Imaging myocardial inflammation by CMR mapping: good getting better?

In this issue of the European Heart Journal: Cardiovascular Imaging, Luetkens et al.¹ publish data on the diagnostic performance of cardiovascular magnetic resonance (CMR) tissue markers, specifically myocardial magnetic relaxation times, in detecting acute myocarditis. In 34 patients with clinical evidence for acute myocarditis and 50 controls, the diagnostic accuracy was found to be excellent for both current CMR markers (‘Lake Louise criteria’²) and myocardial relaxation times T1 and T2 (including the extracellular volume fraction, derived from post-contrast T1). The Lake Louise criteria yielded a diagnostic accuracy of 92% [sensitivity 82%, specificity 98%, area under the curve (AUC) 0.90]. Diagnostic accuracies of 96% were also achieved by combining relaxation times with high-signal-intensity areas in late gadolinium enhancement (LGE) images. Albeit not sensitive itself, the addition of longitudinal strain to native T1 and T2 also showed very good accuracy (sensitivity 91%, specificity 100%, accuracy 96%, Positive predictive value 100%, Negative predictive value 94%, AUC 0.99), without using contrast-enhanced images at all.

CMR is the prime diagnostic tool for non-invasively diagnosing acute myocarditis, which is also one of the most frequent indications for CMR,³ reflecting the unique ability of CMR to characterize myocardial pathology in vivo.⁴ It is especially important in patients with non-ischaemic acute myocardial injury, where myocarditis is frequently identified.^5,6 While their diagnostic value is very good when combined, some of the Lake Louise criteria have limitations: early gadolinium enhancement (EGE) images suffer from inconsistent quality and post-bolus timing issues, and T2-weighted signal intensity is subject to coil profiles and arrhythmia, while high-signal-intensity areas in LGE images, typically only evaluated visually, are subject to artefacts and cannot differentiate acute from chronic or ‘healed’ myocarditis. Moreover, the need for using skeletal muscle as an internal reference for a quantitative signal intensity analysis can be problematic.^7,8 Finally, contrast agents are required for EGE and LGE, adding complexity to the procedure, a (small) risk of adverse effects, and cost. Therefore, a CMR protocol without the need for reference tissue or contrast agents would represent a significant advance, if diagnostic accuracy could be preserved or even improved.

The paper overlaps with a previous report from the same group,⁹ confirming previous evidence for the utility of native T2¹⁰ and T1 mapping^11,12 in acute myocarditis. The authors also used longitudinal strain as a functional parameter, which added specificity when combined with native T1 and T2.

CMR tissue parameters have to be studied with respect to their specific value for identifying inflammation-specific diagnostic targets (see Table 1). Imaging myocardial oedema by water-sensitive CMR sequences apparently could be replaced by native T2 mapping or T1 mapping, although the sensitivity of T1 mapping to oedema varies between sequences.¹³ T1 is not very specific to oedema, as it is also increased in fibrosis. A recent study by Hinojar et al.¹² indicated that T1 is more increased in acute than in convalescent myocarditis, but we still need confirmative data, especially on the discriminative value in a clinical setting and suitable cut-off values between acute and non-acute inflammation. Previous data have demonstrated a lower sensitivity of the Lake Louise criteria during later or chronic stages of the disease.¹⁴ Therefore, the ability of novel protocols to differentiate acute from chronic disease requires attention. While EGE as a marker for acute inflammation, i.e. an increased contrast inflow due to inflammatory vasodilation (hyperaemia) and increased vascular permeability due to microvascular injury, is not targeted by any of the other markers, omitting this information may not critically affect the diagnostic value of CMR.¹⁵ Furthermore, the study by Luetkens et al. did not include patients with other non-ischaemic cardiomyopathies. The addition of a functional marker such as myocardial strain broadens the spectrum of parameters, although the added/high specificity observed in this cohort may not be reproducible in a clinical environment as other cardiac diseases may easily equally alter strain.

Table 1.

Diagnostic targets and related CMR parameters

Diagnostic target	Lake Louise criteria2	Novel parameters
Function, pericardial effusion	Regional or global dysfunction, pericardial effusion in cine images	Regional or global dysfunction, pericardial effusion in cine images
Oedema	Locally or globally high signal in T2-weighted images, normalized to skeletal muscle	T1 ↑↑ or T2 ↑↑
Hyperaemia	Signal intensity increase in pre- and post-contrast EGE images, normalized to skeletal muscle	–
Necrosis, scar	High signal in post-contrast LGE images (visual assessment)	T1 ↑

Comparison of current CMR criteria (Lake Louise criteria) and novel parameters.

EGE, early gadolinium enhancement: LGE, late gadolinium enhancement.

Before myocardial magnetic relaxation time mapping achieves the status of being ready for prime time, T1 mapping needs further standardization and the following questions need answers: Can a combination of novel markers reliably detect both, acute (oedema, hyperaemia) and chronic injury (necrosis, scar)? Can native T1 ‘do it all’ or do we need T2 mapping? Do we need a functional parameter?

While such research and confirmative data during the course of the disease and in mixed, ‘real-life’ patient cohorts are required, the study adds to a strong body of evidence showing that mapping magnetic relaxation times may replace current criteria and especially, that a contrast-free CMR protocol for advanced myocardial tissue characterization in patients with acute myocarditis is in reach.

There is some work to do, but it seems that a very good diagnostic tool is about to get even better.