Ventricular Function and Tissue Characterization By Cardiac MRI in Children Following Hospitalization for Multisystem Inflammatory Syndrome in Children (MIS-C): A Prospective Study

Background Multisystem Inflammatory Syndrome in Children (MIS-C) is a severe life-threatening manifestation of SARS-CoV-2 infection. Acute cardiac dysfunction and resultant cardiogenic shock are common in children with MIS-C. While most children recover rapidly from acute illness, the long-term impact on the myocardium and cardiac function is unknown. Methods In this prospective study, cardiac MRI (CMR) was performed on patients <21 years of age with a history of MIS-C, 6–9 months following hospitalization. Per institutional protocol, patients with any history of LVEF<50%, persistent cardiorespiratory symptoms, or ECG abnormalities underwent clinical CMR. Research CMRs were offered to all others >10 years old. Native T1 and T2 mapping values were compared with 20 children with normal CMR examinations. Results We performed CMRs on 13 subjects at a median age of 13.6 years (interquartile range [IQR] 11.9–16.0) and a median time from hospitalization of 8.2 months (IQR 6.8–9.6). Twelve subjects displayed normal ventricular function with a median left ventricle ejection fraction (LVEF) of 57.2% (IQR 56.1–58.4) and median right ventricular (RV) EF of 53.1% (IQR 52.0–55.7). One subject had low normal EF (52%). There was normal T2 and native T1 as compared to normal controls. There was qualitatively no evidence of edema by T2 weighted imaging. One subject had late gadolinium enhancement (LGE) at the inferior insertion point and mid-ventricular inferolateral region, with normal EF, no evidence of edema or perfusion defects, and normal T1 and T2 times. When stratifying by a history of abnormal LVEF (LVEF <55%) on echocardiography, there was no difference in or parametric mapping values, though LVEF and LVEDV approached significance (p=0.06 and 0.05, respectively). Conclusions Although many children with MIS-C present acutely with cardiac dysfunction, myocardial recovery is overall excellent with minimal to no evidence of residual cardiac dysfunction or myocardial involvement. LVEF by CMR at 6–9 months among children with history of echocardiographic LV dysfunction is slightly lower, though does not meet statistical significance and is still within normal range. The long-term functional implications of this finding and the cardiac implications of MIS-C more broadly are unclear and warrant further study.


Abstract
Background Multisystem In ammatory Syndrome in Children (MIS-C) is a severe life-threatening manifestation of SARS-CoV-2 infection. Acute cardiac dysfunction and resultant cardiogenic shock are common in children with MIS-C. While most children recover rapidly from acute illness, the long-term impact on the myocardium and cardiac function is unknown.

Methods
In this prospective study, cardiac MRI (CMR) was performed on patients <21 years of age with a history of MIS-C, 6-9 months following hospitalization. Per institutional protocol, patients with any history of LVEF<50%, persistent cardiorespiratory symptoms, or ECG abnormalities underwent clinical CMR. Research CMRs were offered to all others >10 years old. Native T1 and T2 mapping values were compared with 20 children with normal CMR examinations.

Results
We performed CMRs on 13 subjects at a median age of 13.6 years (interquartile range [IQR] 11.9-16.0) and a median time from hospitalization of 8.2 months (IQR 6.8-9.6). Twelve subjects displayed normal ventricular function with a median left ventricle ejection fraction (LVEF) of 57.2% (IQR 56. 1-58.4) and median right ventricular (RV) EF of 53.1% (IQR 52.0-55.7). One subject had low normal EF (52%). There was normal T2 and native T1 as compared to normal controls.
There was qualitatively no evidence of edema by T2 weighted imaging. One subject had late gadolinium enhancement (LGE) at the inferior insertion point and mid-ventricular inferolateral region, with normal EF, no evidence of edema or perfusion defects, and normal T1 and T2 times. When stratifying by a history of abnormal LVEF (LVEF <55%) on echocardiography, there was no difference in or parametric mapping values, though LVEF and LVEDV approached signi cance (p=0.06 and 0.05, respectively).

Conclusions
Although many children with MIS-C present acutely with cardiac dysfunction, myocardial recovery is overall excellent with minimal to no evidence of residual cardiac dysfunction or myocardial involvement. LVEF by CMR at 6-9 months among children with history of echocardiographic LV dysfunction is slightly lower, though does not meet statistical signi cance and is still within normal range. The long-term functional implications of this nding and the cardiac implications of MIS-C more broadly are unclear and warrant further study.

Background
While children were relatively spared from severe acute disease secondary to the initial strains of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection, some developed a severe hyperin ammatory process associated with SARS-CoV2-2 infection, termed multisystem in ammatory syndrome in children (MIS-C).(1-4) This syndrome was characterized by high fevers, elevated in ammatory markers and in some, respiratory failure, signi cant cardiovascular compromise, including vasodilatory shock, severe valve regurgitation, ventricular dysfunction, and coronary artery dilatation.(3) Fortunately, most children recover rapidly from their acute presentation, with near-resolution of cardiac dysfunction and coronary dilatation by echocardiography by 4-9 months. (3) While mid-term clinical and echocardiographic reports to date have been encouraging, echocardiography cannot characterize the myocardium for the presence of brosis or edema that may be associated with long-term cardiac dysfunction or poor outcomes. Considering the similar degrees of systemic in ammation and myocardial disease noted in other viral-induced myocarditis during initial hospitalization, evaluating for the presence of signi cant long-term myocardial involvement is necessary. (5, 6) Despite signi cant improvement in symptomatology and echocardiographic markers in patients with MIS-C, understanding myocardial involvement and persistent myocardial in ammation may have implications towards exercise restrictions, returning to play, and long-term wellbeing. Current joint AHA/ACC recommendations for competitive athletics recommend restricting sports participation until in ammation has subsided or at least 3-6 months following the diagnosis of myocarditis (5), and the American Academy of Pediatrics and others have used this to design guidelines for patients following MIS-C diagnosis. (6) There remains limited data on persistence of cardiac in ammation after MIS-C, and expert consensus to date has relied heavily on extrapolation from related conditions. A more nuanced understanding of long-term cardiac in ammation post MIS-C hospital discharge may aid in counseling patients and families on return to sports and long-term implications.
Cardiac magnetic resonance imaging (CMR) is the gold standard non-invasive tool for tissue characterization, and is now mainstay for diagnosing the presence of edema and brosis in the diagnosis of viral induced myocarditis. (7,8) More recently, parametric mapping, including T1 and T2 mapping, has been employed as a method to identify and quantify diffuse brosis and edema.(9-11) T1 mapping has been shown to correlate with collagen volume fraction and extracellular matrix expansion as measured by histologic samples. (12)(13)(14) Clinically, this has been correlated with severity of diastolic dysfunction among adult patients with heart failure with preserved ejection fraction, as well as all-cause mortality, and parametric mapping has been incorporated into updated guidelines in the diagnosis of myocarditis. (15,16) To date, CMR publications in children recovering from MIS-C are predominantly within the initial 3 months following hospitalization, and larger studies on midterm outcomes have not incorporated quantitative markers of brosis or edema. (17)(18)(19)(20)(21)(22) We performed a prospective CMR study of pediatric patients hospitalized with MIS-C 6-9 months following hospitalization versus healthy controls to assess for residual myocardial disease.

Subject acquisition and consent
This study was approved by the Institutional Review Board at Columbia University Irving Medical Center. Children and adolescents <21 years of age with a history of hospitalization for MIS-C and no known prior cardiac disease were recruited for participation in this study. All children with suspected MIS-C are evaluated by the Columbia University Interdisciplinary MIS-C Team and offered longitudinal follow-up by the Columbia University Interdisciplinary MIS-C Follow-up Program. Diagnosis and management of MIS-C at our institution follows our institutional protocol and has been previously described.(1, 3) All suspected cases of MIS-C were evaluated and adjudicated to con rm accurate diagnosis.
By institutional protocol, all children with history of ventricular dysfunction (LVEF <50%), residual cardiac symptoms, including chest pain, palpitations, or exercise intolerance, persistently elevated in ammatory markers, or ECG changes/persistent dysrhythmia were referred for CMR 6-9 months following hospitalization for MIS-C. Families of all other children 10 years of older in the Interdisciplinary MIS-C Follow-up Clinic were approached for research CMR.
This age cutoff was chosen to maximize the likelihood of a successful completion of the CMR exam without the need for sedation. Informed consent and assent were obtained for all subjects.

CMR image acquisition
All CMR studies were performed on a 1.5T GE Explorer CMR scanner (GE healthcare systems, Chicago, IL, USA). The CMR protocol included conventional balanced steady state free procession (bSSFP) cine imaging in long and short axis planes using end-expiratory breath holding or apnea based on anesthesia requirement. Cine imaging slice thickness was modi ed to obtain 10-12 images per short axis stack. Pre-contrast T2 mapping was performed in basal, mid, and apical short axis slices using a breath-held or apnea motion corrected turbo spin echo sequence with a T2 preparatory pulse and bSSFP readout with the following typical parameters: echo times: 11 ms, 41 ms, 71 ms, 100 ms; slice thickness, 8mm; bandwidth 780Hz/pixel; pixel size, 2.2 x 2.2 mm; FOV (mm) 350; phase FOV (%), 70%; temporal resolution, 127 ms; phase resolution, 160 x 160; acceleration factor of 2. and T2-weighted turbo spin echo (TSE) in basal, mid, and apical short axis slices was acquired. Native T1 mapping was obtained using electrocardiogram triggered modi ed Look-Locker inversion recovery (MOLLI) sequences in basal, mid, and apical short axis slices during diastole using a 5(3)3 sequence. Typical sequence parameters were: slice thickness, 8mm; ip angle, 35°; echo time, 1.6 ms, repetition time, 3.8ms, bandwidth 1250Hz/pixel; pixel size, 2.2 x 2.2 mm; FOV (mm) 350; phase FOV (%), 70%; temporal resolution, 261 ms; phase resolution, 160 x 160; acceleration factor of 2. First pass perfusion imaging was performed immediately following the administration of 0.2mmol/kg of Gadobutrol. Phase sensitive inversion recovery imaging was performed in long and short axis planes to assess for late gadolinium enhancement (LGE). All static sequences except for rst pass perfusion, were performed using breath holding or apnea (based on anesthesia needs) at end-diastole.
CMR imaging post-processing Image post-processing was performed using Circle CVI42 (Circle Cardiovascular Imaging Inc, Alberta, CA). Left and right ventricular (LV and RV, respectively) end-diastole and end-systole endocardial and epicardial borders were traced to determine end-diastolic and end-systolic volumes (EDV and ESV, respectively and myocardial mass. Ejections fraction (EF) were calculated by the equation ((EDV-ESV)/EDV)*100 and reported as percentages. Indexed EDV, ESV, and masses were calculated by dividing by body surface area (BSA). Volume and LV mass Z-scores were generated using published data. Z-scores between 2.0 and 3.5 were considered mildly enlarged, between 3.5 and 5.0 were considered moderately enlarged, and >5.0 were considered severely enlarged. (23) T2 and T1 mapping were analyzed in the respective CVI42 modules. Endocardial and epicardial borders were traced with 10% offset to ensure that only myocardium was included in tracings ( Figure 1). T2 and native T1 values (msec) are reported as the mean value for the given basal, mid-ventricular, or apical slice.
T2 TSE, rst pass perfusion, and inversion recovery imaging were reviewed to determine the presence of edema, perfusion defects, and LGE, respectively. All imaging analyses were performed by a single pediatric cardiologist with extensive training and experience in pediatric CMR (MPD).

Controls
Controls for native T1 and T2 mapping were retrospectively identi ed by review of the CMR database for patients with structurally normal hearts undergoing CMR to evaluate coronary anatomy with normal coronary origins, ECG abnormalities, or isolated PVCs/PACs to rule out arrhythmogenic RV dysplasia with normal biventricular size and systolic function and no evidence of LGE. CMR imaging performed solely to assess normal scanner T2 and native T1 times per Society of Cardiovascular Magnetic Resonance (SCMR) recommendation with qualitatively normal biventricular size and function were also included as controls.

Statistical analysis
Descriptive statistics were presented as frequency counts and percentages for categorical variables and mean ± standard deviation (SD) or median (interquartile range; IQR) for continuous variables, as appropriate. Based on the distribution of the data, Student's t-test or Wilcoxon rank sum was used to compare T1 and T2 values between subjects and controls. Spearman rho or Pearson correlation coe cient was used to assess associations between T1, T2, and LVEF as compared to lowest LVEF by echocardiography during hospitalization. The two-sided statistical signi cance level was set at 0.05. All data analyses were performed using STATA 14.1 (StataCorp, College Station, TX).

Results
Thirty-two were followed for at least six months. Eight of these had cardiac dysfunction during initial hospitalization and, therefore, met clinical criteria for CMR. Of these, seven underwent CMR at our institution. Of these seven, one family declined to participate in this research. Two additional patients, hospitalized at an a liated institution, were referred to our center for CMR. Of the eight children 10 years or older with normal cardiac function on admission followed for more than six months at our clinic, ve agreed to participate (Figure 2).
In total, 13 subjects were included in this analysis (eight clinical and ve research CMR) and underwent CMR. Two subjects (15%) required anesthesia with endotracheal intubation for CMR. Of the subjects who underwent CMR, 9 (69%) were male. The median age at hospitalization was 13.1 years (IQR 11.3-15.0), with a median length of hospital stay of 5 days (IQR 4-6). Nine (69%) subjects required ICU admission. Five (41%) subjects required inotropic support and 5 (41%) required respiratory support, including intubation in one subject. All subjects received IVIG, aspirin, and pulse steroids during the acute presentation with a steroid taper following discharge. LV dysfunction and coronary artery dilation or prominence was present in the majority (69% and 77%, respectively). Two subjects had residual ventricular dysfunction on discharge echocardiogram, which resolved by 3 months following discharge. Demographics and hospitalization data are described in Table 1.  .7)). Subjects overall had normal biventricular size and LV mass based on indexed ventricular size and mass and Z-score (Table 2). One subject had low normal LVEF (52%) with diminished apical wall motion, but otherwise normal ventricular size and mass, and no evidence of perfusion defects, LGE, or edema. Two subjects had mild LV dilation (Z-score +2.2 and +2.8) with overall normal LV and RVEF and no evidence of edema or brosis. T1 and T2 times in these subjects were normal. There was no evidence of edema based on T2 TSE imaging and no perfusion defects noted on rst pass perfusion. One subject-different from the aforementioned-had LGE at the inferior insertion point and mid-ventricular inferolateral region, with normal ventricular function, no evidence of edema or perfusion defects, and normal T1 and T2 times. This subject had a prior history of moderate LV dysfunction by echocardiography, with an LVEF of 44%.
Characteristics of the two subjects with diminished function and LGE are presented in Table 3 (PLACE Table 3 Table 2. There was no difference in T2 or native T1 times as compared to controls (Table 2, Figure 3 and 4).
When stratifying subjects by ventricular function during initial admission, LVEF and LVEDV medians on CMR were marginally lower but not statistically different among those with and without a history of ventricular dysfunction (56% vs. 59%, p=0.06; 129mL vs. 199mL, p=0.05, respectively). There was no difference in indexed ventricular volumes, RVEF, or T1 and T2 mapping values (Table 4). Subjects with a history of low LVEF on admission echocardiogram tended to be younger (p=0.03). Among the 11 subjects with available LVEF on echocardiography during admission, there was no association between the lowest available EF and LVEF by CMR (Spearman rho 0.42, p=0.2).

Discussion
The emergence of SARS-CoV-2 has resulted in signi cant morbidity and mortality across the globe. We are only just starting to ascertain the long-term impact that this new virus will have on human health. Acute cardiac dysfunction is a prominent and critical manifestation for children who develop MIS-C following infection with SARS-CoV-2. Whether acute cardiac dysfunction during MIS-C ultimately results in permanent cardiac dysfunction or persistent in ammation as seen in other viral induced myocarditis syndromes remains unde ned. Determining the presence of cardiac in ammation and myocardial damage has signi cant implications towards counseling, as the presence of residual in ammation may necessitate restrictions from sports participation until such in ammation has resolved.Improving our understanding of the potential lifelong impact that MIS-C will have on children's cardiac health is therefore essential to our ability to guide families and determine treatment plans.
We report on the mid-term CMR ndings in children 6-9 months following hospitalization for MIS-C. Subjects overall have normal ventricular size and systolic function, with LV dysfunction present in only one subject. There is no evidence of edema and parametric mapping values were normal as compared to controls.
LGE was present in only one subject with no evidence of perfusion defects and overall normal T1 times in this subject. Importantly, we show nearcomplete recovery in ejection fraction among subjects with a history of ventricular dysfunction.
Our group previously reported on six-month echocardiographic ndings in this population and demonstrated excellent recovery in almost all patients.(3) While echocardiography is a valuable tool to assess global ventricular function, CMR remains the gold standard measure to assess myocardial tissue characterization.(9) Furthermore, LVEF by echocardiography requires numerous assumptions on LV geometry, and assessment of RV function is predominantly qualitative; CMR allows for a more granular calculation of EF.
Beyond presenting prospective data further from hospitalization, our study is novel in that we incorporated the use of a parametric mapping for myocardial tissue characterization in a cohort exclusively comprised of a MIS-C population. Parametric mapping is quickly becoming a standard of care method in the evaluation of myocardial edema and diffuse brosis in patients, and has recently been incorporated as part of revised Lake Louise criteria in the evaluation of myocarditis. (7) Webster et al. evaluated native T1 and T2 values in pediatric patients with symptomatic COVID; however, they did not administer contrast in their cohort, and therefore were unable to assess for LGE. While Barris et al report on mid-term CMR ndings in patients hospitalized with MIS-C, they did not perform parametric mapping; evaluation of edema and brosis was therefore limited to a qualitative analysis. Furthermore, while LGE can identify focal or replacement brosis, it is unable to identify diffuse brosis. . (11,18,21) There are several limitations to this study. Our sample size is limited, due to an a priori decision to only obtain clinical CMR to patients who had evidence of cardiac dysfunction during initial presentation and few subjects willing to consent to a research CMR with contrast administration. Therefore, we may have been underpowered to detect differences in functional and myocardial tissue characterization parameters. Furthermore, younger subjects were excluded from research CMR; as a result, we may not be able to extrapolate our ndings to those hospitalized at a younger age. That said, our sample is biased to those with initial cardiac dysfunction, and therefore would be biased to those most at risk. Finally, control data were available for native T1 and T2 values as part of recommendations by SCMR to obtain local references; (11) as control data were based on subjects referred for clinical CMR, it is possible that myocardial tissue abnormalities were present. However, we ensured that all subjects used as controls had normal biventricular size and systolic function with no evidence of identi able congenital heart disease or systemic diseases. Therefore, the likelihood of myocardial abnormalities is minimal. As availability of normal control myocardial T1 data were limited, we could not age match in this study, and therefore it is possible that lack of differences between controls and subjects was due to age dependent phenomenon. Furthermore, normal controls were not available for ECV due to the need for contrast administration. While comparisons to published reference values have been deemed acceptable, the availability of published data in children are limited to small single center cohorts. (11)

Conclusions
Our study demonstrates overall excellent midterm cardiac outcomes among this cohort of children with MIS-C. Recovery of ventricular function was nearly universal with minimal evidence of residual myocardial disease in our cohort. Long-term, larger-scale studies are needed to ensure complete recovery in those with residual disease and to identify if any children remain at risk for long-term myocardial disease and dysfunction after MIS-C.