We Only See What We Look For: Imaging Cardiac Inflammation

Myocarditis is responsible for a substantial minority of cases of sudden death and nonischemic dilated cardiomyopathy (DCM). In clinical case series of sudden death, myocarditis is often the 3^rd leading cause after hypertrophic cardiomyopathy and congenital and atherosclerotic coronary artery disease.¹ In autopsy studies of young adults, myocarditis is responsible for 5 to 20% of sudden deaths.² Nine to 40% of acute nonischemic cardiomyopathy cases are due to myocarditis, a rate that varies depending on the histological or clinical diagnostic criteria used.³ A recent analysis based on hospital dismissal ICD-9 codes estimated that between 0.5 and 4% of prevalent heart failure is due to myocarditis.⁴ This value is likely an underestimate as the diagnostic test, endomyocardial biopsy (EMB), is not widely performed outside of tertiary care medical centers.

Timely diagnosis of myocarditis can refine estimates of cardiac risk and permit tailored treatment for heart failure. Several studies suggest that more severe myocardial inflammation predicts progression to DCM, increased risk of arrhythmias and chronic heart failure.^5-7 However, the cost, risks, and lack of availability of EMB limit its widespread clinical use. For epidemiological studies, diagnostic criteria that rely on clinical syndromes, biomarkers, and/or imaging abnormalities have been proposed but these criteria sacrifice diagnostic certainty.⁸ There are no clinically available and highly specific biomarkers or imaging tests that identify the histological type or quantify the severity of myocardial inflammation.

Past efforts to noninvasively diagnose myocarditis utilized various features of myocardial damage related to the inflammatory response. In the late 1980's, anti-myosin scintigraphy to detect myocyte necrosis had a high sensitivity but low specificity for myocarditis.⁹ Echocardiography established the prognostic significance of left and right ventricular failure, but never challenged the histological gold standard. Imaging has been limited in part because most cases of clinical myocarditis have a short duration or minor degree of myocyte damage. Cardiac magnetic resonance imaging (MRI) using a combination of T2, early T1 and late T1 sequences achieved a sensitivity and specificity of about 80%, but only in recent onset cases.¹⁰ Although epicardial and/or mid-myocardial patterns of signal abnormality suggest a post-viral myocarditis or nonischemic scar, specific causes and cellular types such as giant cell or eosinophilic myocarditis cannot be identified by MRI or even ¹⁸FDG- positron emission tomography.

In this issue of Circulation: Cardiovascular Imaging, the study by van Heeswijk et al overcomes previous challenges by demonstrating that intravenous injection of perfluorocarbons (PFCs) are taken up by inflammatory cells in the heart and can be detected using 9.4T cardiac MRI in a male mouse model of autoimmune myocarditis ¹¹. If this method can be successfully applied to clinical disease, it could allow for widespread and earlier detection and a better understanding of the role of inflammation in the development of chronic DCM. Interestingly, PFCs were only detected in macrophages, granulocytes and dendritic cells, but not lymphocytes. Yet, the signal from the PFCs correlated with the severity of myocardial inflammation. Although studies of myocarditis have focused on the role of lymphocytes in disease pathogenesis, and specifically on T helper responses, the majority of cells present during acute myocarditis in autoimmune and viral animal models of myocarditis are monocyte/macrophages and granulocytes.^{12, 13} For this reason, the technique presented by van Heeswijk et al holds particular promise for clinical diagnosis.

Quite a few issues remain to be addressed before ¹⁹F imaging will prove clinically useful. Can PFCs, in an experimental setting, detect inflammation in female mice, who have more lymphocytes, as well as it does in males who have far more severe inflammation comprised primarily of macrophages and granulocytes ?^{12, 14} Will this method work as well in other strains of mice or differ between patients susceptible or resistant to chronic DCM? The BALB/c mice used in these studies are susceptible to develop DCM, but may accurately model only one type of myocarditis patient. Timing for injection of PFCs may also be important. Some mouse models of myocarditis are biphasic, with inflammation disappearing quickly after peak inflammation. If this also occurs in patients with myocarditis, timing of PCF injection may be critical. Ligands that can be detected using cardiac MRI or PET imaging will need to detect inflammation in more than one phase of the disease to be most useful as a diagnostic tool.

Questions that will arise early in the clinical arena include the impact of co-existing conditions such as pericarditis and prior ischemic myocardial damage that may lower the specificity of ¹⁹F imaging for myocarditis. Common clinical 1.5T or 3T magnets may have lower spatial resolution for ¹⁹F than the 9.4T research magnet used by van Heeswijk et al. The risks of ¹⁹F-PFC exposure in humans are not fully known. Recognizing these hurdles, the promise of a widely applicable and sensitive noninvasive test for myocardial inflammation is exciting and worth the attention of the entire myocarditis research community. Hopefully, the foundation provided by Heeswijk et al will open the door to an enduring stream of translational and clinical investigations.