC Pixel = reconstruction pixel size; Area ROI = Segmented area of Tube #1; VNR = velocity-to-noise ratio; Flow with fixed ROI = using the amplitude image for segmentation to lock the ROI at the standard value (70.8 mm2). The flow rate data in bold indicates values outside the cofidence interval.
In Fig. 4, the segmentation area (blue points) and the flow rate (red points) are shown as a function of pixel size. Each variable was measured four times. The blue shading corresponds to the CI for the reference segmentation area (70.8 mm2 ± 10%) and the red shading corresponds to the CI for the reference flow rate (1150 mm3/s ± 10%). The flow rates measurements obtained with EPI-PC were within the CI for pixel sizes of 1.2 mm to 2.8 mm. For Conv-PC, only pixel sizes of 1.2 mm to 2.4 mm provided flow rates within the CI.
Fig. 4Segmentation area (on the right y-axis, in blue) and flow rates (on the left y-axis, in red) for two sequences with different pixel sizes (on the X-axis). a EPI-PC, b Conv-PC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The distributions of the segmentation area (on the x axis) and flow rate (on the y axis) of the two sequences are shown in Fig. 5. The purple and green shadings correspond to the CI of the segmentation area and the flow rate, respectively.
Fig. 5Distribution of the segmentation area & flow rate for EPI-PC and Conv-PC, as a function of the pixel size (depicted by different color levels). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 6 shows the VNR of EPI-PC and Conv-PC with different pixel sizes. For a pixel size of 1.2 mm to 2.8 mm, the VNR increased for EPI-PC but did not change for Conv-PC. For each pixel size, the VNR was higher for Conv-PC than for EPI-PC.
Fig. 6The VNR for EPI-PC and Conv-PC sequences, as a function of the pixel size.
Table 3 shows the parameters and measurement results of Conv-PC and EPI-PC at different VENC.
Table 3Parameters and quantification results of Conv-PC and EPI-PC at different VENC.
Parameters
VENC (cm/s)
5
10
15
20
25
Velocity reference (cm/s)
Conv-PC
−3.2
0.37
3.2
1.4
2.2
EPI-PC
−1.2
−1.4
−1.3
0.73
0.73
SD reference (cm/s)
Conv-PC
0.54
0.92
2
1.88
2.26
EPI-PC
4.9
9.4
13.8
17.7
21.7
Average flow rate (mm3/s)
Conv-PC
1230 ± 10
1112 ± 3
1118 ± 6
1072 ± 14
996 ± 24
EPI-PC
1145 ± 12
1212 ± 18
1139 ± 34
803 ± 4
Area ROI (mm2)
Conv-PC
62.9 ± 1.9
69.7 ± 1.8
69 ± 0.9
61.6 ± 5.6
56.0 ± 4.4
EPI-PC
65.3 ± 2.2
62.6 ± 1.5
54.5 ± 3.8
46.4 ± 0.6
Maximum flow rate (mm3/s)
Conv-PC
1674
1607
1535
1504
1416
EPI-PC
1535
1605
1461
1251
Maximum flow velocity (mm/s)
Conv-PC
43
45
40
38
38
EPI-PC
38
38
42
42
ACQ time (second)
Conv-PC
23.6
22.4
22.4
22.4
22.4
EPI-PC
9.3
8.5
8.2
8.0
7.9
VNR
Conv-PC
27.2
17.4
8.1
9.3
7.9
EPI-PC
3.6
2.1
1.5
1.0
0.8
Flow with fixed ROI (mm3/s)
Conv-PC
1214
1115
1040
1070
1034
EPI-PC
1207
1231
1174
1028
983
Velocity reference = average velocity within the ROI of reference tube; SD reference = standard deviation of velocity within the ROI of reference tube; Area ROI = segmented area of Tube #1; ACQ = acquisition; VNR = velocity-to-noise ratio; Flow with fixed ROI = using the amplitude image for segmentation to lock the ROI at the standard value (70.8 mm2).
Fig. 7 shows the influence of VENC on EPI-PC and on Conv-PC. As the VENC increased, the accuracy decreased more rapidly for EPI-PC than for Conv-PC. With a VENC of 15 cm/s, the segmentation area for EPI-PC was outside the CI. Moreover, with a VENC of 20 cm/s, the flow rate for EPI-PC was also outside of the CI. For a VENC of 25 cm/s or more, the software was not able to segment the ROI. In contrast, with a VENC below 20 cm/s, the flow rates for Conv-PC were within the CI.
Fig. 7Segmentation area (on the right y-axis, in blue) and flow rates (on the left y-axis, in red) for two sequences with five different VENC values (on the X-axis). a EPI-PC, b Conv-PC. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 8 gives the VNR for EPI-PC and Conv-PC with different VENC values. As the value of VENC increased, the VNR for EPI-PC fell from 3.6 to 0.8. For Conv-PC, the VNR fell from 27.2 to 8.1 as the VENC value increase from 5 cm/s to 15 cm/s. For VENC values of 15 cm/s to 25 cm/s, the VNR varied less (8.4 ± 0.76). Overall, the VNR with different VENC values was much greater for Conv-PC than for EPI-PC.
Fig. 8The VNR for EPI-PC and Conv-PC sequences, as a function of the VENC.
Therefore, there may be errors in the background field correction, and it will be increased with the increase of VENC (Figs. S5, S6, Table S1). Moreover, since EPI-PC is more susceptible to eddy currents, in clinical applications, stationary tissue close to the target vessel should be selected as much as possible for background field correction.
The estimated VNR in this study was a pseudo-VNR averaged across all images. This measurement method is likely to be more conveniently and might be sufficient for comparing the VNRs from several different sequences.
These results are not directly transferred to single-shot EPI-PC as this sequence is more sensitive to geometric distortion and has a lower VNR,
and does not apply on non-laminar or asymmetrical flow profiles. The multi-shot EPI-PC is less sensitive to geometric distortions, given the shorter readout time.
This is why we used a multi-shot EPI-PC in this study.
4.1 Comparison of EPI-PC and Conv-PC
Using default parameters, the respective flow rate and segmentation area measurements for the two sequences were both within the CIs and did not exceed 8%
Comparison of fast acquisition strategies in whole-heart four-dimensional flow cardiac MR: two-centre, 1.5 Tesla, phantom and in vivo validation study.
(flow rate error: −2.9% for EPI-PC vs. 7.8% for Conv-PC; segmentation area error: −1% for EPI-PC vs. -1.1% for Conv-PC). Difference in the pulsatility index between EPI-PC and Conv-PC was <10%.
There are many other factors that influence the flow quantification. Even with the same imaging protocol, differences in the placement of the water pipes in the magnetic field can affect the results. Therefore, it is acceptable with a quantification error of no more than 10%.
The Conv-PC sequence was able to complete several phase encodings during each pulse cycle and to fill them into the different phase images' K-space; even though the acquisition time increased, it did not therefore affect the pseudo-sampling interval ∆t. However, for the real-time imaging with EPI-PC, the acquisition time was directly related to ∆t. The EPI-PC continue flow rate signal with default parameters, only 9 or 10 characteristic points were used to describe a pulse cycle. Therefore, EPI-PC is more suitable for flows with gentle fluctuations, such as venous blood and CSF. However, EPI-PC has limitations for reconstructing high-frequency fluctuations, such as certain arterial waveforms. Increasing the EPI-PC sampling frequency is likely to improve the accuracy of the corresponding reconstructed curve (Fig. S1).
4.2 The influence of pixel size
The EPI-PC sequence was less sensitive to pixel size than Conv-PC. Due to the characteristics of laminar flow, the velocity is lower at the boundary of the tube than in its center. When the resolution is high, the flow at the tube wall did not produce a large phase difference. Hence, this area can be considered to be non-flowing on the phase contrast image, and so the true segmented area is smaller than the theoretical area (Fig. 5).
Accurate segmentation area and flow rate measurements were possible with a range of pixel sizes: from 1.8 to 2.8 mm for EPI-PC and from 1.8 to 2.4 mm for Conv-PC. Within these ranges, the measured flow rate was slightly influenced by the segmentation area. As the pixel size continued to increase (above 2.8 mm for EPI-PC and above 2.4 mm for Conv-PC), the flow rate of the two sequences began to exceed the boundaries of the CI (Fig. 5). We hypothesize that a partial volume effect led to overestimation of the velocity.
As a result, the flow rate error was too large - even though the segmentation region was within the CI.
A larger acquisition pixel size can effectively improve the VNR of EPI-PC and reduce the difficulty of image post-processing (Fig. 6); it can also improve the temporal resolution and increase the accuracy of flow waveform quantification (Fig. S1). Moreover, to a certain extent, it avoids the aliasing effect since larger acquisition pixels can average the high velocity area with the surrounding area and thus reduce the maximum flow velocity (Fig. S7).
Therefore, in order to measure the flow rate accurately, the pixel size for Conv-PC should be <25% of the target ROI diameter; in other words, the ROI of Conv-PC should comprise at least 13 pixels. Our result is in line with those obtained by Greil et al. and Tang et al.,
who confirmed that a minimum of 16 pixels within the Conv-PC ROI was required to keep the flow rate error within 10%. For the EPI-PC sequence, the pixel size should be <30% of the diameter of the target ROI diameter; in other words, the ROI of EPI-PC should have at least 9 pixels. On the other hand, while ensuring accuracy, larger pixel sizes can increase the VNR and improve imaging speed (Fig. S1), which is essential for EPI-PC.
4.3 The influence of VENC
The EPI-PC was more sensitive to VENC. Without aliasing, the VNR is inversely proportional to the VENC.
This sensitivity was also reflected in our experiments by the pseudo-VNR (Figs. 8 & S2). Compared to EPI-PC, Conv-PC can provide accurate flow rate with a larger VENC. Firstly, this was because the Conv-PC can use multiple cycles to fill a phase image, and so the influence of noise on the echo signal is smaller. For EPI-PC, multiple phase encodings are needed to complete the K-space during a TR, so the SNR of the echo signal is much smaller and the VNR of the phase image is relatively low. Secondly, in order to increase the sampling frequency in the EPI-PC sequence, the SENSE
value in this study as set to 2.5 which is 66% greater than that of Conv-PC (1.5), and a larger SENSE value decreases the VNR of phase image.
Since the VNR of Conc-PC was higher than that of EPI-PC, the increase in VENC had relatively little influence on the accuracy of Conv-PC flow rate measurements. Even when the VENC increased from 5 cm/s to 20 cm/s, the flow rate measured with Conv-PC was still within the CI. In contrast, the maximum value for VENC in the EPI-PC sequence was 10 cm/s; above this value, the segmentation was inaccurate.
The decrease in VNR mainly affects the segmentation of phase images; segmentation errors can arise when the pixel intensity in the phase image of the target vessel is close to that of the surrounding (non-flowing) tissue. This effect can be reduced if the magnitude image is used for segmentation, because the effect of VENC on the SNR of the magnitude image is much smaller than the effect on the VNR on the phase image.
On the other hand, a smaller VENC can increase the VNR, but at the same time slightly increasing the TR, leading to an increase in the Δt of EPI-PC (Fig. S3). Therefore, it is also feasible to improve the imaging speed by increasing the VENC in clinical applications while ensuring accurate quantification.
4.4 Limitations and perspectives
There are also some potential limitations in our study that should be noted.
Although the acquisition pixel size is the main influence on the measurement accuracy, the difference in reconstructed pixel size still affects the fairness of comparing two sequences and should be considered in subsequent studies.
The phantom in our study presents a sinusoidal waveform with only one frequency component, which is used to simulate the CSF or cerebral vein waveform, whereas there are two higher-frequency harmonics (2–4 Hz) in the cerebral arterial waveform.
The results of this study are only applicable to demonstrate the accuracy of EPI-PC in quantifying CSF and cerebral veins. Hence, the ability of EPI-PC to accurately quantify more pulsatile arterial waveforms (e.g., aorta) with a temporal resolution of 62 ms/image needs further evaluation.
The phantom model did not take account of arrhythmic and respiratory effects, for which the EPI-PC sequence is advantageous because (in contrast to Conv-PC) it does not require synchronization with the cardiac cycle. Including these influences allows a more comprehensive comparison of the characteristics of the two sequences.
5. Conclusion
Our study shows that the calculated error between the reconstructed EPI-PC flow curve and the Conv-PC flow curve was small. Compared with Conv-PC, EPI-PC can be adapted to a lower spatial resolution but is more sensitive to VENC.
In this study, setting the pixel size of the EPI-PC to 30% of the diameter of the water tube gives the best VNR and temporal resolution without causing partial volume effects; and to obtain a higher VNR, VENC should be set small (aliasing needs to be avoided). With the imaging speed of 62 ms/image can accurately represent the venous or cerebrospinal flow rate waveform at 99 bpm.
EPI-PC is more susceptible to eddy currents, so when performing background field correction for clinical applications, the stationary tissue should be selected as close to the target vessel as possible, and a small VENC should be used.
These results provided reference values for the clinical application of EPI-PC.
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.
Funding
This research was funded by the French National Research Agency (reference: Hanuman ANR-18-CE45-0014 and EquipEX FIGURES 10-EQPX-0001) and INTERREG France (Channel) England Programme (REVERT project).
CRediT authorship contribution statement
LP designed the study, developed the post-processing software, and wrote the initial draft of the manuscript. SF performed statistical analysis of the data and revised the manuscript. MA collected and analyzed the in vitro data. Ob developed the post-processing software and revised the manuscript. All authors approved the final version of the manuscript for submission. All authors read and approved the final manuscript.
Declaration of competing interest
The authors declare that they have no conflicts of interest with respect to the research, authorship, and/or publication of this article. The figures are original and have not been published previously.
Acknowledgements
The authors thank David Chechin for scientific support and the staff members at the Facing Faces Institute (Amiens, France) for technical assistance.
Readout-segmented echo-planar imaging in diffusion-weighted mr imaging in breast cancer: comparison with single-shot echo-planar imaging in image quality.
Comparison of fast acquisition strategies in whole-heart four-dimensional flow cardiac MR: two-centre, 1.5 Tesla, phantom and in vivo validation study.
Dr Pan Liu received his PhD in Radiological and Medical Physics from the Jules Verne University of Picardy (Amiens, France) in 2021. He is currently funded as a postdoctoral fellow through the CHIMERE-UR7516 research group. His main research interests are MRI image processing algorithms and image processing software development.
Dr Sidy Fall received a PhD in Biomedical Engineering from the Jules Verne University of Picardy (Amiens, France) in 2009. He is currently an engineer at the university's Preclinical MRI Facility. His main research interests are MRI and physiological data processing and especially functional MRI, DTI, and vascular imaging.
Maureen Ahiatsi received a Master's Degree in Neurosciences – Neurobiology from the University of Grenoble Alpes (Grenoble, France) in 2020. She is currently studying for a Master's Degree in Artificial Intelligence and Digital Transformation at the AI & Digital Transformation School (Paris, France). Her research interests include brain pathophysiology, and medical imaging data processing and analysis.
Dr Olivier Baledent received a PhD on the MRI-based imaging of CSF and blood flow using at the Jules Verne University of Picardy (Amiens, France) in 2001. He is currently an assistant professor at the university and also heads the Medical Image Processing Department at Amiens University Medical Center (Amiens, France), where he is developing flow MRI techniques in the CHIMERE-UR7516 research group. In 2020, he helped to create ‘neurofluid imaging’ as a new section in the International Society of Magnetic Resonance Imaging.
00306-0&id=ga1.jpg){kind=link}
00306-0&id=gr1.jpg){kind=link}
00306-0&id=gr2.jpg){kind=link}
00306-0&id=gr3.jpg){kind=link}
00306-0&id=gr4.jpg){kind=link}
00306-0&id=gr5.jpg){kind=link}
00306-0&id=gr6.jpg){kind=link}
00306-0&id=gr7.jpg){kind=link}
00306-0&id=gr8.jpg){kind=link}