Chapter 8 Magnetic resonance angiography
Introduction
Recent innovations in CMR hardware, such as stronger gradients and faster gradient switching, and software, such as newer pulse sequences and ultrafast MRA techniques, continue to ensure that it is the noninvasive imaging modality of choice in many diseases affecting the great vessels.
MRA techniques
Contrast-enhanced MRA
The basic pulse sequence for ultrafast CE-MRA is a 3D GE acquisition which is usually designed to acquire data encoding for contrast first—centric-ordering. The use of gadolinium contrast agent shortens the T1 value of blood so that it appears bright irrespective of flow patterns or velocities (Figure 8.1).
Time-of-flight (2D or 3D)
This is an older GE technique in which flow signal is enhanced by inflow effects. Stationary tissue within a region of interest is saturated by rapid application of RF pulses. This results in fresh blood flowing into this plane being non-saturated and therefore appearing bright. Unwanted venous signal can be eliminated by application of specific saturation bands. Time-of-flight methods are sensitive to flow velocity; for example, slow flow near the vessel wall appears less bright with this technique than faster flow in the middle of the vessel. Importantly, the 2D time-of-flight sequence has the disadvantage of overestimating vessel stenosis due to a phenomenon known as phase dispersion. This effect is less pronounced using 3D time-of-flight imaging.
Post-processing methods
Post-processing algorithms are used to create 3D images. However, they must always be interpreted together with the source images, the ‘raw data’, to avoid misinterpretation from post-processing induced artefacts (Figure 8.2).
Maximal intensity projection
In this ray tracing post-processing algorithm, the desired viewing plane is specified first. Then the maximum intensity encountered in parallel rays along this viewing plane is assigned to the displayed pixel. MIP reconstructions are applicable to both time-of-flight and CE-MRA data and have the tendency to overestimate the degree of vessel stenosis (Figure 8.2b).
Volume rendering technique and shaded surface display
This is a ray casting algorithm which selects visible voxels by tracing rays from an instantaneous viewing position. The surfaces are identified by a threshold technique and the resulting SSD provides 3D appreciation of the vessel. Resulting images can also overestimate the degree of vessel stenosis (Figure 8.2c).
Specific Sites
Carotid arteries
The most commonly used sequences for carotid MRA are time-of-flight and CE-MRA (Figure 8.1). For imaging of the aortic arch and cervical arteries, CE-MRA is the only option. Both 3D TOF and CE-MRA are as accurate as conventional X-ray angiography in the measurement of internal carotid artery stenosis, and because of the small but significant risk of stroke with the invasive technique (0.5–1.0%), CMR is recommended as the optimal method of evaluating carotid artery disease. Data interpretation requires careful evaluation of the raw data to avoid overestimation of stenosis severity. MIP projections will aggravate signal loss and cause vessels to appear narrower because the algorithm selects brightest intensities both within the vessel and in the background. Such overestimation with MIPs is more of a problem in 3D time-of-flight than with CE-MRA, in which background intensities are suppressed.