Since the introduction in the early 1980s of magnetic resonance imaging (MRI) in clinical practice, contrast agents (CAs) have been part of MR studies enhancing the capabilities of this diagnostic technique. CAs enhance the contrast between different types of tissues. Tissue contrast depends largely on proton concentration and longitudinal (T1) and transversal (T2) relaxation times inherent to each tissue. External CAs act by favouring hydrogen nuclei relaxation (decreasing T1) or by increasing asynchronism and dispersion in hydrogen nuclei relaxation within the voxel (decreasing T2, T2*). The final effect of a given CA in an MR signal will depend on the CA used (magnetic element, concentration, structure, size), tissue studied (water concentration, mobility and distribution, voxel situation regarding the CA influence area) and technique (sequence, time).1 Most CAs have two components: an active principle (metallic ion) and a quelate substance that decreases toxicity and determines pharmacokinetics and biodisponibility. Metals and molecules containing unpaired electrons are used as active principles. Based on their behaviour when placed on a magnetic field, CAs are further classified as paramagnetic (gadolinium (Gd), dysprosium and manganese) or supermagnetic (iron oxide). Classification of CAs is based on their distribution (extracellular, intravascular), their effect on the signal intensity of the tissue or specificity on target tissue (see Table 1).
Cardiac Applications of Contrast Agents
Over the last five years, cardiac MRI has emerged as a non-invasive imaging technique with multiple applications for the evaluation and management of patients with cardiovascular diseases. Despite the useful information obtained by anatomical and functional sequences without the use of CA, most clinical cardiac MR studies require its administration for angiography, coronary artery perfusion and viability assessment or evaluation of cardiac masses.2 MR angiography (MRA) has proved useful not only for the evaluation of the morphology and dimensions of the thoracic aorta or pulmonary arteries, but also for the evaluation of the anatomy of pulmonary veins. Candidates for atrial fibrillation ablation procedures or patients with suspected anomalous pulmonary vein drainage are best suited for MRA. Usually, MRA is performed with standard extracellular CA (ECA). Angiography during first pass provides enhancement of the vessel of interest (see Figure 1). However, due to the rapid extravasation of ECA, there is a marked decrease in signal intensity in such a way that MRA during steady state is not useful. Mandatory precise bolus timing and patient co-operation are useful. For this reason, new CAs with predominant intravascular distributions and longer-lasting effects may play a role in MRA and in coronary angiography, where long acquisition times are needed.3
On the other hand, perfusion and a delayed enhancement pattern after CA administration are part of everyday clinically routine cardiac MR examinations (see Figure 2). ECAs are used in clinical practice to assess perfusion. However, with these CAs there are substantial variations from patient to patient, and perfusion defects are visible on just a few images as a transient phenomenon. Blood pool CAs that last for longer in the intravascular space may be useful for a more accurate and reproducible myocardial perfusion assessment. Less technically demanding is myocardial viability (see Figure 3), where ECAs have proved useful in detecting myocardial scarring due to their extravascular distribution and delayed clearance from the infarcted myocardium.4
Blood Pool Contrast Agents
ECAs rapidly reach an equilibrium during the first pass between the intravascular and interstitial spaces due to their low molecular weight, which enables their diffusion through the capillary wall. Active research has been made in this field in order to obtain intravascular CAs, the so-called blood pool agents (BPAs), which are macromolecular or non-diffusible agents with longer plasma half-life and low tissular extraction.5,6 Different approaches have been tested, such as ultra-small iron oxide particles, paramagnetic Gd-based macromolecules and Gd-based small molecules with reversible protein binding. Many different types of compounds are being tested, but are not yet approved for clinical use (see Table 2).
There are two classes of supermagnetic particles: super-paramagnetic iron oxide particles (SPIOS) and ultrasmall super-paramagnetic iron oxide particles (USPIO). Due to their smaller size (near 20nm) and molecular composition, USPIOS remain in the intravascular space for a prolonged amount of time. They have a strong T1 and T2 shortening effect depending on other parameters on the dose used. Several USPIOS have been investigated in humans for different applications, but none has been approved for clinical use.7
Iron Oxide Blood Pool Agents
Iron-based BPAs are coated SPIOS that are insoluble in water and are without magnetic properties outside an external magnetic field. They are characterised by a large magnetic moment in the presence of a static external magnetic field. They have a specific uptake by the monocyte–macrophage system, which explains the fact that if they are not entirely captured by the liver and spleen they become MRI markers of inflammatory and degenerative disorder associated with a high macrophage phagocytic activity.7
Gadolinium-based Blood Pool Agents
At present there is only one Gd-based blood pool agent approved for clinical use: gadofosfosveset trisodium (Vasovist®, Schering, Berlin). It is a non-specific CA of low molecular weight (975, 88 Dalton (Da)) that is retained in the blood pool due to a non-covalent binding to serum albumin. This reversible binding slows down the tumbling of the molecule and increases the paramagnetic effectiveness of Gd, resulting in a decrease in dose needed. These properties have also proved to be particularly useful and safe at the recommended dose for contrast-enhanced 3D MRA. With prolonged T1 shortening of blood, MR angiographic resolution and anatomical coverage are improved because a longer imaging window is available after a single injection. Compared with other currently used Gd ECAs (e.g. gadopentetate dimeglumine), the plasma relaxivity at 0.5 Tesla (T) of gadofosfosveset trisodium is six to 10 times greater.5,8
Gadocoletic acid, B-22956/1 (Bracco Imaging SpA, Milan), is a relatively low-molecular-weight Gd chelate with high affinity for serum proteins that provides high T1 relaxivity values. In humans, T1 values of 35 and 55 milliseconds (ms) have been found at five and 30 minutes after administration, respectively, of 0.1mmol/kg of B-22956 at 1.5T.9 Other Gd-based BPAs remain in the intravascular space due to their molecular weight. Examples such as Gadomer 17 and P792, though not approved for clinical use, are the most extensively studied. Gadomer 17 (SH L 643 A, Schering, Berlin) is a water-soluble polymeric dendrimeric complex chelating agent with 24 Gd atoms. Its intermediate molecular size (17.45kDa) is responsible for its predominantly intravascular distribution without significant diffusion into the interstitial space.10 Similar to other BPAs, it has a higher T1 relaxivity than other Gd-ECAs, making it suitable for MRA. Several trials have addressed the utility of this agent in cardiac MR with promising results. P792 (Vistarem, Laboratoire Guebert, Aulnay-sous-Bois) is a Gd chelate with a molecular weight of 6.48kDa and is 5nm in size, with minimal interstitial distribution once administered.11 Due to an unrestricted glomerular filtration rate it is known as a BPA with rapid elimination. However, its plasma half-life is greater than that of classic ECAs, with a concentration five minutes after administration four times higher than that of ECAs and a plasma half-life of 41 minutes.
Blood Pool Agents in Cardiac Magnetic Resonance
MR aortoiliac angiography is the application for vasovist use in clinical practice.8,12 In 174 patients with suspected peripheral vascular disease, vasovist 0.03mmol/kg proved to be safe and showed a significant improvement in sensitivity, specificity and accuracy for diagnosis of clinically significant (>50%) stenosis compared with unenhanced MR.8 No studies so far have addressed the utility of vasovist in aortic root or upper extremities and head and neck angiography.
Regarding cardiac MR, there is only one study of vasovist in humans. In 24 healthy volunteers, vasovist was administrated at the recommended dose to evaluate signal-to-noise ratio, contrast-to-noise ratio and the image quality with two types of 3D sequence for evaluating contrast-enhanced breath-hold coronary angiography.13 The application of vasovist to myocardial perfusion or delayed enhancement has not been tested in humans. However, data from animal studies are available. One porcine model demonstrated its utility in estimating myocardial blood flow, analysing the rate of myocardial signal enhancement during the first-pass perfusion (upslope). The estimated myocardial blood flow was in agreement with the invasive measurement. Detection of ischaemia has also been noted. In 19 pigs with inducible left anterior descending artery stenosis, ischaemia was detected during first-pass perfusion at peak dypiridamol stress and was comparable to both nuclear imaging and invasive measurements with labelled microspheres. A greater than 40% decrease in blood flow was caused in eight animals, and in 92% of them hypoenhancement on first-pass perfusion was detected. Agreement with nuclear imaging was 85%. The intravascular nature of vasovist seems to facilitate detection of myocardial perfusion defects during first-pass perfusion.14
Gadocoletic acid (B 22956) has been tested in both healthy volunteers and patients for the evaluation of coronary arteries. In six healthy adults, 3D coronary angiography with gadocoletic acid at a dose of 0.075mmol/kg proved to be superior to baseline 3D coronary angiography with T2 preparation without exogenous contrast when visualising the left coronary tree. Both the signal-to-noise ratio and contrast-to-noise ratio increased significantly.15
The same group confirmed the same results in 12 healthy volunteers after five minutes and 45 minutes of 0.075mmol/kg B 22956 administration. Both objective and subjective parameters of image quality improved compared with non-contrast images. The increase in signal-to-noise ratio, contrast-to-noise ratio and sharpness highlights the possible clinical use of this approach for the evaluation of coronary artery anatomy with high-resolution MRA.16 Based on these promising preliminary results, the utility of contrast MR coronary angiography was tested in 18 patients with chest pain and suspected coronary disease who were referred for coronary invasive angiography. Three different doses of gadocoletic acid were evaluated (0.05, 0.075 and 0.10mmol/kg) and imaging of the right and left coronary arteries was obtained. The best diagnostic accuracy was obtained with a 0.075mmol/kg dose (94%), greater than that obtained in images without contrast, with a sensitivity of 86% and specificity of 95%. Evaluable segments in the post-contrast images were above 80% with all three doses compared with 67% in the pre-contrast images.17 These results confirm the feasibility of MR coronary angiography with BPAs. Further studies should be undertaken to confirm the reproducibility of these results and clinical applications.
No viability or myocardial perfusion studies have been reported so far with these BPAs in humans. A small trial has been carried out in which perfusion was evaluated in six pigs with induced coronary artery stenosis. Gadocoletic acid for perfusion was administered at a dose of 0.015mmol/kg at 3ml/s flow rate followed by a 10ml saline flush at the same rate. Differences in increased signal intensity between normal and ischaemic myocardium were detected. Upslope was significantly lower in segments with ≥75% coronary stenosis compared with normal myocardium (no stenosis or <50%), with a decrease in the ratio between them.18 Further studies are needed to evaluate the optimum contrast dose, reproducibility and utility in humans.
The main focus research is myocardial perfusion and coronary angiography. Two papers have addressed the utility of this contrast agent in the evaluation of myocardial perfusion in animal models. In the first,19 Gadomer 17 0.05mmol/kg at 5ml/s was compared with ECAs and single photon emission computed tomography (SPECT) in the detection of ischaemia compared with labelled microspheres in 12 pigs with induced coronary artery stenosis/occlusion.
All three methods detected the presence of myocardial perfusion defects after dypiridamol stress test. Correlation with SPECT and blood-flow reduction measured by labelled microspheres was suitable for both contrast MR agents. However, Gadomer 17 allowed differentiation of ischaemic from non-ischaemic myocardium until 20 minutes after contrast material injection by analysing the percentage increase in signal intensity in ischaemic and remote myocardium. In contrast, differentiation of ischaemic from non-ischaemic myocardium was possible for only 55 seconds after injection of gadopentetate dimeglumine. This preliminary report favours the use of gadomer 17 in myocardial perfusion studies, as it detects myocardial perfusion defects accurately, avoids two contrast doses in stress test studies and may allow a better image quality by planning sequences with higher spatial resolution not limited by temporal resolution due to its longer plasma half-life.
In the second study, gadomer 17 proved to be useful in detecting microvascular obstruction in rats 24 hours after transient coronary artery occlusion.10 Based on clinical experience in animal models,20 gadomer 17 was tested for coronary angiography in 12 healthy volunteers. Non-contrast inversion–recovery 3D coronary angiography was obtained in all volunteers, followed by contrast-enhanced coronary angiography. The contrast-to-noise ratio and image quality improved significantly in contrast-enhanced images, lasting at least 30 minutes after contrast injection. No contrast side effects were noted acutely or 24 hours after exposure to gadomer 17.
As with gadomer 17, this macromolecular BPA is the focus of active research not only in coronary angiography and myocardial perfusion but also in myocardial viability detection. Dewey and colleagues21 compared the utility of P792 0.013mmol/kg with conventional Gd-dotarem (DOTA) in the evaluation of myocardial perfusion and viability in five pigs with induced acute myocardial infarction. MR images in all cases were obtained 48–72 hours after the instrumentation. Myocardial perfusion defects correlated well between both CAs on first-pass perfusion images, but only hypoperfused areas were detected in the equilibrium phase – 10 minutes after injection – with P792. However, there was a higher myocardium–blood ratio of the percentage signal intensity change peak during the first pass in ECAs, reflecting its extravascular distribution and rapid clearance of blood pool.
Infarct size determined by delayed enhancement with P792 showed good agreement compared not only with ECAs but also with the pathological results. However, inversion times to null normal myocardium were significantly longer, as were times to onset of enhancement of the infarcted myocardium (25–30 minutes) compared with Gd-DOTA. Recently, similar results were published.11 Myocardial perfusion and infarct size were analysed in six mini-pigs using a double dose (0.026mmol/kg) injected at a rate of 3ml/s and compared with ECAs. Myocardial perfusion defects could be evaluated with high spatial resolution on first-pass perfusion and after 30 minutes of contrast injection. Infarct size was accurately detected with both contrast media at eight and 31 minutes for Gd-DOTA and P792, respectively, and correlated well with pathological data. Regarding MR coronary angiography, preliminary results on animal models comparing ECA and P792 (dose of 13 micromol Gd/kg) showed an increased blood-to-myocardium signal-difference- to-noise ratio and better-quality imaging in evaluating coronary arteries.22,23 Further studies are needed to confirm these results and establish its possible clinical applications.
Ultrasmall Superparamagnetic Iron Oxide Particles
USPIOS can be used withT1, T2 and T2*. With high doses the effects on T2 and T2* dominate, whereas low to moderate doses shorten T1. Based on these properties and the longer plasma half-life (45 minutes), research has focused on MR coronary angiography, first-pass perfusion and cardiac transplant rejection evaluation. Most trials use feruglose (CLARISCAN™, NC100150 Injection, Nycomed Amersham Imaging, Germany), which has proved safe at different doses, as a CA. First reports from 199924 and 200125 demonstrated the utility of this agent in coronary anatomy evaluation in five patients with coronary artery disease. It improved image quality by increasing the length of vessel visualisation with breath-hold 3D fast low-angle shot sequence. More recently, at a dose of 1–5mg/Fe/kg, and using a free-breathing, navigator-assisted, high-resolution 3D MR sequence, feruglose improved the contrast-to-noise ratio in distal coronary arteries at a dose of 3mg/Fe/kg. Visible vessel length was also improved compared with non-enhanced MR angiography, with better coronary stenosis identification.26 Preliminary reports on pigs have confirmed its potential utility with other MRA sequences (spiral excitation pulses), with increase in contrast-to-noise and signal-to-noise ratio at low doses.27
There are limited published data regarding myocardial perfusion with USPIOS. Feruglose at 3mg/Fe/kg during first-pass perfusion was used in eight patients with coronary artery disease with a T2-weighted turbo spin echo sequence to quantify signal intensity decrease over time in normal myocardium.28 Using the same agent at doses of 2mg/Fe/kg and 5mg/Fe/kg during steady state at five minutes after administration, hypoperfuse myocardium could be accurately depicted in a porcine model. Normal myocardium showed a significant decrease in signal intensity in T2-weighted spin echo images with the higher-contrast dose and compared well with fluorescein images. Using T1-weighted spin echo images, microvascular damage could be identified in 10 rats after a transient (45-minute) occlusion of the coronary artery. There was a statistical difference between injured myocardial signal intensity and normal myocardium up to 45 minutes after administration, with an overestimation of infarct size compared with necropsy.29 In a recent study with rats, MR after administration of feruglose at a dose of 2mg/Fe/kg was able to detect acute rejection and proved useful in evaluating response to therapy. Images were taken from minute one to minute 45 after contrast injection and signal intensity curves over time were analysed.30 Signal intensity increases over time in cases of rejection due to the contrast extravasation in an inflamed and injured vascular bed. The utility of this approach for early detection and treatment monitoring in humans needs to be tested.
Clinical Applications and Future Directions
As mentioned above, there is only one BPA licensed for clinical application. Many others are being tested not only in animal models but also in phase II and phase III clinical trials. In the next 10 years, several of these new CAs will be available for clinical use and it will be necessary to decide not only the sequences to make but also the contrast best suited for the MR exploration based on patient pathology.
One of the most promising fields for BPAs is cardiac MR. Currently, the long acquisition time for coronary angiography makes BPAs the CAs of choice for visualisation of coronary vessels. Currently, clinical non-invasive coronary evaluation is performed by cardiac CT with the administration of iodine contrast and exposure to a high radiation dose. With improvements in MR technology and new BPAs, MR coronary angiography will become a reality. Nevertheless, the diagnostic accuracy and safety of CAs should be tested before their application in clinical practice for decision-making.
Myocardial perfusion is currently performed with ECAs. Evaluation of perfusion defects has to be performed during first-pass imaging and all of the images covering the left ventricular volume need to be taken quickly in a short period of time. BPAs, with their longer plasma half-lives, have proved useful in the evaluation of perfusion defects. Images can be obtained with higher spatial resolution over a longer period of time and left ventricular volume covered entirely. This way, perfusion defects could be more accurately depicted with these CAs. Preliminary reports comparing them with ECAs show a greater diagnostic accuracy. Along with CA development, software for myocardial perfusion quantification needs to be implemented, rendering perfusion quantification reproducible, easy to perform and feasible in clinical practice.
Myocardial viability evaluation is one of the major clinical indications for cardiac MR. BPAs are capable of detecting myocardial infarction compared with ECAs. The main problem is that the sequence acquisition needs to be carried out on time, at least 30 minutes after the contrast injection compared with 10 minutes with ECA. This time delay poses a problem, since cardiac MR studies are already long and make patients uncomfortable, tired and unwilling to collaborate with breath holding. Unless BPAs demonstrate better myocardial infarct delineation or the time delay is decreased, the role of BPAs in myocardial viability studies is doubtful.
Unresolved matters remain regarding BPAs. First is the clinical utility of these apparently superior new-generation CAs. Their role in coronary angiography is clear, since the long acquisition time with actual sequences requires an intravascular CA with long plasma half-life. However, the utility of BPAs compared with ECAs is not yet defined. They seem to be superior in detecting perfusion anomalies in both animals and humans, but does this translate to better diagnostic accuracy? Does it make any difference in everyday clinical practice decision-making? Is it necessary to use BPAs in myocardial perfusion and then wait for 30 minutes in order to obtain delayed enhancement images for myocardial viability evaluation? Further studies are needed to clarify the utility of BPAs in clinical practice. Besides these concerns, are all BPAs the same? Do they provide the same information and can they be used in the same way? Future research should also focus on comparing the different CAs under development, with special emphasis on their safety profile, adverse effects and main clinical utility.
CAs are part of MR studies in everyday clinical practice. BPAs, with their longer plasma half-lives, seem to overcome many of the classic weak points of CAs in that they allow longer acquisition times, leading to optimisation of spatial resolution and full anatomical coverage with greater signal-to-noise and contrast-to-noise ratios. Cardiac MR examinations, perfusion studies and coronary angiography may benefit from these BPAs, increasing the diagnostic accuracy of cardiac studies and their clinical applications. Ôûá