Myocarditis: a Whole Heart Involvement Revealed by Cardiac Magnetic Resonance Mapping. A Case-control Study

Purpose: To ascertain if differences in T1 and T2 mapping exist between ventricular wall segments with and without late gadolinium enhancement (LGE) in patients with myocarditis, and to compare these ndings to those of control subjects. Methods: In a case-control design, 22 patients with Cardiac Magnetic Resonance (CMR) evidence of myocarditis and 18 controls with normal CMR were assessed. The workup included: (1) T1 mapping (ShMOLLI); (2) LGE; (3) T2 mapping (SSFP); and (4) the T2 signal intensity of the myocardium divided by that of skeletal muscle (T2 ratio). T1 and T2 mapping of affected (LGE+) and unaffected (LGE-) ventricular segments of cases were compared, as were those of controls versus cases. Results: On within-group comparison, ventricular segments with evidence of late enhancement (LGE+) in cases showed a mean T1 value signicantly different from that of unaffected (LGE-) ventricular walls (1057±30 vs. 1028±48; p=0.0001). On between-group comparison, the mean T1 value of LGE- segments in cases was signicantly different from the mean of the corresponding walls in controls (1028±48 vs. 996±10; p<0.0001). The mean T2 maps of LGE- walls in cases were not statistically different from those of controls (49±4 vs. 49±1; p=0.9229). Conclusions: This case-control study suggests that T1 mapping demonstrates signicant alterations in the myocardium of patients with myocarditis even in the absence of late gadolinium enhancement. Specically, T1 mapping revealed diffuse myocardial involvement not evidenced by LGE imaging. T2 mapping was noncontributory.


Introduction
The current incidence of myocarditis is unknown. 1 The epidemiology of this condition is poorly documented, due to the heterogeneity of clinical presentation and challenging diagnosis.
According to the Dallas criteria, myocarditis is defined histologically by the presence of an inflammatory infiltrate in the myocardium, alongside degenerative and/or necrotic changes in adjacent cardiomyocytes, which differ from the ischemic damage associated with myocardial infarction. 2 Etiologically, it may be secondary to infectious or noninfectious processes.In developed nations, the leading cause of myocarditis is viral infection, while in developing countries the main causes are rheumatic carditis, Chagas disease, and HIV-related. 3ong several methods available for diagnosis, cardiac magnetic resonance (CMR) is the noninvasive modality best able to characterize the inflamed myocardium, demonstrating edema, necrosis, and fibrosis.The Lake Louise Criteria (LLC) for diagnosis of myocarditis on CMR are based on techniques such as T2-weighted imaging, early gadolinium enhancement, and late gadolinium enhancement (LGE). 4However, these CMR sequences have some limitations, such as an inability to identify diffuse fibrosis and the need for paramagnetic contrast.Another

Pereira et al. T1 and T2 Mapping in Myocarditis
CMR method, the T2 ratio, defined as the T2 signal intensity (SI) of the myocardium divided by that of skeletal muscle, has also been losing ground to newer and more objective imaging techniques.These include T1 mapping, contrast-enhanced T1-weighted imaging, characterization of the extracellular volume (ECV) fraction, and T2 mapping, all of which have demonstrated superior diagnostic accuracy compared to the LLC. 5 Furthermore, T1 and T2 mapping do not require gadolinium contrast, 6 while contrast-enhanced T1-weighted imaging and ECV mapping do.
Considering advances in imaging modalities and confirming the need for change in diagnostic criteria, a proposed update to the LLC was published in December 2018. 5 view of the limitations of the current criteria for CMR diagnosis of myocarditis and given uncertainties surrounding the putative diagnostic superiority of T1 and T2 mapping, as well as the advantage of not requiring gadolinium contrast, we designed this study to test the hypothesis that T1 and T2 map values would be altered both within the myocardial wall segments affected by late gadolinium enhancement (LGE+) and in seemingly unaffected regions (LGE−).Within this context, the objective of the present case-control study was to compare T1 mapping, T2 mapping, and T2 ratio between affected (LGE+) and non-affected (LGE−) wall segments in patients with myocarditis and controls without myocarditis.

Study population profile
This retrospective case-control study included 22 cases with acute myocarditis (age 34 ± 16 years; 13% female) and 18 controls (age 42 ± 12 years; 16% female).Study participants underwent CMR at Hospital Moinhos de Vento, located in Porto Alegre, Rio Grande do Sul, Brazil, between January 2017 and June 2019.Analysis of CMR reports was performed consecutively, based on the date of the scans.
The criteria for inclusion of cases were presence of mesocardial and/or subepicardial LGE, which is currently the gold standard method, in addition to European Society of Cardiology criteria for clinically suspected myocarditis. 7ll cases had (a) symptoms of chest and/or abdominal pain, dyspnea, or palpitations; (b) elevation of cardiac troponin I or T levels > 160 pg/mL; and (c) presence of LGE in the expected anatomical region on CMR.
The criteria for inclusion of controls were (a) symptoms of chest pain, dyspnea, or palpitations; (b) normal/unavailable troponin I or troponin T values; and (c) no evidence of edema, necrosis, fibrosis, or ischemia on CMR.
The ratio of controls to cases was 1:1, and they were matched by age and sex.Exclusion criteria were: CMR demonstrating LGE pattern suggestive of other conditions, such as ischemic cardiomyopathy, hypertrophic cardiomyopathy, idiopathic dilated cardiomyopathy, amyloidosis, aortic stenosis, or pulmonary hypertension; and contraindications to CMR.In all patients, the workup included LGE (132 case wall segments and 108 control wall segments), T1 mapping (132 case wall segments and 108 control wall segments), T2 mapping (130 case wall segments and 108 control wall segments), and T2 ratio ≥ 2:1 (20 affected case wall segments and 21 unaffected case wall segments).
Ventricular walls were analyzed according to the presence or absence of LGE.The T1 and T2 maps of the affected (LGE+) walls of cases were compared with the T1 and T2 maps of the contralateral unaffected (LGE−) walls of the same patients.In addition, the T1 and T2 maps of the LGE− walls of cases (patients with myocarditis) were compared with the T1 and T2 maps of the same walls in non-myocarditis controls.The mean T1 and T2 values of the LGE+ walls, the mean T1 and T2 values of the LGE− walls of the patients with myocarditis, and the mean T1 and T2 values of controls were compared.Figure 1 shows image analysis among myocardial walls.
For analysis of T1 and T2 maps, the average values obtained in controls were considered the reference range for normality.
Ethical approval was granted for all study procedures.As the study was purely observational, there were no physical or biological risks.There was also no personal contact or contact via telephone or social media with the participants.Data analysis was confidential, and the participants' names, addresses, and other contact information were not disclosed.In view of the foregoing and of the impossibility of accessing the participants' contact information in medical records, pursuant to National Health Council Resolution 466/2012, the institutional Research Ethics Committee waived the usual informed consent requirement.The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations were used as a guide for case-control studies.

Cardiovascular magnetic resonance
CMR imaging was performed at 1.5 Tesla in a Siemens Healthcare AERA 45 mT scanner, using an 18-channel coil.Briefly, cine images were obtained in three long-axis sections (four-chamber, three-chamber, two-chamber) and in the short-axis plane, from the base to the apex of the heart.Tissue characterization was performed in a mid-ventricular short-axis view of the left ventricle, with T1 and T2 mapping, turbo spin-echo (TSE) T2-weighted, and short tau inversion recovery (STIR) sequences.LGE images were acquired by sectioning the whole heart, in a manner similar to the cine acquisition along the same axis.For T1 mapping, the shortened modified Look-Locker Inversion recovery (ShMOLLI) acquisition method was used before administration of the contrast agent.LGE images were acquired in the long-and short-axis planes, using a T1weighted phase-sensitive inversion recovery (PSIR) sequence, 10 minutes after intravenous administration of gadobutrol (Gadovist -Bayer, total 0.20 mmol/kg).

Image analysis
Left ventricular volumes and ejection fraction were analyzed on cine images using Argus software (Siemens Medical Solutions).Short-axis T1 and T2 mapping images, T2 ratio, and LGE were subsequently analyzed.On T2weighted dark blood images, edema is diagnosed when the T2 ratio is ≥ 2:1.The ratio was obtained as the T2 SI of the myocardial region of interest (with LGE) divided by the SI of skeletal muscle.T1 and T2 values were obtained from quantitative analysis of all T1 and T2 maps, rather than on visualization of color mapping.When delimiting the endo-and epicardial contours, care was taken to avoid contamination by the ventricular cavity and extramyocardial structures to minimize the partial volume effect on T1/T2 values of the myocardium.In acute myocarditis, identification of remote myocardium can be challenging, because the inflammatory process is often global; thus, a myocardial region without LGE was chosen to represent the myocardium least affected by the disease process, with care also taken to avoid regions of abnormally low SI.Acute myocardial inflammation was considered when the mean T1 value was > 996 ms and the mean T2 value was > 49 ms.Focal areas of LGE were defined as those with a standard deviation of SI ≥ 2.0 above the mean SI of the remote myocardium.To calculate the extent of myocardial injury detected by tissue characterization techniques, the percentage of abnormal myocardium, as defined above, was determined for each segment and then averaged for that patient.

Assessment of CMR image quality
Each myocardial segment of the left ventricle was rigorously assessed for image quality before inclusion in the final analyses.Only segments with minimal or no artifacts were included.Three controls were rejected due to artifacts.Four controls were excluded due to unavailability of map data.

Statistical analysis
Data were expressed as mean ± standard deviation or median (confidence interval).A paired Student's t test for continuous distribution with normal distribution was used to compare the walls.For categorical variables, the chi-square or Fisher's exact tests were used.The significance level was set at 5%.Analyses were carried out in the SPSS 21.0 (SPSS, Chicago, IL) and MedCalc 2020 software environments.Due to the unclear prevalence of myocarditis, sample size calculation was not performed.The initial sampling plan provided for 20 cases of myocarditis; 22 cases were found and ultimately included.This number is consistent with the existing literature. 8

Results
The most common symptom reported as the reason for suspicion of myocarditis and performance of CMR was chest pain (91%).The included patients had few comorbidities.Almost 70% of cases underwent CMR as hospital inpatients, while 95% of controls were scanned in an outpatient setting.Clinical and anatomic profile of myocarditis cases and controls are expressed in Table 1.

T2 ratio
For analysis of the T2 ratio, values ≥ 2:1 were considered abnormal.
The mean T2 ratio for LGE+ regions in cases were 2.75 ± 1, which is indicative of myocardial wall edema.
The mean T2 ratio for LGE− regions in patients with myocarditis was 1.50 ± 0.2, which represents a normal value for these walls.

T1 mapping
On analysis of T1 mapping, values ≥ 996 ms were considered abnormal.By this parameter, the number of affected segments was 111 of 132 (84%) in cases.The LGE+ segments of the cases (patients with myocarditis) showed a mean T1 value significantly different from the LGE− segments of the same patients.On between-group comparison, the mean T1 maps of LGE− ventricular walls in the patients with myocarditis were significantly different from the mean of the corresponding walls in controls.The mean T1 values in each group are given in Table 2.
Observing the most frequently abnormal region in our patient population, the inferolateral wall, the mean T1 value of the affected segment in cases was 1068 ± 47 ms, which is significantly different from all unaffected contralateral segments in these same cases.This change remained significant when we compared all unaffected segments of cases to those of controls.Figure 2 shows a representative image of the inferolateral wall.

T2 mapping
On analysis of T2 mapping, values ≥ 49 ms were considered abnormal.By this parameter, the number of affected segments was 69 of 130 (53%) in cases.The LGE+ segments of the cases showed a mean T2 value significantly different from the LGE− segments of the same patients.The mean T2 values of LGE− segments among cases were not significantly different from the mean T2 values of controls.The mean T2 values in each group are given in Table 3.
Regarding the inferolateral wall, T2 values again showed a significant difference between abnormal myocardial segments and unaffected walls in the same patients.Figure 3 shows a representative image of the inferolateral wall, and Figure 4 shows the relationship between different imaging methods.

Discussion
The present case-control study demonstrates that T1 mapping allows a more comprehensive, in-depth assessment of supposedly normal myocardium in patients with CMR-proven myocarditis.Comparison of LGE+ versus LGE− segments in cases and of cases versus controls revealed not only a regional inflammatory process, but also diffuse myocardial involvement.
Although T1 mapping has been progressively used as an adjunctive tool in the diagnosis of myocarditis, 9 the present study was designed to investigate this method as a means of detecting myocardial involvement in areas of the heart that are considered unaffected by myocarditis when evaluated by LGE alone.Comparatively, we found that even seemingly normal myocardial segments in patients with myocarditis could be compromised by inflammation.A mean T1 value of 1028 ± 48 ms in the unaffected walls of the patients with myocarditis proved to be statistically different from that of controls, which confirms the hypothesis that supposedly unaffected segments were in fact not normal.On the other hand, myocardial edema, as assessed by T2 mapping, showed no difference between the LGE− segments of cases and those of controls.The findings of this study are consistent with the existing literature. 10e proportion of affected segments in cases was 44% when analyzed by LGE alone and 84% when assessed by

Pereira et al. T1 and T2 Mapping in Myocarditis
T1 mapping.This result was interpreted as demonstrating a significant diffuse involvement of the myocardium, to the extent that almost the entire heart could be considered impaired in our patients with myocarditis.
The contribution of this finding to our knowledge of myocarditis is twofold: a) by enhancing the diagnostic performance of CMR in patients with myocarditis, particularly in borderline or difficult-to-diagnose cases; and b) by introducing a novel concept in the diagnosis of myocarditis which allows objective, numerical, and quantifiable assessment of myocardial involvement, unlike current LGE-based criteria, in which the diagnosis is subjective and operator-dependent.In addition, it should be noted that T1 mapping could obviate the use of gadolinium-based contrast agents, which eliminates the risk of allergic reactions, allows use in patients with renal failure, and reduces cost.
A multicenter observational study showed that T1 and ECV values were strong predictors of poor prognosis in non-ischemic dilated cardiomyopathy. 11Nevertheless, it remains unknown whether this altered myocardium is in itself a predictor of cardiovascular events in myocarditis and other cardiovascular diseases, due to a lack of studies with sufficient follow-up.
The diffuse T1 abnormalities in seemingly unaffected myocardial segments described in our study may have a major prognostic impact in the long term.Taylor et al note that diffuse fibrosis has been identified as an etiological factor in diastolic dysfunction, heart failure, and sudden death. 12garding T2 mapping, some studies have shown that this method might be able to locate areas involved in myocarditis with better sensitivity than conventional T2weighted images alone.In 1.5-Tesla CMR, a cutoff value of > 59 ms demonstrated 94% sensitivity and 97% specificity for identification of affected myocardium. 13Using a mean of > 49 ms, our study confirmed a statistically significant difference between cases with affected LGE+ walls and controls (Table 3).
The mean T2 ratio in LGE+ segments was 2.75 ± 1, which is an abnormal value, whereas, in segments without late gadolinium enhancement (LGE−), this ratio was normal.While T1 mapping was able to demonstrate  that, in addition to a regional inflammatory process, myocarditis is characterized by widespread, diffuse myocardial involvement, the T2 ratio was consistent with the presence or absence of LGE and did not reflect this diffuse inflammation.T2 ratio failed to detect the alterations suggestive of diffuse involvement detected by T1 mapping.
A systematic review assessed the prevalence of abnormal CMR findings in recovered COVID-19 patients.Almost 47% of recovered patients exhibited one or more abnormal CMR findings, which included elevated native T1 or T2 values. 14Another review that assessed data of 199 patients with the same profile showed that the most common imaging findings were abnormalities in T1 (73%) and T2 mapping (63%) and edema on T2/STIR sequences (51%).
LGE was observed in only 43% of cases.Similar to our previous pandemic data with non-COVID-19 patients, this study revealed that new quantitative mapping techniques are essential to detect diffuse myocardial inflammation also associated with COVID-19. 15erefore, T1 mapping was the only CMR technique capable of identifying diffuse changes in myocardial tissue, demonstrating abnormalities even in apparently normal ventricular walls.
Considering that the reference values of maps are determined by the value of the controls, and these are related to the characteristics of the patients at the study center and the magnetic resonance imaging device used, we believe that the work has internal validity.

Study limitations
Patients included in this study were selected at the time of CMR and not necessarily at the time of diagnosis with myocarditis.This may have slightly reduced the diagnostic accuracy of CMR, considering that the diffuse changes in the myocardial tissue could be in a healing curve after a few days, and T1 map values may be lower than in the acute phase.However, as our main objective was to compare myocardial segments in the same patients, this limitation may actually have enhanced, rather than jeopardized our analysis.
Although a T1 map has its own normality value according to the center in which it is evaluated, by way of comparison, our mean T1 map was slightly higher than reported in previous studies. 8,16This may have reduced the odds of finding significant differences between cases and controls.However, even considering this unexpected finding, we were able to detect a significant difference between the T1 values of controls and the LGE− segments of cases.We thus believe this was a conservative bias.This limitation may also have decreased the statistical significance of the analysis of T2 map values between groups.
Other limitations include the small sample size, the retrospective design, and the absence of endomyocardial biopsy to confirm the imaging findings.

Conclusion
This study suggests that, in patients with myocarditis, even ventricular wall segments with no LGE are abnormal on T1 mapping.The abnormal T1 map values found in LGE+ LGE-

Figure 1 -
Figure 1 -Image analysis among groups of walls of cases and controls.Designed in MS PowerPoint.LGE: late gadolinium enhancement.

Figure 2 -
Figure 2 -Inferolateral wall and the intermediate T1 map value of all unaffected walls of cases (blue bar: case LGE−).The T1 values of supposedly normal myocardium actually differed from that of the controls and from that of the affected walls of the same cases.* All p < 0.05.Values expressed as means.Designed in GraphPad Prism 9. LGE: late gadolinium enhancement.

Figure 3 -
Figure 3 -Analysis of the inferolateral wall by T2 mapping showed significant difference between LGE+ walls of cases (blue bar) and LGE− walls of the same cases (green bar).No significant difference was found compared to LGE− walls of controls (yellow bar).Values expressed as means.Designed in GraphPadPrism 9. LGE: late gadolinium enhancement.

Figure 4 -
Figure 4 -Relationship between different imaging methods: a) LGE+ in anterolateral wall and LGE− in septum; b) T2 ratio (3.38) shows edema in anterolateral wall and normal value (1.82) in septum; c) Affected T1 map in anterolateral wall (1121 ms) and in the septum (1041 ms); d) Affected T2 map in anterolateral wall (57 ms) and normal in the septum (42 ms).Designed in MS PowerPoint.LGE: late gadolinium enhancement.

Table 3 -Between-group comparison by T2 mapping in affected (LGE+) walls of cases versus all other unaffected (LGE−) walls in these cases and the respective unaffected (LGE−) walls of controls
a: analysis between T2 mapping in cases by affected wall and T2 mapping in all unaffected walls of the same patients; b: analysis between T2 mapping in all case's unaffected walls and T2 mapping of respective wall in controls; c: analysis between T2 mapping in cases by affected wall and T2 mapping of respective wall in controls.LGE+: presence of late gadolinium enhancement; LGE-: absence of late gadolinium enhancement; SD: standard deviation; n*: number of patients; n** number of walls assessed; †: one wall lost due to image artifact.