CHAPTER 22: MAGNETIC RESONANCE IMAGING AND COMPUTED TOMOGRAPHY OF THE VASCULAR SYSTEM

Imaging of the vasculature has evolved greatly in the past 20 years. Although the current gold standard remains invasive x-ray angiography, noninvasive modalities such as magnetic resonance imaging (MRI) and computed tomography (CT) are becoming routine for the evaluation of patients with vascular diseases. In many instances, either MRI or CT has replaced x-ray angiography as the imaging modality of choice in the assessment of patients with suspected vascular disease, because of the ever-increasing image quality, the noninvasive application, the ease and comfort for patients, and the clinical versatility of both CT and MRI. In addition to evaluating the degree of luminal stenosis, MRI and CT can now noninvasively detect the presence and composition of atherosclerotic plaques in various arterial beds. This chapter provides a comprehensive and state-of-the-art overview of the clinical indications for MRI and CT in the evaluation of vascular diseases, focusing on both angiographic and plaque-based assessments. It also introduces emerging molecular imaging approaches for the assessment of disease activity.

Techniques/Physics

Magnetic Resonance Angiography

Magnetic resonance angiography (MRA) can be divided into two categories: nonenhanced MRA and contrast-enhanced MRA (CE-MRA).¹ Nonenhanced MRA can be obtained by detecting the effect of blood flow on either the signal amplitude (time of flight [TOF]) or on the phase of moving protons (phase contrast [PC]). TOF angiography relies on differences in signal amplitude between in-slice stationary protons and protons in blood flowing into the slice. In-slice stationary protons become relatively saturated with repeated excitation pulses and produce low signal intensity, whereas inflowing blood protons have not experienced these excitation pulses, are not saturated, and generate high signal intensity. Two-dimensional (2D) or three-dimensional (3D) datasets can be acquired. Limitations of TOF imaging are long acquisition times and the need to position sections orthogonal to the direction of flow. In addition, slow flow or turbulent flow can lead to signal loss. PC angiography derives image contrast from differences in phases accumulated by stationary and moving spins in a magnetic field gradient. Phase data can be used to reconstruct velocity-encoded flow-quantification images or MRA images. With velocity-encoded imaging, phase amplitude is directly proportional to flow velocity, allowing for quantitative assessment of flow velocity and direction. This can assess flow and pressure gradients across stenoses in the carotid arteries, peripheral arteries, and renal arteries, as well as coarctation of the aorta. This technology also permits visualization of thoracic aortic dissection. However, clinical applications are hampered by the long acquisition times required for velocity-encoded imaging.

Other nonenhanced MRA sequences include electrocardiogram (ECG)-gated fast spin echo (FSE or turbo spin echo) and sequences based on steady-state free precession (SSFP). ECG-gated FSE has a shorter imaging time than TOF imaging and is sensitive to slow flow. It has, therefore, been used to assess peripheral and collateral vessels.² Image contrast with SSFP sequences stems from the T2/T1 ratio, which is high in blood compared to other tissues. SSFP sequences produce bright blood imaging without reliance on blood inflow. Both arteries and veins have high signal intensities with balanced SSFP sequences; therefore, venous signal must be nulled, for example, by saturation bands or spin labeling. SSFP sequences are increasingly being used to evaluate the thoracic aorta and coronary arteries.

First-pass CE-MRA with gadolinium-based contrast agents has gained widespread acceptance because of shorter acquisition times compared to noncontrast MRA. The strong increase of luminal signal intensity after intravenous injection of T1-shortening contrast agents, such as gadolinium chelates, allows for fast angiographic acquisitions with 3D gradient echo sequences.³ These techniques give an accurate evaluation of intra- and extracranial, thoracic, abdominal, and peripheral vessels. In many centers, CE-MRA has widely replaced conventional x-ray angiography for the evaluation of peripheral arterial vessels.⁴ However, a disadvantage of CE-MRA is the association between gadolinium-based contrast agents and nephrogenic systemic fibrosis.⁵ Nephrogenic systemic fibrosis is a rare, but life-threatening, condition involving the deposition of collagen in the skin and other organs. The etiology of this condition is not fully understood, but gadolinium-based contrast agents have been associated with its development in patients with renal failure (eGFR < 30 mL/m²). Different gadolinium contrasts agents have been divided into low, medium, and high risk for causing this condition by the European Medicines Agency. High-risk agents include the linear chelates of gadolinium, and these are contraindicated in patients with an eGFR less than 30 mL/m², acute renal impairment, and perioperative liver transplantation, as well as in neonates. Low-risk agents include newer cyclic preparations of gadolinium. These are considered safe in patients with an eGFR of more than 30 mL/m², and they can be used in patients with an eGFR below this threshold if the benefit of undergoing contrast MRI outweighs the risk. The volume of contrast agent used in this case should be minimized and repetition within 7 days avoided.⁶ In addition, recent research has identified gadolinium deposition in the brain and bones of patients who have undergone contrast enhanced magnetic resonance (MR), although the clinical implications of this finding remain unclear.⁷

Technical advances, such as synchronization of the arrival of contrast agent with MR acquisition,⁸ moving bed technology for multistation studies,⁹ parallel imaging,¹⁰ and k-space sharing methods,¹¹ have dramatically reduced acquisition times and paved the way for whole-body MRA. Whole-body MRA is well suited for repeated clinical examinations in patients with systemic diseases such as vasculitis or atherosclerosis. Whole-body MRA (Fig. 22–1) is feasible and accurate for simultaneous evaluations from the carotid down to lower limb arteries.¹² Compared with invasive angiography, the sensitivity and specificity of whole-body MRA for the detection of significant arterial stenosis were both 96%, similar to the accuracy of single-station MRA.¹³ However, intracranial and coronary artery imaging still requires dedicated acquisitions. MR scanners at high field strength 3.0 Tesla [T] are now widely available, but have both advantages and drawbacks for angiographic studies. On one hand, the intrinsic higher signal-to-noise ratio of tissues at 3 T allows shorter scan times and improved spatial resolution. On the other hand, higher field strengths increase field inhomogeneity, cause disturbances in the electrocardiogram, and increase acoustic noise and susceptibility artefacts.¹⁴ Until recently, these artifacts rendered SSFP imaging particularly challenging, although this has recently been improved with the development of advanced shimming and parallel imaging techniques. Further research aimed at improving image quality at a wide range of field strengths (≥ 3.0 T) is therefore ongoing. In the future, ECG and respiratory gating will also become less important for cardiovascular imaging as new free breathing techniques are developed.¹⁵

In summary, recent technical advances in MR technology have dramatically reduced acquisition times and allowed for larger spatial coverage. These advances have led to more widespread adoption of MRA for the clinical assessment of the cardiovascular system.

Computed Tomography Angiography

CT technology has advanced rapidly since its inception in the 1970s. CT acquires a 3D volume of data that can then be manipulated using postprocessing software. In step-and-shoot mode, the scanner acquires a single data set before moving, where as in helical (spiral) mode the scanner moves constantly as it rotates. Helical scanning is now used for most applications, initially with single-detector CT and now with multidetector CT (MDCT) systems.¹⁶ Wide-volume multidetector scanners are capable of acquiring a longer length of data in each rotation, thus reducing scanning time and potential motion artifact while increasing coverage.

CT angiography (CTA) involves the injection of iodinated contrast agents to opacify the blood vessels. A variety of iodinated contrast agents are now available with different chemical compositions and iodine concentrations. Care must be taken using iodinated contrast in patients with a history of contrast allergy. For such patients, if the benefit of the scan outweighs the risk, then the scan should be performed under close medical supervision and monitoring with rapid access to treatments for anaphylaxis. Nonionic, low, or iso-osmolar iodinated contrast should be used for high-risk patients. Pretreatment with steroid and antihistamine can be considered; however, the evidence for this is uncertain. Renal impairment is a relative contraindication to the use of iodinated contrast, which may be nephrotoxic. ECG gating allows acquisition of CT data in diastole when the heart is still. It is essential for coronary artery imaging and is also useful for imaging the thoracic aorta where motion artifact may be mistaken for an aortic dissection. In retrospective ECG gating, there is radiation exposure throughout the cardiac cycle, and individual phases of the cardiac cycle are reconstructed from these data. Newer prospective ECG gating involves radiation exposure only at predefined portions of the cardiac cycle, and thus substantially reduces radiation dose.

The resolution of current generations of CT scanners is excellent (0.35–0.4 mm³ with 64-MDCT). More recent advances in CT hardware include wide-volume detectors, dual source systems, new detectors, and faster gantry rotation times. Advances in CT software include iterative or model-based reconstruction algorithms, “partial scan” reconstructions, tube current/voltage optimization, and dose modulation. Dual-energy imaging harnesses the difference in tissue attenuation between images acquired with two different energy spectra. Dual-energy images can be acquired coincidentally with a dual-source scanner or with two separate acquisitions with a single-source scanner. This can be used to provide additional information on material composition and potentially reduce the effect of calcium blooming artifacts.¹⁷ Ongoing advances in CT imaging means that radiation doses will continue to fall. This is an important concern because of the association between cumulative radiation exposure and the lifetime risk of cancer.¹⁸

CTA is, therefore, a rapid and widely available noninvasive technique to assess the cardiovascular system, offering excellent spatial resolution. However, these benefits must be weighed against the risks of radiation exposure, and all attempts possible should be made to keep radiation dose “As Low As Reasonably Achievable,” or ALARA.

Practical Approach

CT and MRI scanners are now widely available in the vast majority of hospitals in the developed world. Often these are located in radiology departments, so that close collaboration is required between radiologists and cardiologists as well as vascular surgeons, cardiothoracic surgeons, and pulmonologists. CT and MRI are both performed in the outpatient and inpatient settings, although the shorter duration of CT scans and the ability to monitor patients closely within the scanner make it more suitable for imaging in emergency situations. Although emergency imaging is possible with MRI, this is complicated by the need for MRI-compatible monitoring and life-support equipment. Prior to CT or MRI, patients should be screened for potential risk for allergy or contrast-induced nephropathy, as discussed above. In order to perform contrast-enhanced imaging, an intravenous catheter should be inserted, ideally into an antecubital fossa vein.

Magnetic Resonance Imaging

Prior to entering the restriction zone around the MRI scanner, patients should be screened for potential contraindications to MRI scanning, such as the presence of orbital metal foreign bodies or implanted devices, stents, or clips that are MRI incompatible. The patient is positioned on the scanner bed with the receiver (+/– transmit) coil positioned over the area to be imaged. Appropriate coil selection and positioning can improve image signal-to-noise ratio. Because of the high acoustic noise of most MRI sequences, ear plugs and/or headphones may be required. The patient can also be provided with an alert buzzer to allow them to contact staff if required. If cardiac images are to be acquired, then ECG leads are also attached.

Computed Tomography Imaging

Patients are positioned on the scanner bed, which moves through the scanner gantry. If cardiac images are to be acquired, ECG leads are also attached. Initial low-dose 2D images (called “localizers”) are acquired, which are used to plan the scan. The scanner tube voltage and current should be optimized to the individual patient in order to obtain diagnostic images at the lowest possible radiation dose. Similarly the scan range should cover only the area of interest. Depending on the indication, noncontrast, contrast-enhanced, and delayed images can be acquired. Contrast-enhanced images can be obtained using either a specific time delay after the contrast injection or using bolus tracking, where the attenuation density in a specific region must reach a target value before the scan commences. Manual triggering of contrast-enhanced images may also be used, particularly for patients with challenging contrast hemodynamics. Specifics of imaging individual vascular beds are discussed further below. Images are reconstructed using filtered back projection, iterative, or model-based reconstruction algorithms, and a 3D volumetric data set is produced for assessment.

Having reviewed the different imaging approaches to noninvasive angiography, the subsequent sections will focus on how these can be applied to different vascular territories within the body: the carotid arteries, the aorta, the peripheral arteries, and the pulmonary arteries.

Carotid Arteries

Stroke represents the third leading cause of death in the United States, accounting for 600,000 cases each year. The risk of stroke associated with carotid stenosis has been evaluated in several trials. The North American Symptomatic Carotid Endarterectomy Trial (NASCET),¹⁹ the European Carotid Surgery Trial (ECST),²⁰ and the Veterans Affairs Trial (VAT)²¹ demonstrated that carotid endarterectomy reduced recurrent stroke in symptomatic patients with severe stenosis (70%–99%) of the ipsilateral carotid artery (see Chap. 95). Therefore, carotid endarterectomy is recommended if the morbidity of surgery is reasonably low (< 6%). Symptomatic patients with moderate stenosis (50%–69% in NASCET) still benefited from surgery, although the overall gains were more modest. For asymptomatic patients with severe carotid stenosis, carotid endarterectomy significantly reduced the 5-year stroke rate in selected subgroups with the highest risk. Carotid artery stenting is an alternative approach for patients at high surgical risk, in whom outcomes are similar after 10 years of follow-up.^22,23 In most of the original surgical trials, intra-arterial digital subtraction angiography (DSA) was the gold standard assessment of stenosis severity. However, the risk of thromboembolic complications from DSA means that noninvasive imaging modalities are now favored, such as Doppler ultrasonography, MRA, and CTA.²⁴

Magnetic Resonance Angiography of Carotid Arteries

MRA has emerged as one of the noninvasive methods of choice for carotid imaging, providing accurate assessment of carotid stenosis without the use of ionizing radiation and the problem of calcium blooming observed with CT. Before the advent of CE-MRA, TOF acquisitions were used to image the carotid arteries, demonstrating good agreement with DSA.²⁵ However, TOF acquisitions are sensitive to turbulent flow, can overestimate stenoses, and have long acquisition times, increasing potential motion artifacts. Although TOF MRA remains useful for patients with contraindications to MRI contrast, such as renal insufficiency, CE-MRA has become the technique of choice for imaging the carotid arteries. Image acquisition needs to be synchronized with the arterial phase of the contrast bolus in order to prevent venous enhancement. Unlike TOF MRA, vascular contrast is less sensitive to turbulent flow, particularly in cases of high-grade stenosis; moreover, a large volume from the aortic arch to the circle of Willis (Fig. 22–2) can be obtained in less than 1 minute. CE-MRA demonstrates high sensitivities and specificities for the evaluation of carotid artery stenosis (Table 22–1) and is currently the most accurate noninvasive imaging for carotid stenosis.^26,27 Although CE-MRA offers two to three times lower spatial resolution than DSA or CTA, this is likely to improve with new technologic developments, such as high field strengths,²⁸ parallel imaging techniques, and the use of blood-pool contrast agents that remain in the circulation longer.²⁹

Computed Tomography Angiography of Carotid Arteries

With the advent of MDCT, the whole length of the carotid artery, from the aortic arch to the circle of Willis, can be imaged with high spatial resolution in a true arterial phase, limiting adjacent venous enhancement (Fig. 22–3). CTA demonstrates higher sensitivities and specificities than TOF MRA, but lower than those for CE-MRA for the evaluation of the degree of carotid stenosis (see Table 22–1).³⁰ Although spatial resolution is higher than for MRA, the lower specificity of CTA can be explained by the difficulty in evaluating the degree of luminal stenosis in the presence of extensive calcification. Dual-energy imaging might potentially reduce this calcium “blooming” artifact and aid assessment of plaque composition. CTA can also be used to assess carotid artery stents for stent integrity and in stent restenosis.³¹

In summary, MRA and CTA of the carotid arteries are readily accessible, and their accuracy is now considered sufficient to outweigh the risks associated with invasive DSA. A combination of screening with Doppler ultrasound followed by MRA or CTA in patients with hemodynamically significant stenosis or poor ultrasound image quality has become clinically routine.

Aorta

Aortic Dissection

Aortic dissection is a life-threatening disease in which early diagnosis and treatment is critical for survival (see Chap. 93).³² The yearly incidence of aortic dissection is 10 to 20 cases per million. Misdiagnosis can have fatal consequences because untreated patients with type A aortic dissection have mortality as high as 25% at 24 hours and 75% by 2 weeks.³³ Retrograde invasive aortography was the first accurate method for the diagnosis of aortic dissection. It requires the presence of an intimal flap and two distinct lumens, with an overall sensitivity of 88% and specificity of 94% for the detection of aortic dissection.^34,35 Major drawbacks are the time required for the procedure and the complications associated with instrumenting an already damaged aorta. Invasive angiography has now been superseded by noninvasive imaging techniques that have greatly improved the early diagnosis and management of aortic dissection.

CTA is now the mainstay for the diagnosis of acute aortic syndromes. It is a rapid, noninvasive diagnostic test that is routinely available on a 24-hour basis. An initial nonenhanced CT is performed in order to identify intramural hematoma that would otherwise be obscured by intravenous contrast. Intramural hematoma appears as a crescent of high attenuation within the wall of the aorta or, less often, as localized thickening of the aortic wall with internal displacement of intimal calcifications. CTA using iodinated contrast can then be performed. If aortic root or thoracic aorta involvement is suspected, ECG-gated images of the thoracic aorta should be acquired, in order to minimize motion artifacts that can mimic the presence of dissection.³⁶ CTA has a sensitivity of 83% to 95% and a specificity of 87% to 100% for the diagnosis of acute aortic dissection.³⁷ It can visualize the intimal flap, extent of dissection, patency of the false lumen, size of the aorta, branch vessel involvement, presence of pericardial fluid, and evidence of end-organ ischemia, and it can potentially identify other unexpected causes for the patient’s symptoms. Limitations of CTA for the evaluation of aortic dissection include the inability to obtain hemodynamic information, as may be necessary when there is concomitant aortic regurgitation.

MRA is an accurate and reliable alternative imaging technique for the diagnosis of aortic dissection. Acquisition times have been significantly shortened by implementation of faster and stronger gradients in recent MR systems. The thoracic aorta and major aortic branches can now be imaged in a few seconds with 3D CE-MRA³⁸ (Fig. 22–4) or in a few minutes using noncontrast bright-blood sequences such as steady-state gradient echo sequences. As with invasive angiography, the diagnosis of aortic dissection is based on visualization of an intimal flap and two distinct aortic lumens. Velocity-encoding sequences can help to differentiate true and false lumens and to determine whether aortic side branches are perfused by the true or false lumen. Black-blood sequences offer high spatial and contrast resolution and are well suited for the analysis of the aortic wall. They can identify intramural aortic hematoma as a crescentic aortic wall thickening (≥ 5 mm) without intimal flap or tear. Of all the imaging techniques, MRI has the highest sensitivity and specificity for detecting aortic dissection (98%).³⁹ However, the use of MRI for the detection of aortic dissection is limited by its lack of availability in urgent situations and the difficulty in monitoring and managing critically ill patients in the MRI scanner.

Transesophageal echocardiography, like CTA and MRA, has demonstrated high accuracy for the diagnosis of aortic dissection and is an alternative diagnostic technique.⁴⁰ Because there is a clear time-dependent mortality in patients with aortic dissection who do not receive treatment, decisions about which imaging technique to use should take into account availability in an emergency.

Aortic Aneurysm

Thoracic Aortic Aneurysm

Thoracic aortic aneurysms (TAAs) are typically discovered in asymptomatic patients (see Chap. 93), with symptoms occurring in the setting of either a complication of the disease (ie, rupture or dissection) or when these complications are imminent.³² The most frequent etiologies of TAAs include degenerative aneurysms, bicuspid aortic valve disease, and aneurysms associated with connective tissue disease (eg, Marfan syndrome). Patients with bicuspid aortic valve disease or connective tissue disease should undergo screening at regular intervals for the presence of associated aortopathy. The risk of complications depends mainly on TAA diameter, location, and etiology. Thresholds for preventive surgical interventions have been defined accordingly. When the diameter of the ascending aorta is more than 55 mm for degenerative aneurysm and more than 50 mm for aneurysm associated with bicuspid aortic valve or Marfan disease, surgery is recommended. These thresholds can be lowered in case of rapid expansion of TAA (> 5 mm/y) or in select connective tissue diseases (eg, Loeys-Dietz syndrome). Expression of aortic diameters indexed to body surface area is preferred, in particular when body surface area is outside the normal range.

CTA and 3D CE-MRA perform equally well in depicting the thoracic vascular anatomy, particularly the relationship of aortic aneurysms to branch vessel origin and the extent of aneurysm.⁴¹ Acquisitions should be performed in diastole and with breath holding to limit motion artifacts in the ascending aorta. CT can combine noninvasive evaluation of coronary arteries and imaging of the thoracic aorta. 3D CE-MRA has demonstrated a diagnostic accuracy of 100% for assessing the size and extent of the aneurysm and its relationship to aortic branches, compared with conventional angiography and surgical exploration.⁴² A disadvantage of MRA is the inability to assess calcification at operative sites. CE-MRA should be complemented by cross-sectional black-blood imaging of the aortic wall to determine the presence of an intraluminal thrombus. In addition, the presence of inflammation, such as in mycotic aneurysm or aortitis, might be suspected when delayed enhancement is detected in the aortic wall and surrounding soft tissues. The aortic valve is frequently involved in ascending aortic aneurysms. Cine SSFP sequences of the valve can help to determine whether the aortic valve is bicuspid and either stenotic or regurgitant. Before surgical intervention involving the descending aorta is undertaken, care should be observed to identify the origin of the artery of Adamkiewicz with MRA to avoid spinal complications after surgery.⁴³ Both MRI and CT can be used to monitor the size of thoracic aneurysms and dissections following surgical repair or conservative management. In these scenarios, the need for repeat imaging often favors MRI, given the absence of ionizing radiation.

Abdominal Aortic Aneurysm

The abdominal aorta is the most frequent arterial location of aneurysms, with abdominal aortic aneurysms (AAAs) usually being defined as a maximal aortic diameter of 30 mm or more. The prevalence of asymptomatic AAAs in men and women over the age of 60 years is evaluated at 4% to 8% and 0.5% to 1.5%, respectively, and increases with age. AAAs share common risk factors with atherosclerosis, except diabetes. At least 15,000 deaths in the United States are caused by ruptured AAAs annually. Once an AAA ruptures, mortality is high (80%–90%). In contrast, elective repair is associated with a low rate of complications (2%–3%). Screening strategies have therefore been developed to identify patients with large AAAs (> 55 mm in diameter), who might benefit from elective repair. Ultrasonography remains the most cost-effective imaging technique for detection and follow-up of abdominal AAA expansion. Main indications for CTA and MRA are preoperative evaluation of AAAs and follow-up of patients who have had open or endovascular repairs. Preoperative assessment of AAAs before open or endovascular repair should include determination of the maximum transverse diameter; the relation of the AAA to the renal arteries; the length, diameter, and angulation of the normal-caliber aorta below the renal arteries before the aneurysm (ie, the infrarenal neck); the presence of iliac or hypogastric aneurysms; and serious occlusive disease in the iliac or renal arteries.

CTA represents the preferred imaging modality for preoperative evaluation of AAAs before open or endovascular surgery, because it allows precise assessment of each of the aforementioned parameters. It provides a volumetric acquisition that allows for multiplanar and 3D reconstructions to be generated perpendicular to the long axis of the aneurysm, resulting in greater accuracy of the measurements of aneurysmal size.⁴⁴ As with other indications of CT, the limitations of CTA are the need for ionizing radiation and iodinated contrast. In addition, extensive calcification can preclude accurate diameter determination of the iliofemoral arteries.

MRI represents an alternative to CTA for the preoperative evaluation of AAA, particularly in cases of renal insufficiency or advanced arterial calcification, although attention should be paid to the risk of developing nephrogenic systemic fibrosis following gadolinium administration in patients with the former. Angiography can be obtained either with nonenhanced TOF sequences that require long acquisition times or, more rapidly, with CE-MRA using fast gradient echo sequences (Fig. 22–5). Accurate dimensions of the aneurysm can be determined with multiplanar reconstructions,⁴⁵ and MRA can also assess extension into the iliac and renal arteries and the presence of concomitant renal artery stenosis. Delayed axial acquisitions can be used to evaluate for intraluminal thrombus.⁴⁶ CT and MRI can also be used to assess outcomes following endovascular repair, in particular for evidence of endoleak (persistent blood flow within the aneurysm sac), which is a common complication. One approach is to perform CT and abdominal x-ray imaging at 1, 6, and 12 months postrepair and then annually. However, this involves substantial cumulative radiation exposure and repeated iodinated contrast exposure. More recent protocols therefore recommend the increased use of ultrasound rather than CT; MRA can be considered as an alternative to CT when ultrasound assessment is nondiagnostic.⁴⁷

Peripheral Vessels

The most common initial investigation in patients presenting with suspected peripheral vascular disease is the ankle brachial pressure index. This is the ratio of the systolic blood pressure measured at the ankle to that measured at the brachial artery. An ankle brachial pressure index of 0.90 or less should then prompt further investigation to confirm a diagnosis of peripheral artery disease. This is commonly performed with either CT or MRA⁴⁸ (see Chap. 96).

Magnetic Resonance Angiography of Peripheral Vessels

MRA is indicated for the preoperative evaluation of patients with lower limb ischemia. Similar to invasive angiography, MRA can depict the location and degree of stenosis of the peripheral arteries. TOF angiography was the first technique used to evaluate the peripheral arteries,⁴⁹ but has been supplanted by CE-MRA owing to shorter acquisition times (a few seconds) and greater accuracy with reported sensitivities of 96% to 98% and a specificity of 96% for the detection of significant stenoses (Table 22–2).⁵⁰ The high degree of agreement between MRA and conventional angiography has led to surgical planning based on MRA alone (Fig. 22–6). CE-MRA is also suitable for the evaluation of lower extremity bypass grafts and has been found to be 100% sensitive and specific in this setting.⁵¹ For CE-MRA, the beginning of the acquisition should be perfectly synchronized with the arrival of the contrast bolus to avoid venous enhancement.

Early detection of renovascular disease has been previously hampered by the lack of an adequate noninvasive diagnostic modality. CTA and CE-MRA were found to be more accurate than ultrasound scanning or captopril renal scintigraphy for the diagnosis of renal artery stenosis.⁵² Sensitivity and specificity values for the diagnosis of renal artery stenosis with CE-MRA range from 82% to 100% and 64% to 100%, respectively. In addition, PC-MRA also provides functional information such as the velocity profile of renal arterial flow.⁵³ MRA can also be used to image the splanchnic vessels; in particular CE-MRA, and true fast imaging with steady-state precession (true-FISP) provide excellent spatial resolution and signal-to-noise.^54,55

Computed Tomography Angiography of Peripheral Vessels

The diagnostic accuracy of CTA for the evaluation of peripheral vessels compares well with MRA and invasive angiography. Shorter acquisition times, thinner slices, and higher spatial resolution have enabled rapid scanning of the whole vascular tree with small volumes of iodinated intravenous contrast (Fig. 22–7). The main clinical indications of CTA for imaging the lower extremities include the evaluation of arterial and aneurysmal disease, patency and integrity of bypass grafts, traumatic arterial injury, and acute ischemia.⁵⁶ More recently, CTA of the peripheral vasculature has been widely incorporated into the assessment of patients considered for percutaneous valve replacement (eg, transcatheter aortic valve replacement [TAVR]) or other percutaneous procedures.⁵⁷ Here, it can assess whether peripheral access is feasible but also help to measure the aortic valve annulus, the height of the coronary arteries above the valve, and the degree of calcification in the left ventricular outflow tract. Studies using CTA in peripheral vascular disease report sensitivity and specificity rates of greater than 95% for the detection of stenosis (see Table 22–2).⁵⁸ Once more this diagnostic accuracy is limited in the presence of severe and diffuse calcifications,⁵⁹ although CTA seems of particular use in the evaluation of calf vessels of patients with proximal occlusions, outperforming conventional invasive angiography.⁶⁰ For the evaluation of renal and mesenteric arterial stenoses, the diagnostic accuracy of CTA is similar to that MRA.⁵²

Pulmonary Arteries

Although pulmonary x-ray angiography is the gold standard for the diagnosis of pulmonary embolism (PE), it is invasive, can yield false-positive results, and is associated with a 1% risk of major procedural complications and a mortality of 0.5%.⁶¹ It is now rarely performed in most centers. Instead a variety of noninvasive tests are preferred, including CTA and MRA, alongside clinical probability scores, plasma D-dimer concentrations, lung scintigraphy (ventilation/perfusion), and compression venous ultrasonography of the legs (see Chap. 109).

Computed Tomography Angiography of Pulmonary Arteries

CT pulmonary angiography (CTPA) is now the mainstay for the diagnosis of PE in many centers. CTPA (Fig. 22–8) has been proven to be an accurate, safe, noninvasive, rapid, and cost-effective technique for the direct detection and demonstration of intraluminal PE.^62,63 In addition to the identification of pulmonary artery thrombus, CTPA can also provide information on secondary right heart strain.⁶⁴ The sensitivity of spiral CT is in the order of 90% for central, lobar, or segmental PE (Table 22–3). The advent of MDCT provides the ability to analyze subsegmental PE, although the clinical importance of this finding remains uncertain.⁶⁵ Current guidelines for the diagnosis of PE recommend CTPA alongside clinical probability assessment and D-dimer measurement.^66,67 Low- or intermediate-risk patients should undergo D-dimer testing in order to avoid unnecessary CTPA, whereas high-risk or hemodynamically unstable patients should undergo direct imaging. Patients with a negative CTPA and clinical suspicion of venous thrombosis can be considered for compression venous ultrasonography. CT venography is no longer recommended because of the significant dose of radiation involved and lack of additional value compared to ultrasound.⁶⁷ More recently, dual-energy CT assessments of lung perfusion have demonstrated potential in improving the diagnosis of acute and chronic PE.^68,69 However, the utility of these novel techniques will need to be balanced against the added radiation exposure.

Magnetic Resonance Angiography of Pulmonary Arteries

MRA of pulmonary arteries can be acquired with several techniques. Large pulmonary arteries can be imaged using TOF sequences,⁷⁰ although CE-MRA is now preferred. 3D CE-MRA can be acquired in a single breath-hold (Fig. 22–9) because of the development of short time-resolved 3D gradient echo sequences and parallel acquisition techniques.⁷¹ A number of clinical studies have reported high sensitivities and specificities for the detection of central PE compared with conventional pulmonary angiography (see Table 22–3). However, the sensitivity for the detection of smaller peripheral emboli is lower.⁷² In addition, MRI sequences have been developed to allow for the evaluation of lung ventilation in addition to perfusion sequences,⁷³ whereas a combined one-stop MRI approach for PE and deep venous thrombosis also seems feasible.⁷⁴ However, the high frequency of inconclusive MRAs and its lack of availability in an emergency setting currently limits the clinical application of this approach.^66,71

Coronary artery angiography is covered in Chapter 16. Please refer to it.

Atherosclerosis is characterized by accumulation of lipids, inflammatory cells, and connective tissue within the arterial wall. It is a chronic, progressive disease with a long asymptomatic phase. The first pathologic abnormality is the fatty streak, caused by accumulation of lipids and macrophages in the subendothelial space. These streaks can be observed in the aorta from as early as the second decade of life,⁷⁵ and they develop over time into mature atherosclerotic plaques consisting of a central lipid core covered by an endothelialized fibrous cap containing vascular smooth muscle cells and connective tissue. As the plaque grows, the affected vessel expands outward so that the lumen diameter and, thus, blood flow are preserved in a process known as positive remodeling.⁷⁶ Consequently, even large plaques can be accommodated without producing symptoms that would result from luminal stenosis. Eventually, the artery can expand no further, and the plaque begins to encroach into the lumen of the vessel, hindering blood flow and causing angina at times of high demand. Atherosclerotic plaques can remain quiescent for years. However, they can become life threatening when they initiate clot formation in the vessel lumen, disturbing blood flow or allowing emboli to break off and lodge in the downstream circulation.⁷⁷ Most commonly, this occurs with sudden fibrous cap rupture, which exposes the thrombogenic, tissue-factor-rich lipid core to circulating blood. Alternatively, plaque erosion of the endothelium overlying the fibrous cap can lead to the formation of a platelet-rich thrombus and adverse clinical events. Endothelial erosion is thought to account for approximately 30% of plaque rupture events overall and seems particularly common in women.⁷⁸ The processes leading to plaque erosion are less well characterized than those implicated in plaque rupture. Both forms of plaque disruption invariably lead to local platelet accumulation and activation at the site of rupture. This may trigger the clotting cascade, with thrombus formation and, if extensive, complete vessel occlusion. However, symptoms are not invariable after fibrous cap disruption; indeed, subclinical plaque rupture appears common with up to 70% of obstructive coronary plaques containing histologic evidence of previous rupture and subsequent repair.⁷⁹ This is particularly likely to occur if high blood flow through the vessel prevents the accumulation of a large occlusive thrombus, as frequently occurs in the carotid system. Additionally, the body’s natural fibrinolytic pathways can lead to resolution of some thrombi, allowing subsequent healing of the cap and overlying endothelium. Whether and how a plaque becomes symptomatic is determined by its macroscopic structure, its microscopic composition, and the thrombogenicity of the blood. Plaques that precipitate plaque rupture usually have several key characteristics, consisting of a thin fibrous cap (< 65 μm) with few smooth muscle cells and a heavy infiltrate of inflammatory cells, principally macrophages, microcalcification, and a large necrotic core accounting for more than half of the volume of the plaque. Each of these characteristics therefore represents a potential imaging target for identifying high-risk plaques. Although prospective data have suggested that the majority of these so-called high-risk lesions either heal or rupture subclinically rather than causing myocardial infarction,⁸⁰ such plaques rarely exist in isolation and can serve as a marker of patients with advanced and active atheroma. On this basis, interest persists in identifying these lesions as a means of pinpointing patients at high risk of cardiovascular events. A limitation, however, is that these same plaque characteristics are not necessarily associated with plaque erosion.

Here we will briefly review how CT, MRI, and hybrid imaging platforms might be used to investigate atherosclerotic plaque, focusing on (1) their ability to assess plaque burden, high-risk plaque characteristics, and increased disease activity and (2) how these techniques might improve patient risk stratification and our understanding about the pathogenesis and progression of atherosclerosis.

Magnetic Resonance Plaque Imaging

Atherosclerotic plaque MRI relies on the same fundamental principles as other MRI techniques, making use of differences in relaxation times (T1 and T2) and proton density to generate soft-tissue contrast and to provide detailed information about atherosclerotic plaque composition. This is increasingly being applied to the large and relatively immobile carotid arteries and thoracic aorta, with intensive research aimed at transferring these techniques into the coronary vasculature (Fig. 22–10).

Plaque Burden

Simple measures of the atherosclerotic plaque burden in different vascular beds can provide powerful prognostic information, presumably on the basis that the more plaques a patient has the more likely one will rupture or erode and cause an event. MRI can effectively image atherosclerotic plaque using high-resolution, black-blood, FSE sequences. Preparatory pulses are added in order to null signal in the blood pool (hence the black blood) and to improve contrast between the plaques and surrounding tissue. Measurements of plaque thickness in 2D and plaque volume in 3D can then be performed, providing quantification of the atherosclerotic plaque burden.

In the aorta, these measurements demonstrate a close correlation with transesophageal echography⁸¹ and with increasing number of cardiovascular risk factors.^82,83 Similar approaches have been used to measure plaque thickness and the plaque volume in the carotid arteries,⁸⁴ providing measures of atherosclerotic burden that predicts major adverse cardiovascular outcomes.⁸⁵ Translating these approaches in to the coronary arteries has proved more challenging. The first description was in 2000, when black-blood, breath-held, single-slice imaging was used by Fayad et al to obtain high resolution, cross-sectional images of individual coronary plaques.⁸⁶ Measurement of plaque thickness was then possible and found to be increased in patients with established ischemic heart disease versus control subjects,⁸⁶ and again associated with cardiovascular risk factors.⁸⁷ Such measurements are, of course, prone to sampling error, whilst the ability to measure plaque burden across the entire coronary vasculature remains elusive.

Carotid MRI plaque burden assessments using MRI have proved useful in assessing the efficacy of lipid-lowering therapies and their ability to induce atherosclerotic plaque regression. Indeed, the excellent reproducibility of these measurements means that relatively few patients are required in order to demonstrate treatment efficacy. This was demonstrated in a recent study comparing high- and low-dose statin regimens where improved atheroma burden reduction was demonstrated with the former using only 29 and 22 patients per group.^88,89

Similar black-blood imaging techniques and wall thickness measurements can be used to identify positive remodeling. Indeed, identification of positive remodeling in the aorta and carotids has been shown to identify patients with an increased risk of future adverse cardiovascular outcomes.⁸⁵ Moreover, positive remodeling is also detectable with MRI in the coronary arteries, although the relationship between this MRI finding and prognosis is yet to be established.⁹⁰

Plaque Composition

A key advantage of MRI of the vasculature is the soft-tissue contrast it provides, potentially allowing detailed examination of the composition of individual atherosclerotic plaques. As with other approaches, this is much easier in the carotid versus the coronary arteries. Yuan et al^91,92 demonstrated that multicontrast MRI of human carotid arteries had a sensitivity of 85% and specificity of 92% for the identification of the lipid necrotic core. MRI is also useful in detecting intraplaque hemorrhage, a process believed to play a key role in triggering plaque rupture and growth.⁹³ T2*-weighted sequences have been used to accurately image intraplaque hemorrhage in carotid atherosclerotic plaques,⁹⁴ which in a prospective study was associated with accelerated plaque expansion over 18 months.⁹⁵ Plaque hemorrhage and luminal thrombosis can also be detected on T1-weighted imaging as high-intensity plaques,⁹⁶ with a recent meta-analysis of 689 patients demonstrating a six-fold increase in cerebrovascular events in patients with high-intensity carotid plaque.⁹⁷ Importantly, T1-weighted imaging has also been successfully used in the coronary arteries, again with prognostic implications. Indeed, a recent large study demonstrated high-intensity plaques in 159 out of 568 patients with known or suspected coronary artery disease. Forty-one of these subjects subsequently developed a coronary event with the presence of a high intensity plaque acting as an independent predictor on multivariate analysis (hazard ratio, 3.96; confidence interval, 1.92–8.17).⁹⁸

Late gadolinium enhancement detects regions of extracellular expansion in the myocardium, and in carotid atheroma similarly identifies areas of interstitial edema, angiogenesis, and fibrosis. In particular, late enhancement toward the adventitia relates to regions of inflammation and angiogenesis on histology,⁹⁹ and in a retrospective study localized to the culprit lesions of patients who had suffered a recent stroke. Late gadolinium enhancement can also improve visualization of the atherosclerotic fibrous cap, allowing estimation of the carotid cap thickness and the size of its necrotic core.¹⁰⁰ Importantly, this approach can also be used to identify regions of carotid plaque rupture or ulceration¹⁰¹ that may or may not be clinically apparent.¹⁰² Attempts have been made to use this technology in the coronary arteries, with early studies showing that the presence of coronary late gadolinium enhancement correlated with the severity of atherosclerosis.¹⁰³

Macrophages play a pivotal role in the destabilization of atherosclerotic plaques, secreting locally abundant quantities of fibrous cap–degrading matrix metalloproteinases, proinflammatory cytokines, and tissue factor.¹⁰⁴ These cells can be imaged using iron oxide contrast agents that have superparamagnetic properties on T2*-weighted sequences. After injection of ultra-small superparamagnetic particles of iron oxide (USPIO), these particles are removed from the circulation by the reticuloendothelial system and accumulate in macrophages present in atherosclerotic plaques. The distortion in the MRI signal that this causes can then be used to estimate the macrophage burden in these areas.¹⁰⁵ Kooi et al¹⁰⁶ studied 11 symptomatic patients scheduled for carotid endarterectomy with USPIO-enhanced MRI and found a 24% decrease in signal intensity in the culprit vessel on T2*-weighted sequences with histological evidence of USPIO uptake in 75% of these plaques postsurgery. The USPIO signal also appears to be modifiable with drug therapy, with high-dose statins inducing a stronger decrease in USPIO carotid plaque uptake than does lower-intensity statin therapy.¹⁰⁷ Current limitations of ferumoxtran-enhanced MRI are the obligatory 36-hour delay after injection before imaging¹⁰⁸ and the artifact on T2*-weighted sequences.

In summary, because of the absence of ionizing radiation, MRI is an excellent imaging technology for the noninvasive detection and serial monitoring of atherosclerotic plaque volumes. High image quality and sensitivity to small changes in plaque size mean that there is little variance between measurements, permitting the use of small sample sizes in comparative mechanistic studies. In addition, MRI can be used to investigate features of plaque composition, including the presence of plaque hemorrhage, angiogenesis, and macrophage burden. In the future, we are likely to see an expansion in molecular MRI contrast agents targeting specific disease processes, similar to current nuclear techniques, but with added potential as drug delivery agents.^109,110,111

CT coronary plaque imaging, including assessments of plaque burden, are covered in Chapter 16. Please refer to it.

Positron Emission Tomography and Disease Activity

Anatomic imaging with x-ray angiography, MRI, and CTA can identify arterial stenoses and plaque composition, but these techniques do not inform about disease activity. This is potentially important given that patients with metabolically active disease and more rapid progression are more likely to have events than patients with inactive stable atheroma.¹¹² Positron emission tomography (PET) is a molecular imaging technique that is highly sensitive in its ability to detect the activity of specific disease processes that have been targeted with specially designed radiotracers. PET imaging has a very high sensitivity but low spatial resolution. As a result, it is generally performed on hybrid PET/CT or PET/MRI systems that combine the functional information of PET with the anatomic detail of CT or MRI.

Arterial inflammation is a key driver of atherosclerosis and in particular the precipitation of acute plaque rupture and adverse cardiovascular events. The radiopharmaceutical agent ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) is a glucose analog and identifies cells with high glucose utilization. These include macrophages, which use substantially more glucose than neighboring cell types in the vasculature,¹¹³ particularly in hypoxic conditions.¹¹⁴ The ¹⁸F-FDG uptake in the arterial wall was first noted in the aorta of patients undergoing PET imaging for cancer staging.¹¹⁵ Since these early studies, it is now established that ¹⁸F-FDG uptake is generally greater in symptomatic carotid plaques compared with asymptomatic lesions.¹¹⁶ Also, the arterial ¹⁸F-FDG signal is linked to levels of inflammatory biomarkers^117,118 and the number of components of the metabolic syndrome,¹¹⁹ and it appears to identify patients with an adverse prognosis.^120,121 More recently, it has been demonstrated that arterial ¹⁸F-FDG uptake can be reduced by drug therapy (Fig. 22–11). Indeed, ¹⁸F-FDG-PET is increasingly being used as an end point in trials assessing the anti-inflammatory effects of novel antiatherosclerosis drugs in the carotid arteries and thoracic aorta.^122,123,124 Studies using this tracer in the coronary arteries have had mixed results, predominantly because glucose is also the predominant energy source of the myocardium, with uptake in the heart muscle often obscuring coronary activity.^125,126 This has led to interest in more specific PET markers of inflammation^127,128 and tracers targeting other disease processes of interest.

In addition, ¹⁸F-fluoride is a PET tracer increasingly being used to measure the activity of calcific processes in the vasculature.^{129,130 18}F-Fluoride preferentially binds regions of vascular microcalcification activity.¹³¹ These are beyond the resolution of CT, a technique that instead detects macroscopic calcium deposits. The difference is potentially important, given that macroscopic calcium imparts stability to the plaque, whereas microcalcification is consistently associated with high-risk coronary lesions and increased risk of rupture.¹³² Two clinical trials have explored ¹⁸F-fluoride uptake in the coronary vasculature, demonstrating that unlike ¹⁸F-FDG, ¹⁸F-fluoride uptake can be readily detected in these vessels with excellent signal-to-noise (Fig. 22–12). In the first trial, increased tracer uptake localized to individual plaques and in so doing identified high-risk patients with increased Framingham risk scores.¹³³ The second demonstrated that the plaques with increased ¹⁸F-fluoride activity had multiple high-risk characteristics, including a large necrotic core, microcalcification, and positive remodeling. Perhaps most interestingly, this same study demonstrated that in patients post–myocardial infarction, ¹⁸F-fluoride localized to the exact site of plaque rupture, the culprit plaque, in 37 of the 40 subjects recruited. These studies indicate that ¹⁸F-fluoride can identify patients with increased metabolic activity in their coronary arteries and at least retrospectively appears to identify the individual lesions responsible for coronary events.¹²⁶ The ability of this marker of disease activity to prospectively identify patients at risk of myocardial infarction is currently being tested in a large prospective multicenter study (PREFFIR; clinicaltrials.gov NCT02278211).

Noninvasive methods such as MRA and CTA are increasingly being used to assess vascular disease at different sites in the body. CTA plays a major role in pulmonary and coronary imaging, whereas in the iliofemoral and carotid arteries, MRA is generally preferred. More generally, noninvasive imaging is undergoing rapid evolution, allowing us to investigate plaque burden, plaque composition, and disease activity in great detail and potentially within a single scan. The challenge will be to translate these exciting advances in to clear-cut improvements in the clinical care offered to patients with cardiovascular disease.

We would like to thank Dr. Fabian Hyafil, Dr. Jean-Christophe Cornily, and Dr. James H. F. Rudd, who contributed to the previous version of this chapter in the 13th edition.