Magnetic Resonance Phase Velocity Mapping Research Articles

A nonlocal maximum likelihood estimation method for enhancing magnetic resonance phase maps

A phase map can be obtained from the real and imaginary components of a complex valued magnetic resonance (MR) image. Many applications, such as MR phase velocity mapping and susceptibility mapping, make use of the information contained in the MR phase maps. Unfortunately, noise in the complex MR signal affects the measurement of parameters related to phase (e.g, the phase velocity). In this paper, we propose a nonlocal maximum likelihood (NLML) estimation method for enhancing phase maps. The proposed method estimates the true underlying phase map from a noisy MR phase map. Experiments on both simulated and real data sets indicate that the proposed NLML method has a better performance in terms of qualitative and quantitative evaluations when compared to state-of-the-art methods.

Reduced Intrathoracic Blood Volume and Left and Right Ventricular Dimensions in Patients With Severe Emphysema: An MRI Study

Left ventricular (LV) filling is impaired in patients with severe emphysema manifesting in small end-diastolic dimensions. We hypothesized that the hyperinflated lungs of these patients with high intrinsic positive end-expiratory pressure will decrease intrathoracic blood volume (ITBV) and ventricular preload. We therefore measured ITBV, and LV and right ventricular (RV) dimensions and function using MRI techniques in patients with severe emphysema.Patients with severe emphysema (n = 13) and matched healthy volunteers (n = 11) were included. The magnetic resonance (MR) examination consisted of three parts: (1) evaluation of RV and LV dimensions and function and interventricular septum curvature using cine MRI; (2) quantification of aortic flow using MR phase velocity mapping; and (3) calculation of the cardiopulmonary peak transit time (PTT) from the pulmonary artery to the ascending aorta using contrast-enhanced, time-resolved, two-dimensional MR angiography.There were no differences between the groups regarding age, height, or weight. In the emphysema patients, ITBV index (- 35%), LV end-diastolic volume index (LVEDVI) [- 21%], RV end-diastolic volume index (- 20%), cardiac index (- 22%), and stroke volume index (SVI) [- 40%] were lower compared to control subjects. LV and RV end-systolic volumes, LV wall mass, septal curvature, and PTT did not differ between the groups. LVEDVI (r = 0.83) as well as SVI (r = 0.82) correlated closely to ITBV index. SVI correlated closely to LVEDVI (r = 0.84).LV and RV performance is impaired in patients with severe emphysema because of small end-diastolic dimensions. One possible explanation for the decreased biventricular preload in these patients is intrathoracic hypovolemia caused by hyperinflated lungs.

Noninvasive quantification of fluid mechanical energy losses in the total cavopulmonary connection with magnetic resonance phase velocity mapping

A major determinant of the success of surgical vascular modifications, such as the total cavopulmonary connection (TCPC), is the energetic efficiency that is assessed by calculating the mechanical energy loss of blood flow through the new connection. Currently, however, to determine the energy loss, invasive pressure measurements are necessary. Therefore, this study evaluated the feasibility of the viscous dissipation (VD) method, which has the potential to provide the energy loss without the need for invasive pressure measurements. Two experimental phantoms, a U-shaped tube and a glass TCPC, were scanned in a magnetic resonance (MR) imaging scanner and the images were used to construct computational models of both geometries. MR phase velocity mapping (PVM) acquisitions of all three spatial components of the fluid velocity were made in both phantoms and the VD was calculated. VD results from MR PVM experiments were compared with VD results from computational fluid dynamics (CFD) simulations on the image-based computational models. The results showed an overall agreement between MR PVM and CFD. There was a similar ascending tendency in the VD values as the image spatial resolution increased. The most accurate computations of the energy loss were achieved for a CFD grid density that was too high for MR to achieve under current MR system capabilities (in-plane pixel size of less than 0.4 mm). Nevertheless, the agreement between the MR PVM and the CFD VD results under the same resolution settings suggests that the VD method implemented with a clinical imaging modality such as MR has good potential to quantify the energy loss in vascular geometries such as the TCPC.

Mean-Average Wall Shear Stress Measurements in the Common Carotid Artery

In this study we determined mean-average wall shear stress values in the common carotid artery and assessed if there is a difference in mean-average WSS between: 1) patients with bilateral carotid bifurcation disease and 2) similar-aged volunteers with no evidence of disease. Sixteen patients with bilateral disease of the carotid bifurcation, and 8 volunteers were included in the study. Magnetic resonance phase velocity mapping was used to determine velocity, flow, vessel cross-sectional dimensions, and mean-average WSS in the common carotid artery in both the patients and volunteers. Mean-average WSS in the common carotid artery was 7.5 +/- 2.5 dynes/cm2 in patients and 8.0 +/- 4.1 dynes/cm2 in volunteers. There was no significant difference in mean-average WSS, average velocity, peak velocity, flowrate, or vessel diameter in the common carotid artery between patients and volunteers.

Comparison of myocardial velocities obtained with magnetic resonance phase velocity mapping and tissue doppler imaging in normal subjects and patients with left ventricular dyssynchrony

To compare longitudinal myocardial velocity and time to peak longitudinal velocity obtained with magnetic resonance phase velocity mapping (MR-PVM) and tissue Doppler imaging (TDI), and to assess the reproducibility of each method. Longitudinal myocardial velocity was measured by TDI and MR-PVM in 10 normal volunteers and 10 patients with dyssynchrony. The reproducibility of MR-PVM and TDI was assessed on repeated measurements in the 10 normal volunteers. MR and TDI measurements of longitudinal myocardial velocity correlated well (r = 0.86) in both normal subjects and patients with dyssynchrony. However, myocardial velocities measured with MR consistently exceeded velocities measured with TDI. MR and TDI agreed strongly in measuring the time to peak velocity (r = 0.97). The reproducibility of TDI and MR-PVM appeared similar in measuring peak velocities (13.1% vs. 11.0%, respectively; P = NS) and time to peak velocity (9.1% vs. 5.7%, respectively; P = NS). Excellent correlation and reproducibility were observed between MR-PVM and TDI in measuring longitudinal myocardial velocity and time to peak velocity in both normal subjects and patients with dyssynchrony.

Evaluation of in‐stent stenosis by magnetic resonance phase‐velocity mapping in nickel‐titanium stents

To evaluate different grades of in-stent stenosis in a nickel-titanium stent with MRI. Magnetic resonance phase velocity mapping (MR-PVM) was used to measure flow velocity through a 9-mm NiTi stent with three different degrees of stenosis in a phantom study. The tested stenotic geometries were 1) axisymmetric 75%, 2) axisymmetric 90%, and 3) asymmetric 50%. The MR-PVM data were subsequently compared with the velocities from computational fluid dynamic (CFD) simulations of identical conditions. Good quantitative agreement in velocity distribution for the 50% and 75% stenoses was observed. The agreement was poor for the 90% stenosis, most likely due to turbulence and the high-velocity gradients found in the small luminal area relative to the pixel resolution in our imaging settings. The accuracy of the MRI velocities inside the stented area renders MRI a modality that may be used to assess moderate to severe in-stent restenosis (ISR) in medium-sized vascular stents in peripheral vessels, such as the iliac, carotid, and femoral arteries. Advances in MR instrumentation may provide sufficient resolution to obtain adequate velocity information from smaller vessels, such as the coronary arteries, and allow MRI to substitute for invasive and expensive catheterization procedures currently in clinical use.

Magnetic resonance phase velocity mapping: A novel technique for evaluation of dyssynchrony in candidates for CRT

Magnetic resonance phase velocity mapping: A novel technique for evaluation of dyssynchrony in candidates for CRT

Blood flow measurements with magnetic resonance phase velocity mapping

Magnetic resonance (MR) phase velocity mapping (PVM) is a non-invasive technique that can measure the flow velocity in any spatial direction in an imaging slice. This technique has wide application in the clinical field in quantifying blood flow, as well as in non-biomedical areas. This review describes the value and/or potential of MR PVM as a diagnostic/monitoring technique in heart valve regurgitation and in the total cavo-pulmonary connection. A single slice placed in the aortic root can accurately quantify the aortic regurgitant volume. A multi-slice control volume method has high potential for the quantification of the mitral regurgitant volume. In the total cavo-pulmonary connection, MR PVM with its unique clinical ability to measure all three directions of blood velocity provides the ability to visualize the two- or even three-directional blood flow patterns. It also promises a non-invasive quantification of the mechanical energy losses of blood as it flows through the connection. New rapid acquisition sequences show accuracy in quantifying flow and will greatly contribute to the increase of the number of applications of MR PVM.

Magnetic resonance phase velocity mapping through NiTi stents in a flow phantom model

To assess constant and pulsatile flow velocity within the lumen of a peripheral NiTi stent using phase velocity mapping for comparison with independent assessments of flow velocity in a phantom model. A 9 x 20-mm stent installed in flexible tubing was placed in a phantom filled with stationary fluid. Constant and pulsatile flow (produced by a pump programmed to produce a simulation of the carotid artery flow) was assessed using phase velocity mapping at 4.1 T (for constant flow) and at 1.5 T (for pulsatile flow). In all cases 256 x 256 gradient echo phase velocity maps were acquired. For the pulsatile flow condition, cine images with acquisition gated to the pump cycle were acquired with 40 msec temporal resolution across the simulated cardiac cycle. Computed flow volume rates were compared with fluid volume collection for the constant flow model, and with ultrasonic Doppler flow meter measurements for the pulsatile model. The data showed that volume flow rate assessments by phase velocity mapping agreed with independent measurements within 10% to 15%. Phase velocity mapping of the lumen of peripheral size NiTi stents is possible in an in vitro model.

Fast Measurements of Flow through Mitral Regurgitant Orifices with Magnetic Resonance Phase Velocity Mapping

Magnetic-resonance (MR) phase velocity mapping (PVM) shows promise in measuring the mitral regurgitant volume. However, in its conventional nonsegmented form, MR-PVM is slow and impractical for clinical use. The aim of this study was to evaluate the accuracy of rapid, segmented k-space MR-PVM in quantifying the mitral regurgitant flow through a control volume (CV) method. Two segmented MR-PVM schemes, one with seven (seg-7) and one with nine (seg-9) lines per segment, were evaluated in acrylic regurgitant mitral valve models under steady and pulsatile flow. A nonsegmented (nonseg) MR-PVM acquisition was also performed for reference. The segmented acquisitions were considerably faster (<10 min) than the nonsegmented (>45 min). The regurgitant flow rates and volumes measured with segmented MR-PVM agreed closely with those measured with nonsegmented MR-PVM (differences <5%, p > 0.05), when the CV was large enough to exclude the region of flow acceleration and aliasing from its boundaries. The regurgitant orifice shape (circular vs. slit-like) and the presence of aortic outflow did not significantly affect the accuracy of the results under both steady and pulsatile flow (p > 0.05). This study shows that segmented k-space MR-PVM can accurately quantify the flow through regurgitant orifices using the CV method and demonstrates great clinical potential.

Chatzimvroudis, GP et al. Clinical blood flow quantification with segmented k‐space magnetic resonance phase velocity mapping. J Magn Reson Imaging 2003; 17:65‐71.

AbstractThe original article to which this Erratum refers was published in Journal of Magnetic Resonance Imaging (2003) 17(1) 65–71.

Clinical blood flow quantification with segmented k‐space magnetic resonance phase velocity mapping

To evaluate the accuracy of segmented k-space magnetic resonance phase velocity mapping (PVM) in quantifying aortic blood flow from through-plane velocity measurements. Two segmented PVM schemes were evaluated, one with seven lines per segment (seg-7) and one with nine lines per segment (seg-9), in twenty patients with cardiovascular disease. A non-segmented (non-seg) PVM acquisition was also performed to provide the reference data. There was agreement between the aortic flow curves acquired with segmented and non-segmented PVM. The calculated systolic and total flow volume per cycle from the seg-7 and the seg-9 scans correlated and agreed with the flow volumes from the non-seg scans (differences < 5%). Sign tests showed that there were no statistically significant differences (P-values > 0.05) between the segmented and the non-segmented PVM measurements [corrected]. Seg-9, which was the fastest among the three sequences, provided adequate spatial and temporal resolution (> 10 phases per cycle). Segmented k-space PVM shows great clinical potential in blood flow quantification.

Accurate quantification of steady and pulsatile flow with segmented k-space magnetic resonance velocimetry

Conventional non-segmented magnetic resonance phase velocity mapping (MRPVM) is an accurate but relatively slow velocimetric technique. Therefore, the aim of this study was to evaluate the accuracy of the much faster segmented k-space MRPVM in quantifying flow. The axial velocity was measured in four straight tubes (inner diameter: 5.6–26.2 mm), using a segmented MRPVM sequence with seven lines of k-space per segment. The flow rate and flow volume were accurately quantified (errors<5%) under steady (r 2=0.99) and pulsatile flow (r 2=0.98), respectively. The measured velocity profiles and flow rates from the segmented sequence agreed with those from the non-segmented (p>0.05). Changing the slice thickness or the field of view did not affect the accuracy of the measurements. The results of this study suggest that fast, segmented MRPVM can be used for accurate flow quantification.

The quantification of pulmonary valve haemodynamics using MRI.

The optimum slice location within the pulmonary root to quantify pulmonary valve haemodynamics was examined using magnetic resonance (MR) phase velocity mapping. MRI was carried out on 15 patients with congenital aortic valve disease. Although the patients had aortic valve disease, all measurements were made on the pulmonary valve. Systolic (Q(SYS)) and diastolic (Q(DIAS)) blood flow volume and cardiac index (CI) were determined at four pulmonary artery locations. The change in diastolic flow volume relative to slice 1, closest to the pulmonary valve, was also calculated. For a change in axial position of 1.5 cm, i.e. from 0.5 to 2 cm from the annulus, there was a change in diastolic flow volume of 4.4 ml. There was a significant increase in the mean diastolic flow from 3.4 to 7.7 ml (p = 0.01 between slice positions 0.5 and 2 cm. However, there was no significant change in CI, 3.4-3.7 l/min/m2 (p = 0.14) over the same distance. We believe that two factors are responsible for these results. The first is that of compliance, whose effects can be minimized by placing the MR slice close to the valve, however, this will not account for the second factor, being that of valve motion, and hence diastolic pulmonary valve flow or regurgitant volume will be underestimated. The degree of underestimation may only be important at mild and moderate levels of regurgitation or if changes in regurgitation are to be temporally measured.

Evaluation of the precision of magnetic resonance phase velocity mapping for blood flow measurements.

Evaluating the in vivo accuracy of magnetic resonance phase velocity mapping (PVM) is not straightforward because of the absence of a validated clinical flow quantification technique. The aim of this study was to evaluate PVM by investigating its precision, both in vitro and in vivo, in a 1.5 Tesla scanner. In the former case, steady and pulsatile flow experiments were conducted using an aortic model under a variety of flow conditions (steady: 0.1-5.5 L/min; pulsatile: 10-75 mL/cycle). In the latter case, PVM measurements were taken in the ascending aorta of ten subjects, seven of which had aortic regurgitation. Each velocity measurement was taken twice, with the slice perpendicular to the long axis of the aorta. Comparison between the measured and true flow rates and volumes confirmed the high accuracy of PVM in measuring flow in vitro (p > 0.85). The in vitro precision of PVM was found to be very high(steady: y = 1.00x + 0.02, r = 0.999; pulsatile: y = 0.98x + 0.72, r = 0.997; x: measurement #1, y: measurement #2) and this was confirmed by Bland-Altman analysis. Of great clinical significance was the high level of the in vivo precision (y = 1.01x - 0.04, r = 0.993), confirmed statistically (p = 1.00). In conclusion, PVM provides repeatable blood flow measurements. The high in vitro accuracy and precision, combined with the high in vivo precision, are key factors for the establishment of PVM as the "gold-standard" to quantify blood flow.

Fluid mechanic assessment of the total cavopulmonary connection using magnetic resonance phase velocity mapping and digital particle image velocimetry.

The total cavopulmonary connection (TCPC) is currently the most promising modification of the Fontan surgical repair for single ventricle congenital heart disease. The TCPC involves a surgical connection of the superior and inferior vena cavae directly to the left and right pulmonary arteries, bypassing the right heart. In the univentricular system, the ventricle experiences a workload which may be reduced by optimizing the cavae-to-pulmonary anastomosis. The hypothesis of this study was that the energetic efficiency of the connection is a consequence of the fluid dynamics which develop as a function of connection geometry. Magnetic resonance phase velocity mapping (MRPVM) and digital particle image velocimetry (DPIV) were used to evaluate the flow patterns in vitro in three prototype glass models of the TCPC: flared zero offset, flared 14 mm offset, and straight 21 mm offset. The flow field velocity along the symmetry plane of each model was chosen to elucidate the fluid mechanics of the connection as a function of the connection geometry and pulmonary artery flow split. The steady flow experiments were conducted at a physiologic cardiac output (4 L/min) over three left/right pulmonary flow splits (70/30, 50/50, and 30/70) while keeping the superior/inferior vena cavae flow ratio constant at 40/60. MRPVM, a noninvasive clinical technique for measuring flow field velocities, was compared to DPIV, an established in vitro fluid mechanic technique. A comparison between the results from both techniques showed agreement of large scale flow features, despite some discrepancies in the detailed flow fields. The absence of caval offset in the flared zero offset model resulted in significant caval flow collision at the connection site. In contrast, offsetting the cavae reduced the flow interaction and caused a vortex-like low velocity region between the caval inlets as well as flow disturbance in the pulmonary artery with the least total flow. A positive correlation was also found between the direct caval flow collision and increased power losses. MRPVM was able to elucidate these important fluid flow features, which may be important in future modifications in TCPC surgical designs. Using MRPVM, two- and three-directional velocity fields in the TCPC could be quantified. Because of this, MRPVM has the potential to provide accurate velocity information clinically and, thus, to become the in vivo tool for TCPC patient physiological/functional assessment.

Nota bene. Contortions of the heart.

T he mammalian heart is an odd-looking organ, with four different-sized muscular chambers (two atria on top and two ventricles beneath) and a spectacular array of curving blood vessels that loop around each other in a contorted plumbing system. In a recent report, Kilner et al. ([1][1]) reveal that blood does not flow in a smooth, continuous stream but rather adopts asymmetric swirling patterns as it moves through the chambers of the heart. These whirlpools of blood may seem to be the antithesis of a well-heeled plumbing system, but the investigators argue that they are in fact beneficial, aiding the heart as its contractions increase during exercise. The heartbeat has two principal phases: systole (contraction) and diastole (relaxation). In systole, the muscular walls of the ventricles contract, squeezing blood from the left ventricle into the aorta and from the right ventricle into the pulmonary artery. Meanwhile, the atrial walls relax and blood flows from the pulmonary veins into the left atrium and from the superior and inferior caval veins into the right atrium. In diastole, the walls of the ventricles relax (those of the atria contract), the tricuspid and mitral valves open, and blood flows from the right atrium into the right ventricle and from the left atrium into the left ventricle. With the help of a noninvasive imaging technique, magnetic resonance phase-velocity mapping, Kilner et al. investigated changes in blood flow during the diastole and systole of successive heartbeats in healthy human volunteers at rest. They found that blood flow was asymmetric and that the streaming patterns that developed during diastole and systole were characteristic for each chamber. For example, as the ventricles contract, blood streaming into the expanding right atrium (RA) from the superior and inferior caval veins (SVC; IVC) does not collide but rather adopts a forward rotating, swirling motion that redirects flow toward the tricuspid valve (TV) (see figure, top left). The ventricles then relax and the blood moves leftward out of its clockwise rotation (away from the viewer) and through the tricuspid valve (top right). As the left ventricle (LV) contracts, blood is ejected into the aorta (bottom left) through the aortic valve (AV); then, as the left ventricle relaxes, blood streams into it from the left atrium through the mitral valve (MV), setting up a whirlpool below the valves' flexible leaflets (bottom right). ![Figure][2] But why has the mammalian heart evolved such a sophisticated pattern of blood flow? Earlier work by Kilner's group using echocardiography to look at heart blood flow in exercising volunteers yields a possible answer. As the heart rate speeds up to at least double its normal rate during vigorous exercise, the contractions of the ventricles and atria alternate rapidly. The authors propose that asymmetries in blood flow conserve energy (by minimizing energy-consuming collisions between opposing streams of blood) and help to redirect blood movement, enabling its smooth passage to the next destination. They propose that, with exertion, the whirlpools of ventricular blood are hurtled in a slingshot motion along the sinuous curves of the blood vessels, minimizing disruption to the flow. Indirect evidence that the contorted curves of the vertebrate heart may be associated with dynamic efficiency comes from studying invertebrates such as the snail (not known for its virtuosity on the race track), which has a simple, linear two-chambered heart (with not a curve in sight). The researchers hope to test their hypothesis with a more advanced magnetic resonance system that should provide the rapid imaging necessary to follow the dynamics of heart blood flow in exercising volunteers. 1. [↵][3]1. P. J. Kilner 2. et al. , Nature vol. 404, (2000), p. 759. [OpenUrl][4][CrossRef][5][PubMed][6][Web of Science][7] [1]: #ref-1 [2]: pending:yes [3]: #xref-ref-1-1 View reference 1 in text [4]: {openurl}?query=rft.jtitle%253DNature%26rft.stitle%253DNature%26rft.aulast%253DKilner%26rft.auinit1%253DP.%2BJ.%26rft.volume%253D404%26rft.issue%253D6779%26rft.spage%253D759%26rft.epage%253D761%26rft.atitle%253DAsymmetric%2Bredirection%2Bof%2Bflow%2Bthrough%2Bthe%2Bheart.%26rft_id%253Dinfo%253Adoi%252F10.1038%252F35008075%26rft_id%253Dinfo%253Apmid%252F10783888%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [5]: /lookup/external-ref?access_num=10.1038/35008075&link_type=DOI [6]: /lookup/external-ref?access_num=10783888&link_type=MED&atom=%2Fsci%2F288%2F5465%2F453.atom [7]: /lookup/external-ref?access_num=000086523600051&link_type=ISI

A semi‐automated method to quantify left ventricular diastolic inflow propagation by magnetic resonance phase velocity mapping

A semi‐automated method to quantify left ventricular diastolic inflow propagation by magnetic resonance phase velocity mapping

The importance of slice location on the accuracy of aortic regurgitation measurements with magnetic resonance phase velocity mapping.

Although several methods have been used clinically to evaluate the severity of aortic regurgitation, there is no purely quantitative approach for aortic regurgitant volume (ARV) measurements. Magnetic resonance phase velocity mapping can be used to quantify the ARV, with a single imaging slice in the ascending aorta, from through-slice velocity measurements. To investigate the accuracy of this technique, in vitro experiments were performed with a compliant model of the ascending aorta. Our goals were to study the effects of slice location on the reliability of the ARV measurements and to determine the location that provides the most accurate results. It was found that when the slice was placed between the aortic valve and the coronary ostia, the measurements were most accurate. Beyond the coronary ostia, aortic compliance and coronary flow negatively affected the accuracy of the measurements, introducing significant errors. This study shows that slice location is important in quantifying the ARV accurately. The higher accuracy achieved with the slice placed between the aortic valve and the coronary ostia suggests that this slice location should be considered and thoroughly examined as the preferred measurement site clinically.

Slice location dependence of aortic regurgitation measurements with MR phase velocity mapping.

Although several methods have been used clinically to assess aortic regurgitation (AR), there is no "gold standard" for regurgitant volume measurement. Magnetic resonance phase velocity mapping (PVM) can be used for noninvasive blood flow measurements. To evaluate the accuracy of PVM in quantifying AR with a single imaging slice in the ascending aorta, in vitro experiments were performed by using a compliant aortic model. Attention was focused on determining the slice location that provided the best results. The most accurate measurements were taken between the aortic valve annulus and the coronary ostia where the measured (Y) and actual (X) flow rate had close agreement (Y = 0.954 x + 0.126, r2 = 0.995, standard deviation of error = 0.139 L/min). Beyond the coronary ostia, coronary flow and aortic compliance negatively affected the accuracy of the measurements. In vivo measurements taken on patients with AR showed the same tendency with the in vitro results. In making decisions regarding patient treatment, diagnostic accuracy is very important. The results from this study suggest that higher accuracy is achieved by placing the slice between the aortic valve and the coronary ostia and that this is the region where attention should be focused for further clinical investigation.