The bright blood phenomenon on spoiled GE cine MRI techniques is based on the laminar inflow of fully magnetized spins. Intravoxel dephasing of the proton spins, due to flow disturbance (turbulence, acceleration) leads to loss in signal. This principle can be applied to assess valvular regurgitation. Imaging over multiple frames throughout the cardiac cycle enables visual assessment of flow disturbance caused by the regurgitation. Depending on the severity a dark signal is visible extending from the diseased valve into the receiving chamber, either pencil-beam-like or fan-like, during a variable time period of the cardiac cycle (Fig. 14.3). Using a combination of contiguous cine MRI slices, the direction of the regurgitant jet can be well visualized. Because of the continuously changing loading conditions, the jet direction may vary over time (Fig. 14.4). Similar to cardiac catheterization and cardiac ultrasound, the extent of signal loss in the receiving chamber can be used to grade the severity of regurgitation. This semi-quantitative approach, however, suffers from similar shortcomings as the approach used by the other conventional imaging techniques. The signal loss is highly dependent on the imaging parameters (e.g., echo time), and several studies have shown poor reproducibility of the technique between centers (5,6). The severity is underestimated when the jet impinges on the cardiac chamber wall, and this method is unable to separate turbulent volumes when dual-valve disease exists. Nevertheless, cine MRI remains an interesting approach to rapidly evaluate the presence or absence of valvular regurgitation. In case of valve regurgitation, other MRI techniques can be applied to quantify the valvular regurgitation severity. The new standard bright blood sequence that is now used to study cardiac function, that is, b-SSFP technique, is designed to be relatively flow-insensitive, yielding a better endocardial/blood pool definition than spoiled GE techniques. As a consequence, however, this results in reduced visualization of flow disturbance secondary to valvular regurgitation. In particular, mild regurgitation may not be seen, or less apparent on b-SSFP cine imaging. Krombach et al. found a good correlation between b-SSFP cine imaging and color Doppler and cardiac catheterization in patients with known valvular dysfunction (7). Although the jet phenomenon was slightly
more pronounced on spoiled GE cine MRI, the b-SSFP cine technique resulted in a significantly better image quality.

Noninvasive quantification of blood flow can be achieved with velocity mapping (also called velocity-encoded MRI or phase-contrast velocity mapping). For an extensive description of the velocity mapping technique, we refer to Chapter 5 on MR blood flow measurements. In brief, spins moving along a magnetic field gradient are exposed to a different magnetic field than the stationary, nonmoving spins. As a result, these moving spins experience a different degree of dephasing than stationary spins, which is proportional to the distance they move in time (and thus the velocity). This phenomenon can be exploited to measure flow velocities and to extract flow volumes that pass through valves. Velocity mapping can be applied in any direction, though for flow quantification of valvular lesions, through-plane imaging is used. Stationary spins are depicted as midgray, while increasing velocities in either direction are shown in increasing grades of white or black (Fig. 14.5). Precise determination of the velocity window as close as possible to the expected peak velocity is important to avoid aliasing while maintaining the highest sensitivity for flow measurements. Measurement of the velocity of all pixels in a region of interest enables calculation of the instantaneous flow volume at any point of the cardiac cycle (Fig. 14.5). Integrating the instantaneous flow volumes for all frames throughout the cardiac cycle enables calculation of the flow volume per heartbeat. Velocity mapping has been well validated in vitro and in vivo and can be considered the best available in vivo technique for flow measurements (8,9). In normal individuals, the output of the left ventricle (LV) and the right ventricle (RV) obtained by quantification of the through-plane flow in the proximal ascending aorta and pulmonary trunk has a 1:1 ratio. Differences in ratio can be caused by left-toright or right-to-left shunt or by valvular dysfunction.

To obtain reliable measurements of flow velocities and volumes through valves with velocity mapping, several technical issues need to be addressed. To improve the accuracy of measurements, it can be necessary to perform a baseline correction using a stationary phantom (10,11). To keep phase errors induced by eddy currents and Maxwell terms as low as possible, the area of interest should be as close to the center of the main magnetic field (B₀) as possible. The imaging plane should be perpendicular to the valve. This ensures that the velocity vectors of most voxels are perpendicular to the imaging plane. This can be achieved by a thorough knowledge of cardiac anatomy and cardiac image-plane positioning. The best position for flow measurements is not exactly at the level of the valve, but just proximal or distal to the valve annulus. Correction for motion of the valve annulus through the imaging plane during the cardiac cycle can be obtained with slice-tracking technique (12). This promising technique is, however, not yet available for clinical use.

Cardiac MRI is currently considered the reference technique for in vivo measurement of ventricular volumes and myocardial mass. For this task, breath-hold cine MRI, using the b-SSFP sequence, is used. Based on the volumetric approach (Simpson’s rule), a set of cardiac short-axis cuts covering the length of the ventricles enables to determine ventricular volumes and myocardial mass (see Chapter 6). In normal conditions, there is a 1:1 relationship between LV and RV stroke volumes. In the absence of a cardiac shunt, any discrepancy between stroke volumes in a patient with regurgitation will identify the regurgitant volume. Unfortunately, only patients with a single regurgitant valve can be assessed. However, combination of velocity mapping with ventricular stroke volume measurements allows for the assessment and quantification of multivalvular disease. Moreover, wine MRI allows for an excellent appreciation of the consequences of volume overload on ventricular/atrial remodeling.

Use of signal loss extending from a stenotic valve into the distal chamber or vessel is helpful for identification of a
stenotic valve (Fig. 14.6). The degree of signal loss, however, is not only dependent on the degree of stenosis but also on the echo time used. For shorter echo times, less spin dephasing occurs. Kilner et al. have shown that in more severe stenoses shorter echo times need to be used to prevent signal loss in the images (13). Similar as for the qualitative assessment of valvular regurgitation, signal loss is less marked on the b-SSFP cine MR images, and therefore it is more difficult to get a feel for the peak velocity with b-SSFP sequences than with conventional spoiled GE cine imaging.

Valve area or valve orifice can be estimated indirectly using the peak systolic transvalvular gradient (cardiac catheterization) and the continuity equation (echo Doppler), or can be calculated by direct planimetry (transesophageal echocardiography and MRI). MR images perpendicular through the stenotic valve, during maximal opening, allows a planimetric quantification of the valve orifice (14,15). It is currently unclear whether b-SSFP sequences with short echo times are better than the conventional spoiled GE techniques to obtain reliable estimates of the valve stenosis (16). While spoiled GE cine MRI, especially when using longer echo times, is sensitive to spin-dephasing artifacts, the b-SSFP technique may have different sources of artifacts, including sensitivity to areas of magnetic field disturbance (e.g., calcifications, tissue interfaces) and sometimes severe artifacts in areas of highly complex flow (16). Using data from velocity-encoded flow curves, the continuity equation (similar as used in echo Doppler) can be used to calculate valve area (17) (see Aortic Valve).

Information from the phase-contrast velocity map, that is, peak velocity across the valve, can be used to obtain an estimate of the gradient across the valve, using the modified Bernoulli equation:

where P is the pressure drop across the stenosis (in millimeters of mercury) and V is velocity (m/s). The technique is comparable with Doppler echo measurements and has an in vitro accuracy of 4% (18). The main advantage of the technique over cardiac ultrasound is that the velocity jet can be aligned easily in any direction without the limitation of acoustic windows or the interference of (peri)valvular calcifications.

The flow through a stenotic valve can be imaged inplane (velocity jet parallel to the imaging plane) or throughplane (velocity jet perpendicular to the imaging plane). The in-plane method visualizes the entire jet and shows the point within the jet of peak velocity. However, peak velocity may not always be accurately depicted, for instance, when the jet is not aligned in a single two-dimensional (2D) plane, or in case of tight narrow jets there may be partial volume averaging and motion within the imaging slice. For throughplane motion, the jet will always pass through the imaging plane, but, as only part of the stenotic jet is sampled, the peak velocity may not be measured. Since both strategies have advantages and disadvantages, it is the best to use a combination of the two strategies with initial definition of the jet in-plane and quantification with through-plane imaging at the site of maximum velocity on the in-plane image.

Use of an appropriate velocity window is essential to maintain sensitivity and accuracy of measurements, while avoiding aliasing. Thus, the velocity window should be adapted to the expected peak velocity within the vessel or the valve of interest. Practically, always check the phase images and repeat the sequence with an adapted velocity window if aliasing is present.