Incremental value of PET and MRI in the evaluation of cardiovascular abnormalities

Myocarditis is characterized by an inflammatory reaction of the myocardium. The aetiology of myocarditis can be endogenous (toxins, autoimmune processes, etc.) or exogenous (bacteria, viruses, fungal, or parasites) [22]. MRI and PET provide complementary information for the assessment of myocardial inflammatory processes, including myocarditis [23]. MRI sequences used for myocarditis include the following: (1) T2-weighted (T2-W) images, which show myocardial oedema, an important hallmark of reversible inflammatory injury. Typically, myocardial oedema is seen locally, in a non-vascular distribution in subepicardial or mid-myocardial location. Occasionally, the oedema is diffuse and can only be detected by quantification of the T2 value, particularly when it is two standard deviations above that of skeletal muscle [24]; (2) early gadolinium enhancement, i.e. T1-weighted (T1-W) fast-spin echo images taken 1–2 mins after the injection of contrast agent, which shows enhancement in acute myocarditis due to hyperemia and capillary leak; and (3) delayed enhanced images taken 10–15 minutes after the injection of contrast agent, which shows enhancement due to fibrosis and necrosis [24‐26], thus representing irreversible myocardial injury. The overall sensitivity, specific, accuracy, positive predictive value, and negative predictive value of MRI for myocarditis is 59 %, 86 %, 68 %, 89 %, and 53 %, respectively (24). On ¹⁸F-FDG PET scans, myocarditis presents as diffuse increased metabolic activity with areas of heterogeneity. Increased metabolic activity in myocarditis is secondary to microvascular and myocyte damage as well as changes in fatty acid metabolism [23]. Increased metabolic activity in myocardium is, however, nonspecific and can also be seen in congestive heart failure, right ventricular strain, and hypertrophy due to pulmonary hypertension [26].

The abnormalities in PET and MRI are based on different mechanisms, and hence they may provide complementary information in myocarditis. Other than functional information on the myocardial activity from MRI, PET can demonstrate activity of the inflammatory process in the myocardium (Fig. 11). MRI also provides functional information of the ventricles and valves. Since myocarditis usually affects younger patients in comparison to neoplastic cardiac masses, an argument could be made for the use of PET/MR as compared to PET/CT due to lower radiation exposure. PET and MRI can also be used for assessment of the response to medical treatment. However, other than short case reports, there is still lack of published reports on use of PET/MR in identification and follow-up of patients with myocarditis [27].

Sarcoidosis is a systemic granulomatous disease of unknown aetiology, likely due to an alteration in the immune response after exposure to an environmental, occupational, or infectious agent in genetically susceptible individuals [28]. Pathologic hallmark of sarcoidosis is non-caseating granuloma [29]. On MRI, acute phase of sarcoidosis manifests with myocardial thickening and oedema. Patchy myocardial oedema is seen due to active inflammation and is more commonly seen in the subepicardial region and less commonly in the mid-myocardial region. The oedema corresponds to areas of active inflammation, may present with ventricular arrhythmia or conduction disturbance, and usually responds to steroid therapy [30]. In the chronic phase, there may be wall thinning. Delayed enhancement is seen in up to 50 % of patients, in a mid-myocardial or subepicardial distribution due to replacement fibrosis [31]. Sarcoidosis often involves the septum (particularly basal portion) and left ventricular wall with rare involvement of the right ventricle and papillary muscles [32]. Wall motion abnormalities are also seen.

¹⁸F-FDG PET is also useful in the identification of cardiac sarcoidosis, with a sensitivity and specificity of 89 % and 78 %, respectively [29]. Recently, published data support the growing role of ¹⁸F-FDG PET-CT in the diagnosis, risk stratification, and assessment of response to treatment of patients with cardiac sarcoidosis [4].

With PET/MR, the MR component evaluates for active inflammation, wall motion abnormalities, and fibrosis. FDG uptake indicates inflammation (Fig. 12). Since the disease detection is based on different mechanisms, the information obtained may be complementary, resulting in possible improved sensitivity in disease detection. PET/MRI also has the advantage of providing valuable information on the exact anatomical extension of sarcoidosis in the cardiovascular system, functional assessment of the myocardium as well as assessment of the disease activity and response to treatment (Fig. 12). This can help to improve the selection of available treatment options including medical management and cardiac resynchronization therapy [33].

PET/MR is also useful in other inflammatory disorders, such as Erdheim Chester disease (Fig. 13), which is a rare disorder characterized by abnormal proliferation of non-Langerhans histiocytes, resulting in fibrosis that affects multiple organs.

Inflammation of the pericardium can have diverse aetiologies including idiopathic, autoimmune diseases, infections, post myocardial infarction, uremia, and radiation. Pericarditis can be of acute, chronic inflammatory, or chronic fibrosing types. MRI has become an important modality in the evaluation of pericardial disease. In acute pericarditis, there is pericardial thickening, pericardial effusion (Fig. 14a), and pericardial inflammation, which is manifested as delayed enhancement. In the chronic phase, pericardial thickening is present, but with lower amount of effusion and inflammation than acute type. In chronic fibrosing type, there is pericardial thickening with/without calcification and features of pericardial constriction may be seen. MRI is valuable in the evaluation of pericardial constriction, since it shows features of ventricular interdependence such as exaggerated septal flattening in real-time imaging (Movie 1), diastolic septal bounce, and abrupt cessation of diastolic filling. MRI is increasingly being used in the evaluation of pericardial inflammation, particularly in the context of transient pericardial constriction, which may be seen in the acute or subacute phase of pericarditis due to impaired pericardial distensibility. Although the standard treatment of pericardial constriction is aggressive pericardial stripping, in transient pericardial constriction, anti-inflammatory therapy (NSAIDs, colchicine, steroids) may be beneficial [34]. The presence of inflammation on MRI, even in the absence of clinical and serologic evidence of inflammation warrants continued therapy [34].

Inflammatory processes involving the pericardium demonstrate a mild-to-moderate FDG uptake (Fig. 14b), but occasionally no FDG uptake can be seen. In contrast, markedly increased FDG uptake is seen in proliferative neoplastic disorders that often corresponds with a localized mass [32]. A preliminary study showed that the FDG uptake is strong in tuberculous pericarditis compared to mild to moderate uptake in idiopathic pericarditis. In addition, hypermetabolic lymph nodes were also seen in tuberculous pericarditis [35]. Other causes of pericarditis can be distinguished based on history and FDG uptake patterns. Following surgery, there is diffuse and mild uptake and following radiation, there is diffuse uptake corresponding to the radiation port. Chemotherapy pericarditis is homogeneous and diffuse. Chronic pericarditis shows thickening and diffuse mild uptake with small or no effusion [35].

Although most cases of pericarditis do not require PET/MR, unique information provided by PET can be superimposed to MRI or CT and thus help to improve identification of inflammatory processes or masses in uncertain cases, to exclude other infectious foci, and aid in evaluation of response to therapy.

Valve abscess is considered a complication of infective endocarditis (IE), either in a native or prosthetic valve, and requires urgent surgery with debridement of all infected and necrotic tissues. Echocardiography identifies only 40 % of surgically proven myocardial abscesses [35]. On MRI, morphological and functional evaluation of the valve can be performed along with quantification of valvular lesions such as regurgitation or stenosis. Inflammation or abscess is seen as soft tissue surrounding the valve. Inflammatory tissue has high signal on T2-weighted sequences (Fig. 15a) and abscesses show fluid signal with peripheral enhancement.

PET-CT has been shown to have higher sensitivity than echocardiography in the detection of valve abscess, showing intense uptake in the affected valves (Fig. 15b) [7]. Hybrid PET/MR imaging can aid in the characterization and localization of the high uptake zone on PET due to the excellent spatial resolution capabilities of MRI. The tissue characterization of MRI can also help to improve assessment of post-surgical cases in which high uptake can be expected due to post-surgical changes or residual/recurrent abscess. Information obtained from hybrid PET/MR can also achieve better differentiation of valve abscess from infective valve vegetations. Although there are some case series on the diagnostic value of PET-CT [36], there is no published report on the additional value of PET/MR in identification and management of cardiac valve abscesses [36]. MR or PET/MRI in patients with prosthetic valves can be challenging due to susceptibility artefacts. Sequences with lower susceptibility artefacts such as a conventional gradient-echo than a steady-state free precession (SSFP) should be used.

Vasculitis is inflammation of the vascular wall and can be classified based on the size of the affected vessels, location of vessels, and underlying cause. Aetiologies for vasculitis include autoimmune diseases, infection, and post chemotherapy [37]. Signs and symptoms are primarily dependent on the affected vessels and severity of disease. Although definitive diagnosis is established using biopsy, inflammatory markers and imaging are also helpful in diagnosis.

On MRI, active large vessel vasculitis presents with wall thickening, which may show oedema on STIR and enhancement with contrast. The oedema seen on STIR images, however does not show a consistent relationship with disease activity [38]. MR angiography provides comprehensive evaluation of the vasculature, able to identify abnormalities such as aneurysms and stenosis, and can be used to screen high-risk patients without any radiation exposure. MRI has been shown to have a sensitivity of 80 % and specificity of 90 % in the diagnosis of giant cell arteritis [39]. Since ¹⁸F-FDG accumulates in the macrophage-rich areas, vasculitis is seen in PET with high uptake in the vascular wall in a linear pattern. However, the tracer uptake for inflammatory disease is lower than for oncologic disease [40]. Also, PET/CT lacks the capability to show the vessel wall in the low-dose non-contrast CT. FDG-PET.CT has been shown to have an overall sensitivity and specificity of 80 % and 89 % for the diagnosis of giant cell arteritis when compared to reference clinical criteria [39].

PET/MR combines the anatomic information provided with MRI with the metabolic activity of PET, providing a reliable diagnosis of vasculitis with lower radiation dose than PET/CT, also at the same time enabling assessment of aneurysms and stenosis in the vasculature. In a pilot study on 14 consecutive examinations, PET/MR was found to be efficient in the evaluation of large vessel vasculitis and that only a combination of PET and MRI revealed the whole disease extent (Fig. 16). In MRI, abnormal segments were seen in 32/149, and in FDG 78/149, for an overall number of 83/149. Also, there was a strong and significant correlation between CRP and the number of vessel segments affected by inflammation in PET/MR (r = 0.86, p = 0.01) in contrast to PET or MRI only. Thus, PET and MRI are complementary for diagnosis and likely to be more accurate to assess the true extent of disease than either modality alone [41]. While the disease activity can be evaluated more accurately with PET, MRA provides a comprehensive evaluation of the vascular tree without the need for ionizing radiation [42].

Atherosclerosis is a diffuse, chronic inflammatory disorder characterized by deposition of lipid and fibrous products in the vessel wall, following complex biological processes. Current imaging techniques focuses on the detection of stenosis, grading of stenosis and evaluation of perfusion abnormalities distal to the stenosis. However, major events such as myocardial infarction (MI) and stroke are not produced by plaques with associated stenosis, but by non-stenotic plaques that get disrupted, the so-called vulnerable plaque. Hence, it is vital to be able to detect, quantify, and characterize the biological activity and stability of plaque at a subclinical stage. Identification of these high-risk features along with plaque burden will significantly help in the diagnosis and management of these patients [43].

High-resolution multi-contrast imaging of the vessel wall (T1-w, T2-w, PD, TOF) and dynamic contrast-enhanced angiography can characterize the specific tissue components of complex plaques, which helps to delineate biologic activity (Table 4) [43]. High-risk features of an atherosclerotic plaque include a large lipid-rich necrotic core, thin fibrous cap, endothelial denudation with superficial platelet aggregation, fissured/injured plaque and active inflammation. Stenosis > 90 %, superficial calcified nodule, intraplaque haemorrhage, outward positive remodelling and adventitial inflammation/neovascularity are the other features of vulnerability. Flow can be evaluated using time-of-flight sequence. Due to the smaller size, evaluation of coronary arteries is challenging, but this technique has been well established in the evaluation of carotid plaque [43].

FDG-PET has also been shown to be useful for assessment of inflammatory activity within plaque since macrophages in an inflamed plaque show avid FDG uptake and accumulation. High uptake has been shown both in symptomatic and asymptomatic plaques [44]. It has also been shown that there is a correlation between FDG uptake and extent of macrophage infiltration in a vulnerable carotid plaque [45]. However, low spatial resolution of clinical PET machines limits identification of small plaques and localization of high FDG uptake zones.

With PET/MR, the high spatial resolution of MRI helps in localizing the metabolic activity obtained from PET (Fig. 17) [46]. Simultaneous acquisition also results in perfect spatial alignment. In addition, MRI also provides additional information on plaque components as described above [43]. Thus, the biological features of plaque obtained by PET provides additional information with multi-parametric plaque morphology assessment with MRI. The lumen can also be assessed at the same time for stenosis/occlusion. There is also potential for using targeted PET isotopes, which hone in on specific molecular targets of atherosclerosis, such as macrophage, or receptors of angiogenesis, vascular adhesion, and apoptosis. Some potential agents include 68 Ga-DOTATATE and 68 Ga- NOTA-RGD [46]. A recent study in HIV patients found congruence between PET/MR and PET/CT SUV values and improved delineation of outer and inner wall of carotid artery from MRI [47, 48]. MRI can be used for morphological follow-up and the combination can be used in drug trials.

Myocardial ischemia occurs when the oxygen requirement of part of myocardium exceeds the oxygen supply derived by coronary arteries, typically caused by coronary atherosclerotic disease. Identification of myocardial ischemia is essential in patients with coronary artery disease for management and for prognosis.

MRI is very useful in the evaluation of myocardial ischemia. Dynamic first-pass perfusion MRI performed at rest and stress has been shown to have high accuracy in the evaluation of myocardial ischemia. The CE-MARC trial showed sensitivity (86.5 %) and negative predictive value (90.5 %) of MRI to be statistically superior to that of SPECT (66.5 % and 79.1 %, respectively), while the specificity (83.4 % vs 82.6 %) and PPV (77.2 % vs 71.4 %) were non-statistically significantly superior [49]. Myocardial ischemia is seen as a hypointense perfusion defect at first pass perfusion imaging only at stress, with normal perfusion at rest.

Rubidium-82, Nitrogen-13 ammonia and 15 O labelled water are the PET isotopes used for perfusion imaging. PET has high accuracy, with sensitivity and specificity of 90 % in detection of obstructive coronary disease [50]. In addition, PET imaging is also valuable in quantification of myocardial blood flow (MBF) and coronary flow reserve (CFR), which is useful in multi-vessel disease, early changes of vasoreactivity and follow-up [9]. Also, PET findings have prognostic value [51].

With an integrated PET/MR system, simultaneous 13 N ammonia PET-MR myocardial perfusion has the advantage of increasing the accuracy of MR perfusion by the PET component, since the N-13 ammonia perfusion imaging has more coverage than stress perfusion MR. 13 N-ammonia PET has been shown to be an important cost-effective strategy in clinical decision making [52]. Utilizing MR attenuation correction data eliminates the radiation associated with CT attenuation data. Also, the examination time is shortened, since rest-phase 13 N-ammonia images are not required and are replaced by delayed enhancement images. In addition, MRI can also help address artefacts caused by motion or partial volume averaging. Ventricular function, global and regional, can be evaluated in the same study. A study which compared MRI with PET showed high accuracy of MRI in evaluating pathology with sensitivity of 91 % and specificity of 94 % compared to PET scan, but with significant underestimation of CFR [53]. Challenges of perfusion PET include lack of a good flow tracer for PET MPI, and short half-life of the isotopes, which necessitates having a cyclotron on site.

Myocardial infarction results from death of myocardial tissue due to blockage of coronary arteries. Imaging is useful in the diagnosis, prognosis, and evaluation of complications. An important contribution of imaging is detection of viable myocardium, which is dysfunctional but not dead, and hence can recover its function following revascularization.

Delayed enhancement imaging is highly sensitive and specific in the evaluation of myocardial infarct, both in acute (sensitivity of 99 %) and chronic phase (sensitivity 94 %) [53]. MI is manifest as a subendocardial or transmural pattern of enhancement confined to a vascular territorial distribution [54]. MRI is accurate in the quantification of myocardial function and scar, which has prognostic value. Dysfunctional segments with scarring > 50 % of myocardial thickness have been shown to have lower probability of functional recovery after revascularization [55]. MRI can also evaluate several complications of MI including thrombus, rupture, ventricular septal defect, pericarditis, etc. MRI has prognostic value and can also provide functional information [56]. The difference between oedema seen in T2-weighted images and the scar seen in delayed enhancement is considered the salvageable area by myocardial reperfusion [57].

PET is also very reliable in the evaluation of myocardial viability, quantifying the degree of viability [58]. With PET, myocardial viability is detected by using a combination of perfusion imaging and FDG study [58]. Infarcted scarred tissue does not have ¹⁸F-FDG uptake (Fig. 18), while a viable hibernating myocardium has ¹⁸F-FDG uptake (Fig. 19). Hence, a matched defect, i.e. a defect in perfusion imaging and FDG scan, is indicative of a scar, while a mismatched defect, i.e. a defect in perfusion imaging, but normal FDG uptake, is indicative of hibernating myocardium [59].

A study using FDG and 13 NH3 PET and MRI showed that MRI can depict areas of non-transmural enhancement in viable areas of PET due to higher spatial resolution [60]. There was good correlation between the modalities for location and extent of infarct. Both modalities have comparable PPV for functional recovery after revascularization using 50 % thickness enhancement and 50 % FDG activity as cut off [58]. Both these techniques have high sensitivity but low specificity (63 %) in predicting functional recovery after revascularization, indicating that segments called viable do not always improve [61].

PET-MRI can also evaluate and characterize the heterogeneity of the scar tissue, since this is a marker for inducible monomorphic ventricular tachycardia [53]. The combination of PET SUV maps, metabolic function, delayed enhancement and T1 mapping may be better predictor of arrhythmia susceptibility than MR alone. The combination of PET and MRI can also assess other features such as decreased wall thickening, regional functionional abnormalities, lack of contractile response to low-dose dobutamine, preserved sub-epicardial myocardium, and preserved perfusion [9]. Both modalities provide prognostic information, although FDG-PET depends on mismatch between flow and glucose uptake, and MRI depends on infarct size.

Cardiac MRI is considered the gold standard in evaluation of myocardial function [60]. PET scan has significantly lower spatial and temporal resolution than MRI, limiting its accuracy in evaluation of cardiac function. Left ventricular function is calculated by algorithm-dependent assumptions on endocardial and epicardial border. Hence, the MRI component of PET/MR provides more accurate quantification of ventricular function.

MRI also provides morphological information on the coronary arterial anatomy, including luminal stenosis [62]. Although inferior to CT, MRI can also be used for the evaluation of the anatomy of coronary arteries and evaluate coronary artery disease, with reasonably good accuracy. A recent multicenter trial showed sensitivity of 88 %, specificity of 72 %, PPV of 71 %, and NPV of 79 % compared to coronary angiography [62]. MR coronary angiography is performed using a 3d- SSFP sequence with navigator gating, images acquired during end-expiration and in diastole, with myocardial suppression by T2 prepared sequence. Thus, MRI provides concomitant anatomic information.