Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kozerke, S.
Right arrow Articles by Boesiger, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kozerke, S.
Right arrow Articles by Boesiger, P.
(Radiology. 2001;218:548-555.)
© RSNA, 2001

Cardiac Imaging

Heart Motion-adapted MR Velocity Mapping of Blood Velocity Distribution Downstream of Aortic Valve Prostheses: Initial Experience1

Sebastian Kozerke, PhD, J. Michael Hasenkam, MD, PhD, Hans Nygaard, PhD, Peter K. Paulsen, MD, PhD, Erik M. Pedersen, MD, PhD and Peter Boesiger, PhD

1 From the Institute of Biomedical Engineering and Medical Informatics, University of Zurich and Swiss Federal Institute of Technology, Gloriastrasse 35, 8092 Zurich, Switzerland (S.K., P.B.); and the Department of Cardiothoracic and Vascular Surgery and MR Center, Institute of Experimental Clinical Research, Aarhus University Hospital, Denmark (J.M.H., H.N., P.K.P., E.M.P.). Received January 6, 2000; revision requested February 28; revision received April 12; accepted June 1. Supported by EUREKA grant E!2061 INCA-MRI; Swiss Commission for Technology and Innovation grant 4178.1; Danish Heart Foundation grant 97-2-1-5-22549; Karen Elise Jensen Foundation, Denmark; Desirée and Niels Yde Foundation, Denmark; and Philips Medical Systems, Best, the Netherlands. Address correspondence to P.B. (e-mail: boesiger@biomed.ee.ethz.ch).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
PURPOSE: To investigate blood flow velocities and shear rates at two distances downstream of an artificial aortic valve in patients.

MATERIALS AND METHODS: Blood velocity was quantified downstream of the valve prosthesis (for replacement after aortic valve stenosis or combined stenosis and regurgitation) in 10 patients by using a magnetic resonance (MR) cine velocity mapping method in which the imaging section position is adapted according to the excursion of the valvular plane of the heart. Two acquisitions were performed to display the blood velocity distributions one-fourth valve diameter and one valve diameter downstream of the valve and to quantify blood volumes and shear rates.

RESULTS: The velocity profiles measured during flow acceleration one-fourth valve diameter downstream were characterized by a distinct pattern of two lateral jets and one central jet of antegrade flow. High shear rates were found along the leaflet tips. The profiles obtained one valve diameter downstream were skewed, with varying velocity patterns among patients. Peak shear rates were found close to the vessel wall. With correction for through-plane motion of the valve, the mean apparent regurgitant fraction (± SD) was 14% ± 6; the mean regurgitant fraction without correction was 9% ± 5.

CONCLUSION: The described noninvasive procedure for velocity mapping enables measurements close to the valve and thus evaluation of blood flow patterns with respect to valve design in humans.

Index terms: Heart, flow dynamics, 51.91, 535.91 • Heart, MR, 51.121412, 51.12144 • Heart, valves, 535.453 • Magnetic resonance (MR), cine study, 51.12144 • Magnetic resonance (MR), physics, 51.121412, 51.12144 • Phantoms


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Late complications in patients with an implanted prosthetic heart valve are considered to be related to fluid dynamic characteristics. Thus, hemolysis and thrombus formation have been associated with the heart valve design (1–3). Complications may derive from elevated levels of turbulent shear stress, which is the random time-dependent fluid mechanical stress that acts tangentially on the surface of an element in the blood (4). Specific blood flow patterns have been considered to propagate thrombus formation (5).

Nygaard et al (6–8) provided early data on velocity fields and shear stress distal to prosthetic valves in humans. However, the invasive measurement procedure used in that study can be performed only during open-heart surgery.

Cine magnetic resonance (MR) phase-contrast measurements enable noninvasive mapping of velocity distributions. The modality has the potential to provide more detailed information with regard to velocity distributions than does echocardiography, which is currently the reference standard for evaluating heart valves in vivo. Furthermore, MR velocity mapping enables quantification of transvalvular blood volumes as important markers of valve performance.

MR velocity mapping is based on the determination of the acquired phase shift of proton spins moving along a magnetic field gradient. Owing to magnetic field inhomogeneity and susceptibility variations, it is necessary to perform two acquisitions with different velocity sensitivities. The subtraction of the phase-contrast images from the two acquisitions yields a phase-difference map, which is proportional to the velocities.

The nature of blood flow is primarily determined by the ratio of inertial forces to viscous forces. Predominant inertial forces result in turbulence. The locations of turbulence are associated with areas of high-velocity gradients (1,2,6,7). Therefore, mapping of velocity gradients might be used to estimate the complexity of flow fields.

Conventional MR imaging techniques involve a spatially fixed imaging section. Because of cardiac motion, measurements in the immediate vicinity of the heart valve prostheses are not possible. However, the highest flow complexity occurs immediately downstream of prosthetic valves. Measurements further downstream, depending on varying geometric conditions, such as aortic curvature and the valve position relative to the outflow axis, are expected to exhibit alterations in velocity fields. To provide reference data regarding velocity fields downstream of valve prostheses, in vivo data should be acquired as close as possible to the valve.

By using a method for MR velocity mapping and adapting the imaging section position to the cyclic motion of the heart, we tested our hypothesis that motion-corrected velocity and shear rate data can be obtained immediately downstream of valve prostheses in humans.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Patient Population
Ten patients (five men, five women; mean age, 58.4 years; age range, 44–66 years) who had aortic valve implants (St Jude Medical, St Paul, Minn) underwent MR examinations. Selection criteria were (a) implantation of the aortic valve prosthesis 25–49 months before our investigation—to circumvent influences from operative trauma around the valve prosthesis and to evaluate the patients before pannus formation, (b) no obesity, and (c) no claustrophobia. The actual time between valve replacement and MR examination was 43 months on average (range, 25–49 months). Seven patients had had aortic valve stenosis, and three had had combined stenosis and regurgitation before surgery. Written informed consent was obtained from all patients. The study was approved by the local ethics committee.

MR Examination
MR measurements were obtained by using a Gyroscan NT 1.5-T whole-body unit (Philips Medical Systems, Best, the Netherlands) equipped with a gradient system that delivers a maximum gradient amplitude of 21 mT/m with a slew rate of 105 mT/m/msec. The patients were examined while in the supine position. A five-element cardiac phased-array coil was used for receiving signal.

Heart Motion Adaptation
Adaptation of the actual section position to the active motion of the heart was achieved by using a recently proposed three-step strategy (9).

Labeling.—A cine imaging sequence with a labeling prepulse was applied within a single breath hold to mark points of the basal plane of the heart. By using a subtraction technique, three points in the basal myocardium, moving during the cardiac cycle, were marked and imaged with a temporal resolution of 30 msec (Fig 1, top). Image processing was used to automatically extract the positions of the three points for each phase of the cardiac cycle. Accordingly, the cyclic motion at the basal plane level of the heart in an individual patient could be quantified.



View larger version (66K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Illustration of the principle of heart motion adaptation. Top: By using a cine MR imaging sequence, with preparation of magnetization at the basal level of the heart, points in the right ventricular (RV), septal, and left ventricular (LV) wall moving during the cardiac cycle are marked. Automatic tracing of the basal markers yields the quantitative trajectory of basal motion, which is mapped to the aortic valve level. Bottom: For velocity mapping, the imaging section perpendicular to the aortic root (white line) is adapted to cardiac motion for every required time frame within the cardiac cycle. The three lines that branch from one vertical line represent the different times (62, 132, and 377 msec) during the cardiac cycle.

 
Valve localization.—Coronal and sagittal cine acquisitions were performed with navigator gating to localize the aortic valve annulus. The end-expiratory position of the patient was determined by acquiring eight navigator echoes, each of which was triggered on the R wave at electrocardiography. By cross correlating the first echo with the consecutive signals, the maximum deviation that corresponded to end expiration was defined. The navigator signal that represented end expiration was used for subsequent cross correlations. By extracting the most frequent diaphragm position from the histogram of diaphragm deviations recorded by using 16 successive navigator echoes, the optimal gating level was determined. The gating window was placed asymmetrically around the optimal gating level.

Velocity mapping.—Adaptation of the imaging section position according to the basal plane motion was used to map blood velocities in the vicinity of the aortic valve (Fig 1, bottom). The trajectory of heart motion was mapped to the initial section position either one-fourth valve diameter or one valve diameter downstream of the prosthesis.

For velocity mapping, a gradient-echo phase-contrast sequence that interleaved the two velocity-encoding segments was applied. In-plane resolution was 1 x 1 mm2 with a field of view of 256 x 180 mm2, imaging matrix of 256 x 180, and section thickness of 5 mm. Velocity data were recorded in intervals of 30 msec throughout the cardiac cycle, starting 62 msec after the R wave. The initial delay after R-wave detection was required for execution of the navigator pulse, processing of the navigator signal, and the algorithm for motion-adapted gating. The aliasing velocity in the section-selecting direction was set at 160 cm/sec. Accordingly, velocities within a range of -160 to +160 cm/sec could be measured without wraparounds. In the frequency-encoding and phase-encoding directions, flow-compensating gradients were incorporated. To reduce dephasing effects in complex flow fields, 75% of the full echo was acquired to yield an echo time of 3.9–4.0 msec. Missing samples were zero padded, and an asymmetric weighting function was applied to central k-space data before Fourier transformation.

Accounting for an average gating efficiency of about 75% and a heart rate of 70 beats per minute, the imaging duration was 3 minutes 30 seconds. The measurements one-fourth valve diameter and one valve diameter downstream were performed successively.

Respiratory Motion Compensation
To compensate for the passive motion of the heart due to respiration, a two-dimensional selective excitation pulse (navigator) was applied to detect the actual breathing state of the patient. The pencil-beam navigator pulse was positioned through the dome of the right hemidiaphragm and applied immediately after R-wave detection.

To overcome the shortcoming of low imaging efficiency associated with conventional gating (10), the motion-adapted gating approach (11) was implemented for the velocity-encoding sequence. In real time, the phase-encoding step was calculated according to the actual shift of the diaphragm with respect to the end-expiration level. The phase-encoding number was determined from the actual diaphragm displacement by using a cubic weighting function. Diaphragm displacements within a 10-mm constant gating window were marked as valid for motion-adapted gating. Larger displacements led to rejection of the acquired data.

In addition to navigator gating, adaptive real-time motion correction of the imaging section position was performed (12). The imaging section was shifted according to the current diaphragm displacement within the gating window. A factor of 0.6 was used to scale the diaphragm motion to the craniocaudal motion of the imaging plane (13).

Data Analysis
Initially, phase errors from concomitant gradient fields were corrected by using the method described by Bernstein et al (14). Phase shifts caused by eddy currents were compensated for by subtracting a linear two-dimensional function from the original phase data. The linear function was obtained from a least-squares fit to the phase of static tissue. Static regions were automatically identified as those pixels, on the phase images, that exhibited a low SD during the cardiac cycle (15). Phase pixels with a time SD lower than 4% of the maximum time SD found were considered to be static. Vessel contours were extracted by using active contour-based segmentation (16).

Because the measured velocity data represent a superposition of axial flow velocities through the valve and the through-plane velocity of the section of interest, the section through-plane velocity component was subtracted to obtain the velocity of the blood flow without the influence of aortic root motion. By integrating the pixel velocity over the contour area, the mean blood flow of each heart phase was calculated. The mean forward and reverse flow values were extracted by integrating either the positive or negative blood velocities over the contour area. To match the different heart rates of the patients, the time axis was normalized by assigning the end systole to 100%.

Systolic volume was defined as the volume ejected between 20% and 100% of end systole. Regurgitant volume was obtained by integrating blood flow during diastole (Fig 2). The regurgitant fraction was calculated as the ratio of regurgitant volume to forward volume.



View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Graph illustrates the calculated systolic volume (Vsys) and diastolic regurgitant volume (Vreg).

 
Peak shear rates (ie, spatial velocity gradients) were calculated to assess areas of high shear stress as follows: By using median filtering with a mask width of three pixels, single isolated peaks were removed from the velocity data. A central difference scheme was used to calculate the first derivatives in the four radial directions of a single pixel. The maximum absolute value of the four derivatives was determined and assigned as the peak shear rate to the single pixel. This procedure was repeated for all pixels within the contour area of each heart phase image.

Processing for phase correction and vessel segmentation on a DEC-Alpha workstation (Digital Equipment Corporation, division of Compaq, Manalapan, NJ) took 15–20 minutes per patient. Values for blood flow, blood volumes, and shear rates were readily available thereafter.

Statistical Analyses
A paired Student t test was performed to determine differences in the following parameters. The regurgitant volume based on velocity data measured one-fourth valve diameter downstream, with correction for through-plane motion, was compared with data in which the through-plane motion correction process was skipped. The ratio of reverse systolic blood volume to forward systolic blood volume measured one-fourth valve diameter downstream was compared with this ratio based on data measured one valve diameter downstream. A P value of less than .05 was considered to prove the statistical significance of the observed differences.

Phantom Validation
To validate the accuracy of MR phase-contrast velocity mapping for measurements downstream of the artificial heart valve, phantom experiments were conducted. A Plexiglas pipe equipped with a bileaflet aortic heart valve was used in a steady flow setup (diameter of valve orifice, 23 mm; pipe diameter downstream of inserted valve, 38 mm; flow rate, 15 L/min). Fluid viscosity and relaxation time constants, T1 and T2, were matched to the values of blood (viscosity, 3.5 mPa · sec; T1, 941 msec; T2, 232 msec). MR velocity mapping was performed 8 mm downstream of the valve, and the data were compared with those measured with laser Doppler anemometry. Linear regression analysis was used for comparison. Laser Doppler anemometry is considered the reference standard for registration of fluid velocity in vitro. The in vitro parameters for MR velocity mapping were identical to those used for the in vivo measurements.

To substantiate the hypothesis of a correlation between high-velocity gradients and regions of highly complex flow spatial velocity, gradients were calculated at MR phase-contrast imaging and compared with the laser Doppler anemometry measurements of fluctuating velocity. Local maximum velocity gradients were extracted from the MR velocity data and compared with local maximums of fluctuating velocity by using regression analysis.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
The trajectory of basal plane motion was automatically extracted from the basal marker in all patients and enabled calculation of the section position and correction for through-plane motion for each phase of the cardiac cycle.

Aortic Root Motion
The mean maximum through-plane motion of the aortic root (± SD) was 7 mm ± 2 (range, 4–11 mm). The peak velocity of through-plane motion of the aortic root was 6 cm/sec ± 1 (range, 4–7 cm/sec) and observed in early diastole (140%–160% of end systole) (Fig 3).



View larger version (39K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Graph illustrates the mean through-plane motion (a) of the prosthetic valve in the 10 patients and the corresponding mean velocity of through-plane motion (b). Most rapid through-plane motion occurs during early diastole. The small leakage flow through the valve is therefore likely to be underestimated. The error bars represent SDs.

 
The maximum rotation of the prosthetic valve around the anteroposterior axis during the cardiac cycle was 5° (range, 4°–9°); the rotation around the right-left axis was 1° (range, 0°–3°).

With correction for through-plane motion, the mean apparent regurgitant volume (± SD) measured one-fourth valve diameter downstream of the valve was 10 mL ± 4; the mean regurgitant volume without through-plane motion correction was 6 mL ± 3. The corresponding mean regurgitant fraction was 14% ± 6 with correction for through-plane motion; the mean regurgitant fraction without through-plane motion correction was 9% ± 5 (Table 1).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Comparison of Systolic and Regurgitant Blood Volumes
 
Blood Velocity Display
The time-resolved blood velocity distributions in the two measured downstream positions for eight consecutive heart phases in a representative subject (patient 8) are illustrated in Figure 4. The temporal resolution was 30 msec, and the first heart phase image shown was acquired 122 msec after the R wave. On the color-encoded maps in Figure 5, forward flow is represented in shades of red and reverse flow velocities are represented in shades of blue to indicate regions of opposite flow during systolic ejection with respect to the valve position, starting with the first heart phase at 152 msec after the R wave.



View larger version (32K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Graphs illustrate the axial velocity distribution during systolic ejection in eight consecutive heart phases at the two measured distances, one-fourth valve diameter (top) and one valve diameter (bottom) downstream of the valve, in a 66-year-old man. The first time frame shown (top and bottom, far left) is 122 msec after the R wave; the consecutive phases are 30 msec apart. Flow through the two major orifices of the valve generates lateral jets (LJ). The central jet (CJ), which corresponds to the central slit of the prosthetic valve, is clearly visible in the third time frame (top, third drawing from left) in the measurement closest to the valve. As blood flows downstream, the central jet migrates into the right lateral jet (bottom). A = anterior, L = left, P = posterior, R = right.

 


View larger version (74K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5. Color-encoded axial flow velocity images measured one-fourth valve diameter (top) and one valve diameter (bottom) downstream of the valve in five consecutive heart phases, each of which is separated by 30 msec, starting 152 msec after the R wave. The red shades represent forward flow velocities, and the blue shades represent reverse flow velocities. Values are given in centimeters per second. A = anterior, P = posterior, R = right.

 
One-fourth valve diameter downstream.—Two lateral jets that corresponded to the two major orifices of the valve were generated during early flow acceleration (Fig 4, top). Later during the acceleration phase, a small central jet associated with the central slit of the valve was observed. The triple-jet pattern close to the valve was a consistent finding in all patients (Table 2). During flow acceleration, reverse flow within the three sinuses of Valsalva occurred (Fig 5, top). As flow decelerated, the velocity profiles became skewed, with regions of reverse flow predominantly near the left vessel wall.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Patient Data
 
One valve diameter downstream.—The evolution of the flow velocity distribution during flow acceleration was characterized by a blunt profile that developed into a skewed distribution with either a single velocity jet or two velocity jets migrating toward the outer curvature of the aorta. The jets that corresponded to the orifices of the valve could no longer be discerned (Fig 4, bottom). Reverse flow occurred in areas near the left vessel wall (Fig 5, bottom), except in two patients, in whom reverse flow was found near the anterior vessel wall. The location of highest velocity and the vessel area occupied by reverse flow were found to be independent of the rotation of the valve (Table 2).

Blood Flow
There was good agreement between the stroke volumes calculated independently in the two downstream locations in all 10 patients (Table 2). The mean difference in stroke volume (± 2 SDs) between the two locations was 0 mL ± 6.

To quantify the differences between forward and reverse blood volumes with respect to the distance from the valve, the forward and reverse blood volumes during systolic ejection were compared separately (Fig 6). At the closest distance from the valve, the systolic forward and reverse blood volumes were substantially greater compared with the volumes obtained further downstream (Table 2). The mean regurgitant volume (± SD) for all patients and all measured positions amounted to 10 mL per beat ± 5.



View larger version (35K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6. Graph illustrates blood flow rates measured at the two distances, one-fourth valve diameter (diam) and one valve diameter downstream of the valve, in a 66-year-old man. Flow is separated into that in the forward (Qant) and reverse (Qret) directions. The smaller ratio of antegrade-to-retrograde flow during systole one-fourth valve diameter downstream, compared with this ratio measurement one valve diameter downstream, indicates greater recirculation.

 
Shear Rates
The time-resolved maps with magnitudes of calculated shear rates encoded into pixel brightness are illustrated in Figure 7.



View larger version (76K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 7. Shear rate maps with measurements one-fourth valve diameter (top) and one valve diameter (bottom) downstream of the valve in five consecutive heart phases, each of which is separated by 30 msec, starting 152 msec after the R wave. The position and size of the valves are indicated by the dashed circles. Values are given in sec-1. Peak shear rates occur near the leaflet tips of the valve (top). As blood flows downstream, velocity jets decay and reattach to the vessel wall (bottom). A = anterior, P = posterior, R = right.

 
One-fourth valve diameter downstream.—Initially, a high shear boundary developed along the two leaflets, as seen in the first time frame illustrated in Figure 7 (top). The peak shear rate value was detected 212 msec after the R wave at a location that corresponded to the right leaflet tip. Later, during flow deceleration, the magnitude of local maximum shear rate values decreased rapidly and the shear boundary was distributed along the valve inner circumference. In general, the separation of the shear boundary from the vessel wall was seen.

One valve diameter downstream.—The peak shear values one valve diameter downstream were considerably smaller compared with those closest to the valve. The shear boundary was reattached to the right vessel wall, and, accordingly, high shear rate values were found along this site (Fig 7, bottom). The maximum shear rate occurred 242 msec after the R wave.

Phantom Validation
In vitro, the difference between velocities measured with MR and those measured with laser Doppler anemometry downstream of the artificial valve was within 12% for the central 75% of the pipe diameter (Fig 8, A). Linear regression analysis yielded a correlation coefficient (r2) of 0.99; the standard error of the estimate was 4 cm/sec. The locations along the pipe radius that exhibited high shear rates, as derived from MR velocity data, correlated well with the peak values of fluctuating velocity measured with laser Doppler anemometry (r2 = 1.00; standard error of the estimate, 1 mm) (Fig 8, B).



View larger version (25K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 8. Graph illustrations of phantom validations. A, Velocities measured with MR imaging (MRI) and laser Doppler anemometry (LDA) 8 mm distal from the artificial aortic valve. B, Comparison of velocity gradients (dv/ds), or shear rates, derived from MR data and fluctuating velocities (ie, root mean square [RMS] values) measured with laser Doppler anemometry (-{circ}-) show agreement in the spatial positions of the peak values.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
REFERENCES
 
High-velocity jets associated with artificial heart valves may cause hemolysis and clot formation and lead to aneurysmal dilatation of the ascending aorta due to increased regional wall shear stress. The implications of regurgitation of prosthetic valves are different from those of regurgitation of normal valves. Reverse blood flow through the narrow central slit of a bileaflet valve can create relatively high laminar shear stresses and thus lead to increased blood cell damage (17). On the other hand, regurgitation might be important to avoid stagnant flow in the hinge areas of the valve.

MR velocity mapping is a noninvasive modality that can provide detailed information on velocity distributions. However, because of the cyclic motion in the valvular plane of the heart, conventional MR velocity mapping is limited to measurements downstream of the sino-tubular junction of the aorta. The described method of MR velocity mapping enabled us to assess the blood flow patterns closer to an artificial aortic valve in patients than has been possible previously.

Through-Plane Motion
The considerable through-plane motion of the prosthetic valve was quantified (mean maximum motion, 7 mm) and thus could not be neglected when measurements in the immediate vicinity of the valve were performed. The through-plane motion measured in this study was smaller compared with that in a previous investigation with young healthy volunteers (9). The differences might be attributed to age-related stiffening of the heart and the fact that the patients underwent open-heart surgery and had consequent scar tissue.

The rapid relaxation of the heart and thus rapid through-plane motion of the valve during diastole led to an underestimation of regurgitant flow. However, regurgitant flow was small (mean, 10 mL) compared with systolic blood flow. Therefore, the velocity-to-noise ratio for the diastolic velocity data was approximately 2.5 times lower than that for the peak systolic velocity data. The use of adaptable velocity encoding (18) would be beneficial in future studies. Nevertheless, the estimate of regurgitation in our study compared well with related data in the literature (19).

Blood Velocity Distribution
The velocity data obtained shortly before peak systole one-fourth valve diameter downstream of the valve clearly disclosed the valve design: Three velocity jets emerged from the two major lateral orifices and the small central slit of the valve. This distinct triple-jet pattern has been verified by using laser Doppler anemometry and MR imaging in in vitro steady-flow experiments with models simulating peak systolic flow (20).

Reverse flow within the sinuses of Valsalva, which was indicative of recirculation, occurred during systolic flow acceleration. This finding has been observed in vitro under pulsatile flow conditions (21). Our comparison of systolic forward and reverse flow indicated the existence of recirculation in the cavities of the sinuses during flow acceleration. Flow quantification revealed substantially more antegrade and retrograde flow at the distance closest to the valve.

During early flow acceleration, the two lateral jets were farther apart than they were later during the acceleration phase. This seemed to be caused by flow that was initially accelerated by the opening leaflets toward the vessel wall. Later during systole, flow was directed more in the axial direction of the aorta, which was indicative of fully opened leaflets. The observed symmetry of velocity distribution that corresponded to the two lateral orifices of the valve suggested an optimal utilization of flow area across the valve during systolic flow acceleration. The velocity profiles obtained one valve diameter downstream of the valve exhibited substantial changes compared with the patterns observed close to the valve, with large heterogeneity among the patients. As blood flowed downstream, the skewness of velocity profiles became more accentuated: Either a single velocity jet or two jets migrated toward the right vessel wall. These alterations were considered to be attributed to differences in aortic root geometry and aortic curvature among the patients.

Asymmetry of the velocity jet pattern has been associated with nonoptimal use of the flow area across the valve (22). However, this conclusion was based on data measured one valve diameter downstream of the valve and therefore might be misleading given the large variation in flow profiles among subjects. In this study, measurements one-fourth valve diameter downstream revealed lateral velocity jets of almost equal magnitude.

Shear Rate Distribution
Peak velocity gradients were detected one-fourth valve diameter downstream immediately distal to the leaflet tips during the deceleration phase in all patients. The highest normal Reynolds stresses measured 1.5–2.0 vessel diameters downstream of the valve in humans typically have been at locations that corresponded to the central slit and the hinges of the valve (7,8). However, in vitro stress measurements performed closer to the valve (one-half diameter downstream) have revealed high turbulent shear stresses also immediately distal to the valvular leaflets (8). As blood flowed further downstream, the peak velocity gradients became considerably smaller, as observed in the present study, and were located along the right posterior vessel wall; this indicated the reattachment of the velocity jet to the vessel wall.

Technical Considerations
MR velocity mapping has limited accuracy when measurements are performed in disturbed or turbulent flow fields (23,24). Turbulence has not been shown to cause distortion of velocity data in straight-tube flow (25). Unlike separated flow, like that found with stenosis, in fully developed turbulent flow the intravoxel phase dispersion can be assumed to be small, because a small random time-dependent velocity component is superimposed on a large mean velocity component. However, flow separation occurs downstream of artificial heart valves owing to the obstruction of flow caused by the hinge mechanisms of the valve and by the leaflets themselves. Total signal voids were not seen; however, slight signal losses were visible on the modulus images, and, thus, phase distortions that affected the accuracy of the velocity encoding were possible.

Nevertheless, in vitro, the differences in velocity profiles downstream of the artificial aortic valve measured with MR and laser Doppler anemometry have been reported to be within 10% (20) with parameter settings (1 x 1-mm2 spatial resolution, 4-msec echo time) similar to those used in the present study. Reports (20,26) indicate that MR velocity mapping with short echo times can be used to assess velocity distributions downstream of artificial heart valves in vivo.

MR measurements closer than one-fourth valve diameter downstream of the valve were not possible because of susceptibility changes around the prosthetic valve and interference of the open-valve leaflets with the imaging plane.

In conclusion, the MR velocity mapping method described in this study enabled the in vivo assessment of blood flow patterns in the near vicinity of the artificial aortic valve and the quantification of regurgitant blood flow through the valve.

The velocity distributions measured in the immediate vicinity of the valve reflected the valve design and provided information about flow distribution through the leaflets. The velocity profiles obtained further downstream of the valve were severely altered and varied among patients owing to effects of the individual aortic geometries.

The described MR velocity mapping method enabled detailed assessment of valve-induced changes in blood flow and quantification of transvalvular blood volumes in humans with an implanted prosthetic valve. Data on in vivo near-valve velocity distributions are an important basis for the evaluation of prosthetic valve designs.


   FOOTNOTES
 
Author contributions: Guarantors of integrity of entire study, P.B., E.M.P., P.K.P.; study concepts and design, S.K., J.M.H., E.M.P.; definition of intellectual content, P.B., J.M.H., P.K.P.; literature research, S.K., J.M.H., H.N.; clinical studies, S.K., J.M.H., E.M.P.; experimental studies, S.K.; data acquisition, S.K., J.M.H., E.M.P.; data analysis, S.K., J.M.H., H.N.; statistical analysis, S.K.; manuscript preparation, S.K.; manuscript editing, J.M.H., E.M.P., P.B.; manuscript review, J.M.H., E.M.P. P.B., H.N., P.K.P.; manuscript final approval, all authors.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 

  1. Woo YR, Yoganathan AP. In vitro pulsatile flow velocity and turbulent shear stress measurements in the vicinity of mechanical aortic heart valve prostheses. Life Support Syst 1985; 3:283-312.[Medline]
  2. Hasenkam JM, Pedersen EM, Ostergaard JH, et al. Velocity fields and turbulent stresses downstream of biological and mechanical aortic valve prostheses implanted in pigs. Cardiovasc Res 1988; 22:472-483.[Medline]
  3. Giersiepen M, Wurzinger LJ, Opitz R, Reul H. Estimation of shear stress-related blood damage in heart valve prostheses: in vitro comparison of 25 aortic valves. Int J Artif Organs 1990; 13:300-306.[Medline]
  4. Sallam AM, Hwang NH. Human red blood cell hemolysis in a turbulent shear flow: contribution of Reynolds shear stresses. Biorheology 1984; 21:783-797.[Medline]
  5. Wurzinger LJ, Blasberg P, Schmid-Schonbein H. Towards a concept of thrombosis in accelerated flow: rheology, fluid dynamics, and biochemistry. Biorheology 1985; 22:437-450.[Medline]
  6. Nygaard H, Paulsen PK, Hasenkam JM, Kromann-Hansen O, Pedersen EM, Rovsing PE. Quantitation of the turbulent stress distribution downstream of normal, diseased and artificial aortic valves in humans. Eur J Cardiothorac Surg 1992; 6:609-617.[Abstract]
  7. Nygaard H, Paulsen PK, Hasenkam JM, Pedersen EM, Rovsing PE. Turbulent stresses downstream of three mechanical aortic valve prostheses in human beings. J Thorac Cardiovasc Surg 1994; 107:438-446.[Abstract/Free Full Text]
  8. Nygaard H, Yoganathan AP. In vivo and in vitro evaluation of turbulent stresses downstream of the St Jude Medical aortic valve. Med Biol Engin Comput 1996; 34:209-210.
  9. Kozerke S, Scheidegger MB, Pedersen E, Boesiger P. Heart motion adapted cine phase contrast flow measurements through the aortic valve. Magn Reson Med 1999; 42:970-978.[Medline]
  10. Wang Y, Rossman PJ, Grimm RC, Wilman AH, Riederer SJ, Ehman RL. 3D MR angiography of pulmonary arteries using real-time navigator gating and magnetization preparation. Magn Reson Med 1996; 36:579-587.[Medline]
  11. Weiger M, Bornert P, Proksa R, Schaffter T, Haase A. Motion-adapted gating based on k-space weighting for reduction of respiratory motion artifacts. Magn Reson Med 1997; 38:322-333.[Medline]
  12. Danias PG, McConnell MV, Khasgiwala VC, Chuang ML, Edelman RR, Manning WJ. Prospective navigator correction of image position for coronary MR angiography. Radiology 1997; 203:733-736.[Abstract/Free Full Text]
  13. Wang Y, Riederer SJ, Ehman RL. Respiratory motion of the heart: kinematics and the implications for the spatial resolution in coronary imaging. Magn Reson Med 1995; 33:713-719.[Medline]
  14. Bernstein MA, Zhou XJ, Polzin JA, et al. Concomitant gradient terms in phase contrast MR: analysis and correction. Magn Reson Med 1998; 39:300-308.[Medline]
  15. Walker PG, Cranney GB, Scheidegger MB, Waseleski G, Pohost GM, Yoganathan AP. Semiautomated method for noise reduction and background phase error correction in MR phase velocity data. J Magn Reson Imaging 1993; 3:521-530.[Medline]
  16. Kozerke S, Botnar R, Oyre S, Scheidegger MB, Boesiger P. Automatic vessel segmentation using active contours in cine phase contrast flow measurements. J Magn Reson Imaging 1999; 10:41-51.[Medline]
  17. Cape EG, Nanda NC, Yoganathan AP. Quantification of regurgitant flow through bileaflet heart valve prostheses: theoretical and in vitro studies. Ultrasound Med Biol 1993; 19:461-468.[Medline]
  18. Buonocore MH. Blood flow measurement using variable velocity encoding in the RR interval. Magn Reson Med 1993; 29:790-795.[Medline]
  19. Yoganathan AP. Cardiac valve prostheses. In: Bronzino JD, eds. The biomedical engineering handbook. Boca Raton, Fla: CRC, 1995; 1847-1870.
  20. Fontaine AA, Heinrich RS, Walker PG, et al. Comparison of magnetic resonance imaging and laser Doppler anemometry velocity measurements downstream of replacement heart valves: implications for in vivo assessment of prosthetic valve function. J Heart Valve Dis 1996; 5:66-73.[Medline]
  21. Giersiepen M, Krause U, Knott E, Reul H, Rau G. Velocity and shear stress distribution downstream of mechanical heart valves in pulsatile flow. Int J Artif Organs 1989; 12:261-269.[Medline]
  22. Houlind K, Eschen O, Pedersen EM, Jensen T, Hasenkam JM, Paulsen PK. Magnetic resonance imaging of blood velocity distribution around St Jude Medical aortic valves in patients. J Heart Valve Dis 1996; 5:511-517.[Medline]
  23. Kilner PJ, Firmin DN, Rees RS, et al. Valve and great vessel stenosis: assessment with MR jet velocity mapping. Radiology 1991; 178:229-235.[Abstract/Free Full Text]
  24. Oshinski JN, Ku DN, Pettigrew RI. Turbulent fluctuation velocity: the most significant determinant of signal loss in stenotic vessels. Magn Reson Med 1995; 33:193-199.[Medline]
  25. Stahlberg F, Sondergaard L, Thomsen C, Henriksen O. Quantification of complex flow using MR phase imaging: a study of parameters influencing the phase/velocity relation. Magn Reson Imaging 1992; 10:13-23.[Medline]
  26. Walker PG, Pedersen EM, Oyre S, et al. Magnetic resonance velocity imaging: a new method for prosthetic heart valve study. J Heart Valve Dis 1995; 4:296-307.[Medline]



This article has been cited by other articles:


Home page
 Eur Heart JHome page
D. J. Pennell, U. P. Sechtem, C. B. Higgins, W. J. Manning, G. M. Pohost, F. E. Rademakers, A. C. van Rossum, L. J. Shaw, and E. K. Yucel
Clinical indications for cardiovascular magnetic resonance (CMR): Consensus Panel report
Eur. Heart J., November 1, 2004; 25(21): 1940 - 1965.
[Full Text] [PDF]

Home page
 CirculationHome page
M. H. Yacoub and L. H. Cohn
Novel Approaches to Cardiac Valve Repair: From Structure to Function: Part II
Circulation, March 9, 2004; 109(9): 1064 - 1072.
[Full Text] [PDF]

Home page
 Am. J. Roentgenol.Home page
S. J. Karlik
Visualizing Radiologic Data
Am. J. Roentgenol., March 1, 2003; 180(3): 607 - 619.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kozerke, S.
Right arrow Articles by Boesiger, P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kozerke, S.
Right arrow Articles by Boesiger, P.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE