Positron emission tomography/MRI for cardiac diseases assessment

Abstract

Functional imaging tools have emerged in the last few decades and are increasingly used to assess the function of the human heart in vivo. Positron emission tomography (PET) is used to evaluate myocardial metabolism and blood flow. Magnetic resonance imaging (MRI) is an essential tool for morphological and functional evaluation of the heart. In cardiology, PET is successfully combined with CT for hybrid cardiac imaging. The effective integration of two imaging modalities allows simultaneous data acquisition combining functional, structural and molecular imaging. After PET/CT has been successfully accepted for clinical practices, hybrid PET/MRI is launched. This review elaborates the current evidence of PET/MRI in cardiovascular imaging and its expected clinical applications for a comprehensive assessment of cardiovascular diseases while highlighting the advantages and limitations of this hybrid imaging approach.

Introduction

Positron emission tomography (PET) is a functional imaging modality that uses several PET tracers to target perfusion, metabolism, innervation and inflammatory conditions.¹ Magnetic resonance imaging (MRI) comprehensively performs cardiac morphological and volumetric imaging without radiation exposure.² The hybrid PET/MRI is a novel instrument that combines MRI sequences with functional PET information in a single scan.³ Simultaneous data acquisition in PET/MRI allows for non-invasive quantification of cardiac perfusion and metabolism, along with anatomical assessment and tissue characterisation in various cardiac diseases, some of which are the leading causes of death and disease burden around the world.⁴ However, this advanced hybrid imaging tool has limitations, and its clinical utility remains to be determined. This review discusses the principles of PET and MRI, and the potential clinical applications of hybrid PET/MRI in patients with cardiac diseases.

PET

Advances in nuclear cardiology, including PET, allow for non-invasive diagnosis and risk assessment in cardiac diseases.^5,6 Radiopharmaceutical tracers are used to assess the perfusion, metabolism, innervation and inflammation in various cardiac conditions.¹ In general, the PET tracers are labelled with short-lived positron emitters produced by a cyclotron or generator. PET is applicable in almost all patients due to fewer side effects from tracers and minimal radiation exposure. The radiotracer ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG), a glucose analogue, is the most frequently used tracer for the assessment of tumour, cardiac viability, brain function and also active inflammation.

PET for ischaemic cardiomyopathy

PET tracers used to assess myocardial perfusion include Rubidium-82 (⁸²Rb), nitrogen-13 ammonia (¹³N-NH₃) and oxygen-15 water (¹⁵O-H₂O).⁷ PET myocardial perfusion imaging with ¹³N-NH₃ and ⁸²Rb provides high diagnostic accuracy in the visual detection of coronary artery disease (CAD). Dynamic scanning permits quantitative assessment of myocardial blood flow (MBF) and myocardial flow reserve defined as the ratio of MBF at peak hyperaemia to MBF at rest. These measurements add value in multivessel disease detestation and risk stratification of CAD patients.^8–11

¹⁸F-FDG PET identifies glucose metabolism in the heart and thus myocardial viability. A region with preserved ¹⁸F-FDG uptake indicates the presence of viable myocardium with the ability for glucose uptake via the glucose transporter.¹ Glucose administration with oral loading or an insulin-glucose clamp enhances ¹⁸F-FDG uptake in the viable myocardium.^1,12

PET for non-ischaemic cardiac disease

The activation of granulocytes and macrophages during inflammation enhances the ¹⁸F-FDG uptake. Thus, ¹⁸F-FDG PET is useful for detecting active cardiovascular inflammation due to cardiac sarcoidosis (CS)¹³ and infections.¹⁴ Furthermore, it can aid in the non-invasive differentiation of benign and malignant cardiac tumours due to increased glucose metabolism in most types of malignant tumours.^15,16 In order to accurately evaluate myocardial inflammatory lesions or tumours using ¹⁸F-FDG PET, it is essential to suppress physiological myocardial glucose metabolism by long periods of fasting combined with a low-carbohydrate diet and high-fat diet.^17,18 Several other radiopharmaceutical tracers have been developed recently to evaluate cardiac disease by targeting myocardial perfusion, metabolism, innervation and inflammation.¹

Several PET tracers have been developed recently to evaluate cardiac disease by targeting metabolism and inflammation, which can be visible on PET/MRI in the future.¹ Myocardial oxidative metabolism can be evaluated by ¹¹C-acetate.¹⁹ Previously, ¹¹C-palmitate and fluorine-18-labelled fluoro-6-thia-heptadecanoic acid have been used to evaluate fatty acid metabolism.^20,21 Fluorine-18 fluorothymidine, which is a promising PET tracer for evaluating tumour proliferative activity,²² and fluorine-18 fluoromisonidazole, which is one of the hypoxia tracers, might have the potential to further improve specificity with regard to diagnosis of CS because they had minimal physiological uptake unlike ¹⁸F-FDG.²³ In the near future, these PET tracers may be used clinically in cardiac PET/MRI.

MRI

MRI for ischaemic cardiomyopathy

Cardiac magnetic resonance (CMR) is a valuable tool for evaluating patients with proven or suspected CAD. CMR is widely accepted as the gold standard for cardiac morphological evaluation and volume quantification. Gadolinium, the contrast agent widely used in MRI, displays the characteristic late gadolinium enhancement (LGE) in myocardial fibrosis. The physiological basis of LGE is based on the increased extracellular distribution volume of the contrast agent and a prolonged wash-out related to the decreased capillary density within the myocardial fibrotic tissue.²⁴ The increase in gadolinium concentration within the fibrotic tissue appears as bright signal intensity due to the T₁ shortening based on conventional inversion recovery gradient-echo sequences.²⁵ When hyperenhancement is present, its pattern needs to be examined. Subendocardial or transmural hyperenhancement is consistent with the presence of CAD. Thus, the pattern of LGE in the myocardium can help determine the underlying aetiology of heart failure.²⁶ The extent of scarring throughout the wall, as well as the wall thickness assessed by LGE, is predictive of functional recovery in patients with CAD.²⁷ The presence and extent of LGE can determine prognosis in both ischaemic and non-ischaemic cardiomyopathies.²⁶ Myocardial perfusion imaging with stress test has been used for the detection of functional ischaemia.²⁸ The Clinical Evaluation of Magnetic Resonance Imaging in Coronary heart disease (CE-MARC) study reported that CMR is increasingly recognised as a tool with high diagnostic accuracy and prognostic value for the assessment of myocardial ischaemia in both single-vessel and multivessel coronary disease.²⁹

MRI for non-ischaemic cardiac disease

LGE is widely used for diagnosis of non-ischaemic cardiomyopathy and as a predictor of major adverse cardiac events. Isolated mid-wall or epicardial LGE hyperenhancement patterns strongly suggest a ‘non-ischaemic’ aetiology.³⁰ For example, in CS, the lesions are commonly localised in the septal, basal, and lateral segments of the left ventricle and papillary muscles, with relative sparing of the subendocardium.³¹ Aortic stenosis is one of the most common valvular diseases that is characterised not only by progressive valve obstruction, but also by the left ventricular (LV) remodelling response with positive LGE, which is associated with adverse outcomes.³²

Tissue characterisation techniques such as T₁, T₂, and T₂* mapping have validated use in identifying myocardial pathology.³³ CMR detects oedema and inflammation of CS with additional T₂ weighted imaging and T₂ mapping.³⁴ Assomull et al showed that the presence of myocardial LGE was associated with a threefold increase in heart failure hospitalisation or cardiac death, and a fivefold increase in sudden cardiac death or ventricular arrhythmias in non-ischaemic dilated cardiomyopathy (DCM).³⁵ However, only 30% of individuals with DCM show positive LGE,²⁶ probably due to the difficulty in the detection of diffuse fibrosis. T₁ mapping can be more sensitive to detect myocardial abnormalities, especially for diffuse fibrosis, since it enables direct myocardial signal quantification (in milliseconds) on a standardised scale, and allows better characterisation of myocardial tissue composition on a global or regional level.²⁵ Hence, tissue characterisation of the myocardium with multiple imaging parameters enables accurate diagnosis.

The myocardium volume is divided among two main compartments: the ‘intracellular volume’, dominated by myocytes and including other cells such as fibroblasts, circulating red blood cells; and the ‘extracellular volume’ dominated by water associated with the extracellular matrix and the intracapillary plasma volume.³⁶ Extracellular volume (ECV) is also recommended for detection of myocardial fibrosis in the 2013 T₁ mapping consensus document.³⁶ T₁ mapping and ECV method are advantageous for the early diagnosis, quantification, and understanding of underlying pathophysiological processes involved in cardiac amyloidosis.^37,38 Cardiac amyloidosis, especially from wild-type transthyretin-related cardiac amyloidosis, may be prevalent in older male patients with aortic stenosis and promote increased risk of mortality. The importance of screening for cardiac amyloidosis in older patients with aortic stenosis and optimal treatment strategies in those with cardiac amyloidosis warrant further investigation, especially in the era of transcatheter aortic valve implantation.³⁹ Both high native T₁ and ECV values with cardiac MRI are specific markers and useful for prognosis in patients with cardiac amyloidosis.⁴⁰

PET/MRI principles

Pros and cons of PET/MRI compared to PET/CT

PET/MRI is an emerging technology that combines the inherent advantages of MRI, which offers several potential advantages over PET/CT, such as high tissue contrast, improved motion correction, and no exposure to ionising radiation with simultaneous functional information from PET as described above.^41,42 Despite the above advantages, hybrid PET/MRI has several limitations such as^43,44 high initial capital cost, limited availability, and longer acquisition time (up to one hour) compared with PET/CT. Protocols and indications for the use of PET/MRI in cardiac disease are limited. Radiologists and technologists with knowledge in both PET and MRI are required to obtain the appropriate scans. Qualitative and quantitative accuracy require correction for attenuation, cardiac, and respiratory motions in the hybrid PET/MRI system. Mismatched PET and attenuation correction (AC) data could be the cause of the artefact. A novel free-breathing radial gradient-recalled echo sequence has been proposed to improve these issues.⁴⁵ However, metallic implants such as prosthetic valves and stents cause severe artefacts and remain to be a major problem to overcome, especially with high magnetic field (3 Tesla) MRI.⁴⁶ Pulmonary disease evaluation is limited with PET/MRI and this technology cannot be applied to patients with implantable devices (Table 1).

Table 1.

Advantages and disadvantages of PET/MRI compared to PET/CT

Advantages
Lesser radiation dose Improved soft-tissue contrast The gold standard for quantifying cardiac volume with MRI Better motion correction due to the simultaneous scan Increased time available to collect PET data
Disadvantages
High initial capital cost Limited availability Longer acquisition time (up to one hour) Less well-defined protocol and indication Requires radiologist and technologist with knowledge in both PET and MRI Qualitative and quantitative accuracy still being determined Limited role in pulmonary disease Not applicable in patients with implantable devices Artefacts due to magnetic resonance attenuation correction maps

PET, positron emission tomography; MRI, magnetic resonance imaging; CT, computed tomography

The PET/MRI system

Current PET/MRI systems are broadly classified into three different types¹: tandem configuration of PET and MRI scanners,² PET gantry inserted inside the bore of a standard MRI scanner, and³ full integration of PET and MRI scanners in a single gantry⁴⁷ (Figure 1). In a precise sense, simultaneous PET and MRI acquisition can be performed by types² and.³

Figure 1.

PET/MRI systems can be the separated PET and MR system (a) or the integrated PET/MRI (b). The latter is classified into two different types, where the PET gantry is inserted in the bore of a standard MRI scanner (c) or the PET and MRI are fully integrated into a single gantry (d).

PET, positron emission tomography; MRI, magnetic resonance imaging; RF, radiofrequency

PET/MRI detector

The strong electromagnetic field of MRI can interfere with the photomultiplier tubes (PMTs) in conventional PET detectors.⁴⁸ PMT detector can be used only in the integrated PET/MR system when the PET and MRI units are placed separately.⁴⁹ Therefore, novel detectors are required for integrated PET/MRI. Detectors that are insensitive to the magnetic fields such as avalanche photodiodes and silicon photon multipliers have been proposed to replace the conventional PMT detector.^50–53

Attenuation correction

The emission data from a PET scan must be corrected for attenuation to quantify tracer uptake. PET/CT systems use the μ-maps from CT data for AC. Since the hounsfield units (HU) correlate with tissue density, the linear attenuation coefficients can be transformed directly for the PET/CT. On the other hand, PET/MRI experiences challenge with AC because MRI data are associated with proton density or relaxation properties that are not linearly correlated with tissue density. Innovative approaches to create AC maps from MRI data include the segmentation-based methods,⁵⁴ atlas registration and pattern recognition methods,⁵⁵ and joint estimation-based methods⁵⁶ for the integrated PET/MRI.⁴⁷ The segmentation-based methods are used for cardiac lesions in which MRIs are categorised as three-tissue (soft-tissue, lungs, and air⁵⁷) or four-tissue (water-equivalent, fat, lungs, and air⁵⁴) groups based on the MRI intensity. Vontobel et al showed that myocardial ¹⁸F-FDG uptake based on four-tissue group-based segmentation and MR-based attenuation is highly comparable to standard CT-based attenuation corrected ¹⁸F-FDG PET.⁵⁸ The total effective radiation dose was significantly lower with PET/MRI compared with PET/CT due to the attenuation derived from MRI.⁵⁹

Motion correction

The complex pattern of cardiac motion due to the heartbeat and respiratory displacement causes artefacts and blurring in the PET data. To reduce motion artefact, electrocardiography gating and respiratory gating techniques have been used in PET/CT imaging.^60,61 For the magnetic resonance-based motion estimation, non-rigid registration of dual-gated magnetic resonance images has been proposed in order to correct both heartbeat and respiratory motion.⁶² Simultaneous PET and MRI scans avoid the misalignment from voluntary patient motion or involuntary respiration which causes the retrospective co-registration of the independent scans like PET/CT. Although cardiac motion can be corrected using the PET data itself, magnetic resonance-based respiratory and cardiac motion correction, such as the postreconstruction registration approach, which requires independent motion correction and reconstruction for each frame, and motion-compensated image reconstruction method, which directly incorporates the motion information into the system matrix of the iterative PET reconstruction algorithm, have the possibility to improve the accuracy of the myocardial tracer uptake and enhance diagnostic ability.^63–65

Workflow for PET/MRI

Several PET/MRI workflows exist depending on the target cardiac disease (Figure 2). Parallel acquisition of PET and MRI data is a significant advantage of the integrated PET/MRI. An MRI scan is started with a localiser and AC. A PET scan can be started right afterwards. For perfusion PET imaging, a 5 to 20 min PET acquisition, which is also dependent on the tracer used, is performed in list mode and with electrocardiographic triggering. The time duration of ¹⁸F-FDG PET might be decided according to the scanning duration of the whole MRI protocol. Gadolinium-enhanced images are acquired at 10 to 20 min after injection of the contrast agent. The total scanning session ranges between 30 and 90 min depending on the wide variety of MRI acquisitions, such as cine images, perfusion, LGE, coronary magnetic resonance angiography, and T₁ and T₂ mapping. Recently, multiple approaches have been proposed in order to reduce the acquisition times that are needed to perform a complete cine MRI study.⁶⁶ Compressed sensing has been successfully used in various cardiac MRI reconstructions.⁶⁷ PET data consist of perfusion or ¹⁸F-FDG or both and are simultaneously acquired throughout the scanning session.^45,68–70 Cine and LGE images are recommended for the assessment of myocardial viability or CS (Figure 2a). Other sequences including T₂WI, T₁/T₂ mapping, and rest perfusion images can be added as required dual PET tracers of perfusion, and ¹⁸F-FDG is a choice for the assessment of cardiac disease (Figure 2b).⁷¹ Stress and rest perfusion imaging is a strategy used for the assessment of ischaemic cardiomyopathy (Figure 2c).⁴⁸

Figure 2.

Sample imaging protocols for cardiac PET/MRI SPAIR, spectral attenuated with inversion recovery; LGE, late gadolinium enhancement; FDG, ¹⁸F-fluorodeoxyglucose; PET, positron emission tomography

Preclinical PET/MRI studies

Few reports exist on the benefits of using hybrid PET/MRI in specific clinical cases. Animal experiments suggest combining simultaneous MRI and ¹⁸F-FDG PET acquisition for assessing myocardial viability/infarct myocarditis.^72–74 Barton et al challenged to assess the dynamic changes in both cardiac metabolism and contractile function with continuous ¹⁸F-FDG infusion in healthy pigs.⁷⁵

Clinical PET/MRI studies

Myocardial viability

Hybrid PET/MRI has a significant role in the assessment of viability in ischaemic cardiomyopathy. Information on perfusion (PET or CMR), metabolism, LGE, and contractility can be obtained from PET/MRI (Figure 3). Nensa et al confirmed that ≥24 h of high-fat, low-carbohydrate, and protein-permitted diet in combination with unfractionated heparin results in sufficient suppression of myocardial ¹⁸F-FDG uptake in 84% of patients for the assessment of inflammatory or malignant myocardial disease using PET/MRI.⁷⁶ Priamo et al reported that ¹⁸F-FDG PET brought additional information to MRI and PET/MRI allows the accurate diagnosis of cardiac viability in a small sample of 12 patients.⁷⁷ Marchesseau et al confirmed the concordance of semiquantitative values from PET/MRI and PET/CT.⁷⁸ They reported that standardised uptake value measured with PET/MRI and PET/CT showed significant correlations for each scar (Pearson’s r = 0.75) and remote myocardium (Pearson’s r = 0.71) in their regional assessment. In addition, the ratio of myocardial tissue to LV blood pool measured on the PET/MRI is higher for infarcted myocardium than for remote myocardium similar to PET/CT. Beitzke et al showed a good correlation between PET and CMR for the LV scar extent and little correlation between the degree of transmural scarring by CMR and hibernation by PET in 39 patients with ischaemic heart disease through a comparison of simultaneously acquired viability parameters from MRI and PET with dual-tracer acquisition of ¹³N-ammonia and ¹⁸F-FDG.⁷¹ Kunze et al quantitatively investigated the relationship of the ¹⁸F-FDG uptake, native T1 and ECV from the simultaneously acquired PET/MRI data in the context of a complex tissue state of inflammation, oedema and cellular tissue damage after revascularisation of acute myocardial infarction.⁷⁹

Figure 3.

Viability assessment A patient with stenosis of the left anterior descending and the right coronary arteries was assessed for myocardial viability. Left ventricular short-axis images of MRI LGE (a), fused image of LGE and PET (b), and FDG PET with insulin-clamp method (c) are displayed. Different patterns can be observed as¹; normal physiological ¹⁸F-FDG uptake and no LGE (red arrow),² reduced or absent FDG uptake and transmural LGE (white arrows), and³ subendocardial LGE but preserved FDG uptake lesions (yellow arrow). FDG, ¹⁸F-fluorodeoxyglucose; PET, positron emission tomography; MRI, magnetic resonance imaging; LGE, late gadolinium enhancement

Non-ischaemic cardiomyopathy

There have been several case reports about the utility of PET/MRI in CS.^80–83 A definitive and accurate diagnosis of CS is essential because cardiac involvement is an important prognostic factor for patients with sarcoidosis.^{84 18}F-FDG PET and MRI are used for the non-invasive CS diagnosis (Figure 4).^85,86 Comprehensive imaging with cardiac PET/MRI provides helpful information such as active inflammation, fibrotic scar, and myocardial oedema from ¹⁸F-FDG PET, LGE, and T₂WI on MRI, which guides the therapeutic strategy of CS.⁶⁸ Wisenberg et al compared PET/MRI to PET/CT in the same day protocol with 10 CS patients⁸⁷ and showed that PET/MR provides similar diagnostic data with ¹⁸F-FDG uptake compared to PET/CT. Also, combining both the modalities simplified the decision making due to the imaging focus on different aspects of the disease (MRI preferentially on scar/oedema and ¹⁸F-FDG PET on macrophage-related inflammation).^13,87 Hanneman et al reported no significant difference in the number of positive cases identified by PET/MRI and PET/CT during the evaluation of 10 patients with CS or myocarditis.⁵⁹ However for PET/MRI, the scan time was significantly longer (81.4 ± 14.8 vs 12.0 min, p < 0.001) and total effective radiation dose was significantly lower (6.9 ± 0.6 vs 8.2 ± 1.1 mSv, p = 0.007) compared with PET/CT. ¹⁸F-FDG accumulation showed good agreement with LGE and T₂ hyperintensity in patients with suspected myocarditis.⁸⁸ Nappi et al investigated the feasibility of using PET/MRI for the early detection of cardiac involvement in 13 patients with Anderson-Fabry disease.⁸⁹ Focal LGE indicative of intramyocardial fibrosis and high signal on short T1 inversion recovery MRI associated with focal ¹⁸F-FDG uptake indicative of active inflammation was seen in several patients, making PET/MRI a possible diagnostic tool in Anderson-Fabry disease.

Figure 4.

Cardiac sarcoidosis A patient in her 70s with systemic sarcoidosis was assessed for a cardiac lesion. Left ventricular short-axis images of MRI LGE (a), fused image of LGE and PET (b), and FDG PET with long-fasting and low-carbohydrate food preparations (c) is displayed. LGEs in the septum (red arrow) and papillary muscle (white arrow) are detected. Physiological myocardial FDG uptake was well suppressed. Focal FDG uptakes were detected in the areas with MRI LGE, which indicate active inflammation. FDG, ¹⁸F-fluorodeoxyglucose; PET, positron emission tomography; MRI, magnetic resonance imaging; LGE, late gadolinium enhancement

Several case reports of hypertrophic cardiomyopathy discuss the utility of simultaneous PET/MRI for detecting structural and metabolic characterisations and for a detailed risk assessment of the hypertrophied heart (Figure 5).^90,91

Figure 5.

Hypertrophic cardiomyopathy eft ventricular short-axis images of MRI LGE (a), fused image of LGE and PET (b), and FDG PET with long-fasting and unfractionated heparin injection (c) is displayed. Septal hypertrophy with patchy LGE and slight FDG uptakes (red arrow) are detected, which may be due to the metabolic energy shift or inflammatory response. FDG, ¹⁸F-fluorodeoxyglucose; PET, positron emission tomography; MRI, magnetic resonance imaging; LGE, late gadolinium enhancement

Cardiac tumours

A pilot study of 20 patients with cardiac tumours by Nensa et al showed that the SUVmax in malignant lesions was significantly higher than in non-malignant cases. Besides, combining PET and MR information, including high intensity of T₂WI, contrast enhancement, and cine imaging, improved the diagnosis of malignant and non-malignant disease.⁹² Elsayad et al described a case series of the usefulness of PET/MRI in the diagnostic and management of primary cardiac angiosarcoma.⁹³ PET/MRI imaging accurately identifies and avoids misdiagnosis of the crista terminalis, which is a less-frequent anatomical variation and is sometimes misdiagnosed as a right atrial mass.⁹⁴ Thus, the assessment of metabolic activity with PET together with the MR morphological information may represent a novel modality for the evaluation of cardiac tumours.

Other new tracers

Other tracers have been researched and reported for use in PET/MRI, such as a novel PET tracer that targets the expression of the chemokine receptor four on inflammatory cells after acute myocardial infarction,⁹⁵ the sodium fluoride labelled with fluorine-18 (¹⁸F-NaF) for the imaging of calcification activity,^{96 11}C-meta-hydroxyephedrine for the assessment of cardiac reinnervation after cardiac transplantation,^{97 18}F-Macroflor, a modified polyglucose nanoparticle for the monitoring of macrophage biology,⁹⁸ and myocardial perfusion quantification to assess ischaemic cardiomyopathy.⁹⁹

According to Dweck et al., ¹⁸F-NaF PET has the potential to assess the disease activity of aortic stenosis.¹⁰⁰ Doris et al reported that motion correction on gated PET/MRI could improve the image quality of ¹⁸F-NaF activity in aortic stenosis.⁹⁶ For cardiac amyloidosis, patchy uptake of ¹⁸F-NaF was observed in the myocardium of patients with transthyretin amyloidosis (ATTR) but not in patients with amyloid light-chain (AL); therefore, hybrid ¹⁸F-NaF PET/MR has the potential to aid in the evaluation of cardiac amyloidosis, assessment of cardiac function and the characteristic of LGE, T1 mapping, ECV and differentiation of AL and ATTR within a single accurately coregistered scan.^101,102

Summary

Integrated PET/MRI opens the way for new comprehensive applications and cardiac disease characterisations that are currently investigational. Most existing studies are limited by small sample size. Further large sample studies are warranted for evaluating the role of PET/MRI in cardiac diseases and for identifying clinical indications for this technique compared to the standalone PET and MRI.