For the majority of CE MRA applications, a single dose of standard, extracellular GBCA (0.1 mmol/kg) is sufficient. Contrast is injected intravenously (IV) through a catheter (preferably through an 18–22 gauge catheter in an antecubital vein) at a variable rate (i.e., 0.5–4.0 mL/s), depending upon the size of the catheter, the quality of the patient’s veins, and the length of the acquisition. However, caution should be used when injecting high relaxivity GBCA because when injecting at a high rate, the concentration of GBCA can be high enough to cause T2-star artifact, reducing the overall signal intensity within the vasculature during the first pass (Fig. 4.4). As indicated previously, the image acquisition is timed such that the center of k-space is acquired during maximum enhancement of the vessels of interest, because the center of k-space contributes to the overall signal and enhancement of the image, while the peripheral of k-space contributes to the detail or sharpness of the images (Fig. 4.5). The time-to-peak enhancement of the vessels of interest can be done using a test bolus injection to determine the transit time of a contrast injection from the peripheral IV to the area being imaged [1]. Alternatively, with centric k-space ordering, where the center of k-space is acquired at the beginning of the scan, real-time bolus tracking can be used so that the acquisition can be started precisely when the contrast bolus arrives in the vessels of interest [2]. When the acquisition is not timed appropriately to the passage of the contrast bolus, a variety of different artifacts can occur, including decreased signal-to-noise ratio (SNR), truncation artifact, and edge blurring [3]. Ideally, the contrast bolus injection duration should be equal to the length of acquisition (Fig. 4.6). A benefit of CE MRA, relative to CT angiography (CTA), is that if the image quality is insufficient to make the diagnosis, the acquisition can be repeated with a second injection (Fig. 4.7) with little to no harm to most patients because GBCA is less nephrotoxic and MRA does not require the use of ionizing radiation.

At our institution, we dilute a single dose of GBCA [i.e., 0.1 mmol/kg of gadobenate dimeglumine (Multihance, Bracco Diagnostics Inc. Princeton, NJ, USA) or 0.03 mmol/kg of gadofosveset trisodium (Vasovist, ABLAVAR, EPIX Pharmaceuticals, Inc. Cambridge, MA, USA)] with normal saline to achieve a total volume of 30 mL to inject in two phases. We inject 5 mL of diluted contrast at 2 mL/s, resulting in a 2.5 s contrast bolus which we expect to prolong to 3–4 s by the time it gets to the pulmonary circulation, for the time-resolved CE MRA (which has a “temporal resolution” of 3–4 s). We then inject 25 mL at 1.5 mL/s, resulting in a 16–17 s contrast bolus which we expect to prolong to ~20 s by the time it reaches the pulmonary circulation, for the high-resolution 3D CE MRA, which is about 17–20 s long.

Conventional three-dimensional (3D) CE MRA can be performed without or with ECG triggering. ECG-triggered CE MRA (Fig. 4.8) is preferred for imaging patients with diseases of the ascending aorta to minimize artifact from cardiac motion [4]. Furthermore, when making measurements of the aorta, or other vessels, it is critical to make double-oblique measurements perpendicular to the vessel long axis (Fig. 4.8b) rather than simply using axial images to avoid overestimating the true size of the vessel [5]. Equally important is to use a consistent method of defining how the measurements are being made, especially when using MRA-derived measurements in comparison with prognostic data obtained using other imaging modalities [6]. A limitation of ECG-triggered CE MRA techniques is that they require more time to acquire and is therefore potentially problematic in patients who have difficulty holding their breath. Acquisition times for ECG-triggered CE MRA can be decreased by limiting the coverage to a thin slab over the aorta or by using parallel imaging.

A major advance in improving the speed of MR image acquisition has been the use of parallel imaging [7–9]. This has been enabled through changes in image reconstruction from undersampled k-space data and use of multichannel receiver coils [10]. With these multichannel receiver coils, each individual coil is used to simultaneously receive the MR signal. The use of parallel imaging to accelerate data acquisition can be used to shorten the scan times to image the same volume [11] or to increase anatomical coverage during the same acquisition time [12]. Using advanced parallel imaging techniques, it is now possible to scan the entire thoracic vasculature within 15–20 s with high, isotropic spatial resolution (Fig. 4.9).

A major motivation for scanning at higher field strengths is that the SNR is proportional to the field strength. Therefore, at 3.0T SNR is approximately twice that at 1.5T. The advantage in SNR at 3.0T, relative to 1.5T, can be used to increase spatial resolution, decrease scan time, or a combination of these while obtaining images of relatively similar or better quality as those at 1.5T [13]. In addition, because of the differences in T1 effects of GBCA at 3.0T, there is actually greater vascular contrast, relative to surround soft tissues, at 3.0T compared to 1.5T. Therefore, CE MRA at 3.0T can be performed with smaller doses of GBCA [14–17]. In fact, recent data suggests smaller volumes of GBCA result in higher-quality arterial phase MRA than with larger volumes of GBCA [18].

MRI at 3.0T results in significant challenges and limitations due to the stronger magnetic field that have not yet been fully overcome. At 3.0T, the resonance frequency of water protons is double that at 1.5T. Because the frequency and wavelength are inversely related, the radiofrequency (RF) wavelength of water protons at 3.0T is half that at 1.5T. This shortened wavelength can span the dimensions of the field of view for body imaging, especially in those with a large body habitus [19]. When RF waves overlap within the imaging field, destructive or constructive interference results in areas of darkening or brightening, respectively. RF field homogeneity can be improved by (A) using advanced coil designs, such as multi-coil transmit coils, which can help to suppress eddy currents [20] or (B) implementing new pulse sequences, such as 3D RF pulses [21].

Another potential limitation with performing MRA at 3.0T is that the specific absorption rate (SAR) increases at higher field strengths due to constructive and destructive interference due to RF field inhomogeneity. RF pulses transfer energy to protons within the patient and ultimately generate heat as energy is released. To avoid overheating within the patient, which can have detrimental physiologic effects, the SAR is carefully monitored during scanning. The FDA currently limits the total body heating to 4 W/kg during a 15-min period [22, 23]. The SAR estimates the energy transfer to the tissue by the RF pulse and is proportional to the square of the resonance frequency. Therefore, in going from 1.5 to 3.0T, where the resonance frequency is doubled, the SAR increases fourfold for the same sequence, assuming all parameters are equal [24]. As a result, the pulse sequences, acquisition techniques, and hardware designs are modified to work within the allowable constraints of the increased SAR. Parallel imaging helps alleviate this problem by decreasing scan time and the number of RF pulses necessary to acquire an image.

Most CE MRA, as with most CE MRI, is performed with GBCA that are not purely intravascular. Rather, most GBCAs do distribute throughout the extracellular space. For vascular imaging, however, this can be problematic in achieving the greatest contrast between the vessels and their surrounding soft tissues. Intravascular GBCAs are, therefore, another recent advance that has enabled improvement in CE MRA spatial and contrast resolution. Gadofosveset trisodium, the currently only FDA-approved blood-pool GBCA, is a protein-binding intravascular contrast agent that is approved for the evaluation of patients with aortoiliac occlusive disease [25]. Protein binding is what prolongs the intravascular half-life of gadofosveset trisodium relative to other GBCAs. Another factor contributing to its advantage for CE MRA is that gadofosveset trisodium has a higher relaxivity than other GBCAs [26]. Due to its intravascular distribution and greater contrast enhancement, smaller volumes of gadofosveset trisodium are necessary than other GBCAs to achieve the same image quality [27, 28]. Furthermore, the prolonged intravascular half-life extends the amount of time available for scanning during vascular enhancement. Therefore, very high spatial and contrast resolution CE MRA images are feasible during the steady state (Fig. 4.10) which may be beneficial in detecting additional findings not present on first-pass CE MRA [29, 30]. In addition, the prolonged intravascular enhancement with blood-pool contrast agents implies that high-quality imaging can be performed during free breathing using respiratory navigators or bellows for triggering to the same part of the respiratory cycle. This is particularly useful in patients that have difficulty holding their breath [29]. CE MRA during the steady state can also be particularly helpful in patients with cavopulmonary shunts because during the first pass, unenhanced blood from the inferior vena cava will dilute the contrast entering the pulmonary circulation from the superior vena cava.

For cardiothoracic NCE MRA, 3D bSSFP sequences (Figs. 4.14 and 4.15) are most often used because of (a) inherently high signal of blood compared to surround soft tissues and (b) relative blood flow independence [45]. Image contrast of bSSFP sequences are determined by the ratio of T2 to T1 of the tissues, which results in the very bright blood-pool signal. Standard 3D bSSFP techniques will produce images that have bright arterial and venous signal, which is typically not a problem for assessment of the larger, central vessels like the aorta, main pulmonary arteries, or pulmonary veins [46–49]. 3D bSSFP MRA is also widely used for coronary MR angiography [50, 51]. These sequences are typically performed using respiratory triggering or navigation to minimize respiratory motion artifacts and ECG triggering to minimize cardiac motion artifacts. However, a limitation of the bSSFP sequence is its susceptibility to field inhomogeneities which occur at the lung–soft tissue interfaces and around metallic implants or devices.

For quantitative measurements of vessel size in the thorax, 3D bSSFP MRA performs very well in comparison with CE MRA. In a comparison of 3D bSSFP and CE MRA, François et al. [46

Stay updated, free articles. Join our Telegram channel

Join