JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY
20, 2019
Listen to this manuscript’s
audio summary by
Editor-in-Chief
Dr. Valentin Fuster on
JACC.org.
VOL. 74, NO.
ª 2019 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION
PUBLISHED BY ELSEVIER
ORIGINAL INVESTIGATIONS
Prognostic Value of Repeating
Cardiac Magnetic Resonance in
Patients With Acute Myocarditis
Giovanni Donato Aquaro, MD,a Yacob Ghebru Habtemicael, MD,a Giovanni Camastra, MD,b Lorenzo Monti, MD,c
Santo Dellegrottaglie, MD,d,e Claudio Moro,f Chiara Lanzillo, MD,g Alessandra Scatteia, MD,h Mauro Di Roma, MD,i
Gianluca Pontone, MD,j Martina Perazzolo Marra,k Andrea Barison, MD,a Gianluca Di Bella,l on behalf of the
“Cardiac Magnetic Resonance” Working Group of the Italian Society of Cardiology
ABSTRACT
BACKGROUND Cardiac magnetic resonance (CMR) is widely used to confirm the diagnosis of acute myocarditis (AM) in
the acute setting. CMR is often repeated after 6 months to assess the evolution of myocardial involvement. However, the
clinical and prognostic role of 6-month CMR is unknown.
OBJECTIVES This multicenter study aimed to evaluate the clinical and prognostic role of 6-month repetition of CMR in
patients with AM.
METHODS In a subgroup of 187 patients from the ITAMY (ITAlian study in MYocarditis) registry, CMR was performed
within the first week after symptom onset (CMR-I) and repeated after 6 months (CMR-II).
RESULTS Myocardial edema was detected in all the patients at CMR-I and persisted in 31 (16%) at CMR-II. LGE was
detected in 182 (96%) patients at CMR-I and in 164 (86%) at CMR-II. At CMR-II, 20 (11%) patients presented a complete
recovery from edema and LGE, 30 (16%) patients had edema with LGE, and 137 (73%) presented LGE without edema.
LGE disappeared completely in 18 (10%) patients, the number of LGE segments decreased in 87 (46%), unchanged in 58
(31%), and increased in 26 (14%). During a median clinical follow-up of 7 years (25th to 75th percentile: 6 to 8 years)
cardiac events occurred in 22 patients. At Kaplan-Meier curves, patients with LGE and without edema had worse prog-
nosis than others (p < 0.0001). Patients with increased extent of LGE (p ¼ 0.02) had a worse prognosis than those with
decreased/unchanged LGE. At multivariate Cox regression analysis, the midwall septal pattern of LGE and the presence of
LGE without edema at CMR-II were independent predictors of a cardiac event.
CONCLUSIONS In the acute setting, LGE does not mean definite fibrosis, and it may disappear at 6 months. The presence
of LGE without edema at 6-month CMR is associated with worse prognosis, particularly when distributed with a midwall
septal pattern. LGE without edema could represent definite fibrosis whereas the presence of edema suggests a residual
chance of recovery. (J Am Coll Cardiol 2019;74:2439–48) © 2019 by the American College of Cardiology Foundation.
From the aFondazione Toscana G.Monasterio, Pisa, Italy; bCardiac Department, Vannini Hospital Rome, Roma, Italy; cRadiology
Department, Humanitas Research Hospital, I.R.C.C.S., Rozzano, Milan, Italy; dDivision of Cardiology, Villa dei Fiori, Acerra,
Napoli, Italy; eMount Sinai School of Medicine, New York, New York; fU.O. Cardiologia e UTIC, ASST Monza, P.O. Desio, Desio,
Italy; gCardiology Department, Policlinico Casilino, Rome, Italy; hDepartment of Advanced Biomedical Sciences, Federico II
University, Naples, Italy; iRadiological Department, European Hospital, Roma, Italy; jCardiac Department, Centro Cardiologico
Monzino, Milano, Italy; kDivision of Cardiology, Department of Cardiac, Thoracic, and Vascular Sciences, University of Padua,
Padua, Italy; and the lClinical and Experimental Department of Medicine, University of Messina, Messina, Italy. The authors have
reported that they have no relationships relevant to the contents of this paper to disclose. Dudley Pennell, MD, served as Guest
Associate Editor for this paper.
Manuscript received July 9, 2019; revised manuscript received August 29, 2019, accepted August 30, 2019.
ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2019.08.1061
2440 Aquaro et al.
Prognostic Role of 6-Month Follow-Up CMR in Myocarditis
C
ABBREVIATIONS ardiac magnetic resonance (CMR) is
AND ACRONYMS the best noninvasive imaging mo-
dality for the diagnosis of acute
AM = acute myocarditis
myocarditis (AM) in hemodynamically stable
CMR = cardiac magnetic
patients with preserved left ventricular (LV)
resonance
ejection fraction (EF) (1,2).
EDV = end-diastolic volume
In the clinical setting, CMR is performed
EF = ejection fraction
within the first week following symptom
ICD = implantable
onset to confirm the diagnosis of AM while
cardioverter-defibrillator
evaluating myocardial edema; late gadolin-
LGE = late gadolinium
enhancement ium enhancement (LGE); and novel T1, T 2,
LV = left ventricular
and ECV mapping techniques (3,4). In the
clinical setting, CMR is often
SSFP = steady-state free
precession 6 months after symptom onset to assess the
evolution of myocardial damage and to
perform a functional evaluation. However, few data
are available regarding the clinical role and meaning
of CMR repeated 6 months after symptom onset.
Anzini et al. (5) showed that left ventricular (LV)
dysfunction at admission and at 6 months is a strong
predictor of poor prognosis in patients with AM.
Recently, the presence of LGE in the midwall layer of
the interventricular septum, in an acute setting, was
associated with a worse long-term prognosis in a large
series of patients with AM and preserved LVEF (6).
LGE is generally considered a marker of irreversible
myocardial damage in myocardial infarction and
cardiomyopathy, as it was in the original Lake Louise
criteria for CMR diagnosis of AM (3). However, in a
population of patients with AM, Mahrholdt et al. (7)
repeated CMR 6 months after the first diagnosis and
found a complete disappearance of LGE in 19 of 71
patients. This may suggest that in AM patients, LGE is
not necessarily a marker of myocardial fibrosis but
may be associated with acute inflammation, and dis-
appears after a few months in a significant number of
patients. Thus, in the evolution of CMR signs of
myocardial damage, its clinical and
meaning may be unpredictable in the acute setting.
SEE PAGE 2449
The aim of this study was to: 1) evaluate the evo-
lution of myocardial tissue abnormalities in patients
with AM and preserved LVEF by performing CMR
within the first week of symptoms (CMR-I) and after
6 months (CMR-II); and 2) assess the clinical and
prognostic implications of 6-month CMR results for
AM.
METHODS
This study is a substudy of the ITAMY (ITAlian study
in MYocarditis) multicenter investigation on prog-
nostic value of CMR in AM (6). A total of 200
JACC VOL. 74, NO. 20, 2019
NOVEMBER 19, 2019:2439–48
consecutive hemodynamically stable patients with
clinically suspected AM and preserved LVEF were
enrolled in the study from January 1, 2008, to
December 31, 2014. We excluded patients with heart
failure and arrhythmic presentation because of the
demonstrated low sensitivity of CMR for detecting
acute AM in these kinds of patients. Based on the
recent European Society of Cardiology position paper
(8), clinically suspected AM was defined when
symptomatic patients with chest pain (pericarditis or
pseudoischemic pain) fulfilled 1 or more diagnostic
criteria (new electrocardiogram modification,
repeated elevated troponin, wall motion abnormalities with
preserved LVEF on echocardiography) or in asymp-
tomatic patients with 2 or more diagnostic criteria. A
definite diagnosis of AM was then made in the pres-
ence of 2 or more original CMR Lake Louise criteria
(myocardial edema, hyperemia, and LGE) (3). Endo-
myocardial biopsy was performed when CMR was
inconclusive (#1 CMR criterion). To exclude obstruc-
tive coronary artery disease, coronary artery angiog-
raphy was performed on all patients, with the
exception of those younger than age 30 years with
low risk of coronary artery disease. At hospital
admission, all patients underwent clinical evaluation,
electrocardiogram, echocardiography, and laboratory
testing (including leucocytes, C-reactive protein,
erythrocyte sedimentation rate, and troponin T).
Informed consent was obtained from all patients at
the time of first CMR examination (CMR-I).
A follow-up CMR examination (CMR-II) was per-
formed 6 months (median 177 days) after the first
CMR. From the initial enrolled population, 13 patients
were excluded because of suboptimal images at
CMR-I and/or CMR-II. The final population included
187 patients (82 men, mean age 33  13 years).
CMR ACQUISITION PROTOCOL. CMR was performed
prognostic
with 1.5-T systems (CVi, HD release, GE Healthcare,
Milwaukee, Wisconsin; Magnetom Avanto, Siemens
Medical Systems, Erlangen, Germany; Gyroscan NT,
Philips Healthcare, Best, the Netherlands) using
dedicated cardiac software, a phased-array surface
receiver coil, and vectocardiogram triggering. Ac-
cording to the protocols recommended by the Society
for Cardiovascular Magnetic Resonance, we acquired
cine steady-state free precession (cine-SSFP) images,
T 2 -weighted imaging, and LGE at 10 min after gado-
linium injection in the short-axis (9 to 13 images
covering the entire LV), 2-chamber, and 4-chamber
planes. Short-axis cine-SSFP images were acquired
immediately after gadolinium injection for hyperemia
assessment. The same protocol was repeated at
CMR-II.
JACC VOL. 74, NO. 20, 2019 Aquaro et al. 2441
NOVEMBER 19, 2019:2439–48 Prognostic Role of 6-Month Follow-Up CMR in Myocarditis

myocardium in each short-axis image. A region of

T A B L E 1 Patient Characteristics (N ¼ 190)
interest was placed in a region of myocardium
Men 155 (82)
without LGE. The mean SI and SD were measured in
Age, yrs 33  13
this region of interest. LGE was defined as myocar-
Weight, kg 78  14
Height, cm 174  8 dium with an SI higher than the average SI of the
Risk factors region of interest plus $6 SD. LGE extent was
Hypertension 8 (5) expressed as % of the LV mass. The increase of extent
Hypercholesterolemia 5 (3) in LGE was defined as an augmentation of $1% of the
Diabetes mellitus 1 (0.5) LV mass. Myocardial hyperemia was evaluated as
Smoking 13 (7)
previously reported (10) using post-contrast cine-
Family history of CAD 10 (5)
SSFP images. The occurrence of edema, hyperemia, or
Obesity 10 (5)
Clinical presentation
LGE was evaluated in each of the 17 LV segments
Chest pain 190 (100) (10,11).
Fever (2 weeks from symptom onset) 81 (43) CLINICAL FOLLOW-UP. After the CMR examination,
Pericardial effusion 45 (24)
a follow-up was performed for all patients by the in-
ECG ST-T abnormalities 144 (76)
vestigators at the 10 Italian hospitals involved in the
ECG negative T 23 (12)
patient registry. A clinical questionnaire was
ECG ST elevation 121 (64)
Coronary artery angiography 170 (89) completed for each patient by a clinical physician
Endomyocardial biopsy 18 (9) during a periodic ambulatory visit at each hospital, by
Laboratory test contacting patients’ relatives by telephone, by a
Leukocytes, 103 cells/m2 9.7  4 general practitioner, or by consulting the office of
ESR, mm/h 12 (1–28) vital statistics at the patient’s place of residence. The
CRP, ng/ml 6 (2–24)
clinical questionnaire evaluated the occurrence of the
Elevated troponin 190 (100)
following major events: cardiac death, resuscitated
Values are median (25th to 75th percentile), mean  SD, or n (%). The data refer cardiac arrest, ventricular assist device, trans-
to the entire population of 190 patients.
plantation, and appropriate implantable cardioverter-
CAD ¼ coronary artery disease; CRP ¼ C-reactive protein;
ECG ¼ electrocardiographic; ESR ¼ erythrocyte sedimentation rate. defibrillator (ICD) shock. It also included minor
events (heart failure hospitalization). A complete
analysis of the ICD was performed by the referring
CMR ANALYSIS. All CMR studies were analyzed physician to confirm the appropriateness of
offline using a workstation with dedicated cardiac the shock.
software, and consensus was obtained among STATISTICAL ANALYSIS. Values are presented as the
three experienced observers who were blinded to mean  SD or as the median (25th to 75th percentile) for
the clinical presentation results. To evaluate the variables with normal and non-normal distributions,
LV global and regional function and calculate the respectively. Values with non-normal distribution
LV mass, the endocardial and epicardial borders according to the Kolmogorov-Smirnov test were
were manually drawn in the end-diastolic and logarithmically transformed for parametric analysis.
-systolic short-axis cine-SSFP images. Papillary Qualitative data are expressed as percentages.
muscles and trabeculations were not included in Categorical variables were compared using the chi-
the myocardium. LV end-diastolic volume (EDV), square test or Fisher exact test when appropriate.
LV end-systolic volume, EF, and LV mass Continuous variables were compared by the Stu-
were determined. dent’s independent t-test and analysis of variance or
On T 2-weighted images, edema was considered by the Wilcoxon non-parametric test when appro-
present when the ratio of signal intensity (SI) be- priate. The Kaplan-Meier time-to-event method was
tween the myocardium, and the mean SI of the skel- used to calculate and compare longitudinal curves
etal muscle was $2 (3,9). LGE was qualitatively between groups. Logistic regression analysis was
evaluated and presented as a nonischemic pattern of used to explore the effect of each significant variable
distribution (i.e., subepicardial or midventricular in univariate analysis to predict the occurrence of
enhancement), and the number of myocardial seg- cardiac events evaluated as a combined endpoint
ments with LGE was counted (10,11). The extent of (cardiac death, appropriate ICD-firing, resuscitated
LGE was measured using a previously validated cardiac arrest, and hospitalization for heart failure).
method. Briefly, the endocardial and epicardial A p value <0.05 was considered statistically
contours were manually traced to identify LV significant.
2442 Aquaro et al. JACC VOL. 74, NO. 20, 2019

Prognostic Role of 6-Month Follow-Up CMR in Myocarditis NOVEMBER 19, 2019:2439–48

patients, remained unchanged in 58 (30%), and

T A B L E 2 Cardiac Magnetic Resonance Data
increased in 26 (14%). An example of a patient with
CMR-I CMR-II p Value complete disappearance of LGE at 6 months is shown
LVEDVi, ml/m2 82 (72–93) 80 (73–90) 0.56 in Figure 1. The distribution of LGE did not change
LVESVi, ml/m2 31 (24–36) 30 (25–34) 0.99
from CMR-I to CMR-II. It was subepicardial in the
LVEF, % 61  7 62  8 0.28
2
inferior and/or lateral segments of 139 patients (76%),
LVMi, g/m 75  11 69  13 <0.0001
LV dysfunction 0 5 (3%) 0.06
and in the midwall involving the interventricular
RVEDVi, ml/m2 82 (71–93) 84 (73–94) 0.58 septum in 43 (24%) patients. All 18 patients with
RVESVi, ml/m2 32 (25–38) 33 (27–40) 0.55 complete disappearance of LGE had a subepicardial
RVEF, % 61  10 59  10 0.28 inferior and/or lateral pattern at CMR-I, whereas LGE
Edema T2w 190 (100) 30 (16) <0.0001 did not disappear in patients with a mid-wall septal
Hyperemia 42 (22) 12 (6) <0.0001
pattern (p ¼ 0.007). Patients with LGE disappear had
LGE 182 (96) 164 (86) 0.03
greater extent of LGE at CMR-I than others: 10.2%
Absence of edema and LGE 0 20 (11) <0.0001
(25th to 75th percentile: 8.3% to 14.5%) versus 4.2%
LGE without edema 0 137 (73) <0.0001
Edema without LGE 8 (4%) 0 0.007
(25th to 75th percentile: 2.1% to 8.3%) of LV mass,
Segments with edema, n 3 (2–5) 2 (2–4) 0.02 respectively (p < 0.01). There was no significant
Segments with hyperemia, n 2 (0–3) 0 (0–1) 0.02 relationship between the extent of LGE at CMR-I and
Segments with LGE, n 3 (2–5) 3 (1–4) 0.0003 the number of days between symptom onset and
Extent of LGE (% of LV mass) 6.2 (2–10) 4.1 (2–8) <0.0001 CMR-I (r ¼ 0.02; p ¼ 0.78). The average number of
Pericardial effusion 45 (24) 15 (8) <0.0001
days between symptoms and CMR-I was not different
Pericardial enhancement 12 (6%) 7 (4%) 0.34
between patients with an increased amount of LGE at
RV involvement (LGE and/or edema) 6 (3%) 2 (1%) 0.28
CMR-II and those with decreased/unchanged extent
Values are median (25th to 75th percentile), mean  SD, or n (%). The data refer to the entire population of 190 of LGE (5.6  2 days vs. 5.4  1 day; p ¼ 0.68).
patients.
The LV mass index decreased at CMR-II compared
CMR ¼ cardiac magnetic resonance; LGE ¼ late gadolinium enhancement; LV ¼ left ventricular; LVEDVi ¼ left
ventricular end-diastolic volume index; LVEF ¼ left ventricular ejection fraction; LVESVi ¼ left ventricular end- with that at the basal examination (average LV mass
systolic index; LVMi ¼ left ventricular mass index; RVEDVi ¼ right ventricular end-diastolic volume index;
RVEF ¼ right ventricular ejection fraction; RVESVi ¼ right ventricular end-systolic index.
index 75 g/m 2 vs. 69 g/m 2; p < 0.0001). The LV mass
index decrease was greater in patients with complete
healing (no LGE and edema) than in the other patients
RESULTS (median decrease was 10 g/m 2 [25th to 75th percen-
tile: 7 to 16 g/m 2] vs. 8 g/m 2 [25th to 75th percentile: 4
The final population included 187 patients (82 men, to 12 g/m 2]; p ¼ 0.009). In contrast, no significant
mean age 33  13 years) who completed CMR-I and changes in LVEF, LVEDV index, right ventricular EF,
-II. The patient characteristics are reported in and right ventricular EDV index were detected be-
Table 1. tween the examinations. Pericardial effusion was
CMR-I was performed a median of 4 days (2 to present in 45 (24%) patients at CMR-I and in 15 (8%)
7 days) after the onset of AM symptoms and repeated patients at CMR-II.
(CMR-II) after a median of 177 days (100 to 233 days). In patients with persistent edema at CMR-II, the
A comparison of the findings between CMR-I and -II is examination was repeated after 6 more months:
reported in Table 2. Myocardial edema was present in among the 30 patients with edema at CMR-II, only 3
all patients at CMR-I and in only 31 (16%) patients at (10%) patients had persistent edema at 12 months and
CMR-II. LGE was detected in 182 (96%) patients at 1 (3.3%) at 18 months.
CMR-I and in 164 (86%) patients at follow-up. At CLINICAL FOLLOW-UP. During the median follow-up
CMR-II, 20 (11%) patients presented with a complete of 7 years (25th to 75th percentile: 6 to 8 years),
recovery from edema and LGE, 30 (16%) patients endpoint cardiac events were found in the following
presented with edema and LGE, and 137 (73%) pre- 22 patients: 8 major cardiac events (3 sudden cardiac
sented with LGE without edema. All 8 patients with deaths, 2 resuscitated cardiac arrests, 3 appropriate
edema without LGE at CMR-I presented with a com- ICD therapies) and 14 hospitalizations for heart fail-
plete recovery at follow-up. A total of 43 patients ure. The characteristics of the patients with and
(23%) had a midwall septal pattern of LGE. without cardiac events are reported in Table 3. Pa-
As evident in Table 2, a significant decrease of tients with cardiac events had more frequent in-
myocardial segments with edema (p ¼ 0.02) and LGE creases of LGE extent segments than other
(p ¼ 0.0003) was found at CMR-II. Overall, at CMR-II, patients (p < 0.0001).
LGE disappeared completely in 18 (10%) patients; the As evidenced by the Kaplan-Meier survival curves
extent of LGE segments decreased in 87 (46%) on the Central Illustration, patients presenting with
JACC VOL. 74, NO. 20, 2019 Aquaro et al. 2443
NOVEMBER 19, 2019:2439–48 Prognostic Role of 6-Month Follow-Up CMR in Myocarditis

F I G U R E 1 An Example of LGE Disappearing 6-Months After Symptom Onset Using CMR

In this case of a male patient with acute myocarditis, CMR (cardiac magnetic resonance) performed in the first week following symptom onset
(CMR-I) showed myocardial edema (arrows in T2-weighted [T2w] images of top row) and late gadolinium enhancement (LGE) (middle row).
After 6 months (CMR-II) (bottom row) LGE was negative.

LGE without edema at CMR-II had a worse prognosis DISCUSSION

than those with LGE and persistence of edema
(p < 0.0001). No cardiac events occurred in the pa- The main results of the present study may be sum-
tients with absence of both LGE and edema (complete marized as follows:
recovery) at CMR-II.
1. At 6 months, signs of myocardial edema dis-
In Figure 2, Kaplan-Meier survival curves demon-
appeared in the majority of patients (84%), and a
strated that patients with an increased LGE extent
complete healing (absence of edema and LGE) was
had worse prognosis than those with decreased/ab-
found in 11% of patients.
sent LGE at CMR-II (p ¼ 0.02).
2. LGE disappeared in 10% of the patients, and its
Univariate and multivariate Cox hazards-
extent, measured as number of myocardial seg-
proportional regression analysis are reported in
ments involved, decreased in 46% of patients,
Table 4. In univariate analysis, the LVEF at CMR-I, the
remained unchanged in 30%, and increased in 14%.
extent of LGE at CMR-I and CMR-II, the presence of
3. Patients with LGE without edema had a worse
midwall septal pattern of LGE, the persistence of LGE
prognosis than those with persistent edema.
without edema at CMR-II, and the increase of LGE
4. Patients with increased extent of LGE at CMR-II
extent at CMR-II were associated with cardiac events.
had a worse prognosis than those with un-
In the multivariate Cox hazards-proportional
changed/decreased LGE.
regression analysis, only the midwall septal pattern
5. The midwall septal pattern of LGE and the pres-
of LGE (2.8; 95% confidence interval: 1.1 to 7.2;
ence of LGE without edema were independent
p ¼ 0.028) and the persistence of LGE without edema
predictors of cardiac events.
at CMR-II (4.5; 95% confidence interval: 1.3 to 14.5;
p ¼ 0.008) were independent predictors of cardiac CMR is widely accepted as the preferred imaging
events. modality in the acute phase of clinically suspected
2444 Aquaro et al. JACC VOL. 74, NO. 20, 2019

Prognostic Role of 6-Month Follow-Up CMR in Myocarditis NOVEMBER 19, 2019:2439–48

phase of myocarditis is not necessarily synonymous

T A B L E 3 Characteristics of Patients With and Without Cardiac Events
with irreversible damage, because in 11% of our pa-
Cardiac Event tients, LGE completely disappeared at follow-up.
Yes (n ¼ 22) No (n ¼ 168) p Value Actually, this finding is not novel. In 2006, Mahr-
Men 16 (73) 139 (82) 0.28 holdt et al. (7) repeated CMR 6 months after symptom
Age, yrs 35  14 32  13 0.33
onset in 71 patients with acute myocarditis and
Hypertension 2 (10) 6 (4) 0.23
presence of LGE. They found that LGE was present in
Dyslipidemia 1 (5) 4 (3) 0.55
only 52 patients during follow-up CMR and dis-
Diabetes mellitus 1 (5) 0 (0) 0.006
Smoking 1 (5) 12 (8) 0.64 appeared in 26.7% of patients. In a recent study,
Family history of CAD 1 (5) 9 (6) 0.87 Martinez-Naharro et al. (12) found regression in LGE
Obesity 0 (0) 10 (7) 0.24 in 42% of patients with amyloid light-chain
Fever 12 (67) 69 (55) 0.34 amyloidosis after specific chemotherapy, supporting
Leukocytes, 103 cells/m2 10  4 94 0.47 the observation that LGE is not a marker of irrevers-
ESR, mm/h 13 (4–21) 18 (4–39) 0.70
ible myocardial damage. Gadolinium-based contrast
CRP, g/ml 6 (1–18) 6 (2–21) 0.39
agents are multicompartmental molecules that move
CMR-I
LVEDVi, ml/m 2
85  19 82  18 0.48
from capillary vessel to the interstitial space with a
LVEF, % 61  8 63  7 0.40 rapid wash-in but a slower wash-out, which is nor-
LVMi, g/m2 74  15 75  16 0.93 mally completed within 5 to 8 min of the injection.
RVEDVi, ml/m2 84  25 83  18 0.95 LGE is caused by all conditions that enlarge intersti-
RVEF, % 59  8 57  11 0.46 tial space, such as replacement fibrosis, edema, and
Pericardial effusion 7 (33) 38 (25) 0.41
proteins overload (amyloidosis), and consequently,
Segments with edema 3 (1–5) 2 (1–4) <0.05
increase the volume of distribution of gadolinium in
Segments with hyperemia 2 (0-3) 2 (0-3) 0.89
myocardium, causing a much slower wash-out (13). In
Segments with LGE 3 (2–5) 2 (1–4) 0.09
Extent of LGE (% of LV mass) 6.2 (3.5–10.9) 6 (2–4) 0.28 the case of acute myocarditis, interstitial space could
CMR-II be increased not only by replacement fibrosis but also
Increased LV EDVi 12 (55) 73 (44) 0.36 by the presence of edema and inflammatory cell
Increased mass index 6 (29) 34 (23) 0.55 infiltration. Gadolinium contrast agents cannot enter
Pericardial effusion (CMR-II) 1 (5) 14 (11) 0.83 freely into viable cell through intact membranes, but
Patients without both LGE and edema 0 (0) 20 (12) 0.13
when macrophages phagocytize necrotic myocytes,
Patients with edema and no LGE, n 0 (0) 0 (0) —
interstitial fluids, including gadolinium, might be
Patients with LGE and edema 3 (14) 28 (17) 0.99
Patients with LGE and no edema 19 (86) 118 (71) 0.13
incorporated into phagosomes. Inflammatory cells
Segments with edema 5 (3–11) 2 (2–4) 0.19 might also obstruct lymphatic vessels slowing the
Patients with increase of edema segments 3 (14) 28 (17) 0.99 wash-out of gadolinium from interstitial space and
Segments with LGE at follow-up 4 (2–5) 2 (1–4) 0.13 participating in the genesis of interstitial edema.
Extent of LGE (% of LV mass) 6 (3.6–10.4) 4 (2–8.3) 0.037 Kolman et al. (14) demonstrated the presence of
Patients with increased LGE extent 12 (54) 33 (19) <0.0001
dilated lymphatic vessel in myocardial segments with
LGE and edema, which was absent in segment with
Values are n (%), mean  SD, or median (25th to 75th percentile).
CMR ¼ cardiac magnetic resonance; EDVi ¼ end-diastolic volume index; EF ¼ ejection fraction; other abbre- LGE and without edema in a case of hypertrophic
viations as in Tables 1 and 2.
cardiomyopathy. Scar shrinkage may also participate
in reduction of LGE extent at follow-up. Finally,
myocarditis for hemodynamically stable patients myocardial hyperemia, which is a common finding in
with preserved EF, enabling the detection of myocarditis, may increase the amount of gadolinium
myocardial edema, hyperemia, and LGE using both that arrives at the involved myocardial segments (15).
conventional and novel mapping techniques (1,2). As Whatever the cause is, in the acute setting of
demonstrated in the present study, at 6 months after myocarditis in the presence of coexisting edema, LGE
symptom onset, myocardial edema disappeared in cannot be considered a definitive marker of irrevers-
the majority of patients, remaining detectable in only ible myocardial damage, and in a non-negligible
16% of them. Interestingly, the LV mass index percentage of patients, it could completely disap-
decreased significantly at follow-up, which may be pear after 6 months. For this reason, myocarditis
related to edema reabsorption and less inflammatory differs from myocardial infarction, where LGE is al-
cell infiltration. The decrease in LV mass was higher ways associated with irreversible myocardial damage,
in patients showing a complete recovery from edema the extent of which usually decreases at follow-up
and LGE than in those with persisting edema and/or due to scar shrinkage, although it never disappears
LGE. Our data demonstrated that LGE in the acute completely (16). In our population, 18 of 139 patients
JACC VOL. 74, NO. 20, 2019 Aquaro et al. 2445
NOVEMBER 19, 2019:2439–48 Prognostic Role of 6-Month Follow-Up CMR in Myocarditis

C ENTR AL I LL U STRA T I O N Prognostic Role of 6-Month Follow-Up CMR in Myocarditis

T2w-STIR LGE Edema & Late Gadolinium Enhancement at

Cardiac Magnetic Resonance-II
CMR-I
100
No Edema and No LGE
90
CMR-II

Survival Probability (%)

80 LGE and Edema

No Edema and No LGE at CMR-II

70

CMR-I 60
LGE with No Edema
50
CMR-II
40
LGE and Edema at CMR-II P < 0.0001
30
0 500 1,000 1,500 2,000
CMR-I Time (Days) from CMR-II
Number at risk:
No Edema & No LGE 20 12 10 8 3
CMR-II
LGE & Edema 30 21 17 12 10
LGE with No Edema at CMR-II LGE with No Edema 137 45 22 14 7

Aquaro, G.D. et al. J Am Coll Cardiol. 2019;74(20):2439–48.

Cardiac magnetic resonance (CMR) was performed within the first week following symptom onset (CMR-I) and after 6 months (CMR-II). At CMR-II, 3 different pre-
sentations were found: 1) the complete absence of edema and late gadolinium enhancement (LGE) (no edema and no LGE, left top); 2) the presence of both edema
and LGE (LGE and edema, left middle); and 3) LGE without edema (LGE with no edema, left bottom). The red arrows identify the presence of LGE, and the blue arrows
indicate the presence of edema. (Right) The Kaplan-Meier survival curves demonstrate that patients with LGE but without edema at CMR-II had a worse prognosis
than those with edema and LGE and than those with complete healing from edema and LGE.

(13%) with an inferior and/or lateral subepicardial
pattern of LGE at CMR-I had a complete disap-
pearance of LGE, whereas this did not occur in any
patients with a midwall septal pattern. In a large
number of patients with acute myocarditis and
preserved EF enrolled in the ITAMY registry, the
midwall septal pattern of LGE was associated with a
worse prognosis than the inferior and/or lateral
subepicardial pattern (6). The results of the present
study may suggest that the midwall septal pattern
could be associated with more aggressive myocar-
ditis and with a minor likelihood of complete re-
covery at follow-up, which may partially explain the
prognostic difference of septal pattern. In other
conditions, such as dilated cardiomyopathy, the
presence of midwall septal LGE may be a secondary
chronic structural change of advanced heart disease,
due to more dilated ventricles with increased
wall stress, and it is generally a marker of worse
prognosis. However, we enrolled stable patients
with preserved LVEF and normal volumes, permit-
ting us to exclude the presence of increased wall
stress. Moreover, the presence of midwall septal
LGE was always associated with edema with the
same pattern distribution at CMR-I, suggesting
acute damage rather than chronic structural
myocardial damage.
Interestingly, in 26 patients (14%), the extent of
LGE even increased at follow-up compared with the
extent in the acute phase, and this was also found in
the absence of myocardial edema. Patients
with increased LGE demonstrated worse survival
free from cardiac events than those with unchanged/
decreased LGE. These findings may suggest that, in
some patients, myocardial damage might continue as
the result of an autoimmune response or multiple
relapse of myocarditis and be associated with disease
progression and worse outcome. Additionally, in
2446 Aquaro et al.
Prognostic Role of 6-Month Follow-Up CMR in Myocarditis
F I G U R E 2 Kaplan-Meier Survival Curves
100
90
80
Survival Probability (%)
70
60
50
40
30
20
10
0 500
Number at risk
LGE Increased 26 16
LGE Decr.\Unch. 161 59
LGE Increased
These Kaplan-Meier survival curves demonstrate that patients with increased late gadolinium enhancement (LGE) had a worse prognosis than
those with decrease or unchanged LGE. Decr ¼ decreased; unch ¼ unchanged.
some patients, the peak of myocardial damage may
not have been in the first week from symptom onset,
but could have had a slow progression in the subse-
quent months. The majority (21 of 26) of patients with
increased LGE had negative edema at CMR-II. Edema
detected by CMR is only the “tip of the iceberg” in the
inflammation process, and we postulate that in the
chronic phase of AM, the autoimmune response
might induce myocardial injury with slower progres-
sion and less aggressiveness than it does in the acute
phase (17). In the chronic phase of AM, CMR may not
be sufficiently sensitive to detect edema using the
conventional T 2-STIR technique. As shown by recent
studies, the T 2 mapping technique with a quantitative
measurement of myocardial T 2 could improve the
detection of myocardial edema and increase the
sensitivity of CMR in both the subacute and chronic
phase of AM (17).
Another interesting result of this study was that
patients with LGE and no edema had a worse prog-
nosis than those with persistent edema. Our data
cannot completely explain this finding. However, we
JACC VOL. 74, NO. 20, 2019
NOVEMBER 19, 2019:2439–48
P = 0.02
1,000 1,500 2,000 2,500 3,000
Time
9 8 4 1 0
39 26 15 5 0
LGE Decreased/Unchanged
can postulate that in patients with LGE and edema
there was active inflammation with the potential to
heal in the future, whereas in those without edema,
LGE represents a definite fibrosis.
This hypothesis is supported by Vermes et al. (18),
who observed that among the CMR criteria of acute
myocarditis (edema, hyperemia, and LGE), only evi-
dence of myocardial edema was associated with an
improvement of systolic function after 12 months,
likely reflecting a recovery of reversibly injured
myocardium.
These data suggest that definite myocardial fibrosis
is associated with a worse outcome in myocarditis as
was demonstrated in all other cardiac conditions.
Finally, in the multivariate analysis, the best inde-
pendent predictors of cardiac events were the pres-
ence of LGE without edema and a midwall septal
pattern of LGE. Other studies evaluated the prog-
nostic role of edema and LGE in myocarditis (6,19).
Particularly, in the study by Gräni et al. (19), the
presence of septal LGE, midwall LGE, and an
abnormal T 2-weighted ratio ($2.0) were predictors of
JACC VOL. 74, NO. 20, 2019 Aquaro et al. 2447
NOVEMBER 19, 2019:2439–48 Prognostic Role of 6-Month Follow-Up CMR in Myocarditis

MACE in multivariate analysis. The present study

T A B L E 4 Univariate and Multivariate Cox Regression Analysis
differs from that of Gräni et al. (19) because we
enrolled only patients with preserved EF, and we Univariate Analysis Multivariate Analysis

performed the CMR-I examination in the first week HR (95% CI) p Value HR (95% CI) p Value

from symptom onset. We enrolled patients with Age 0.60 (0.20–1.80) 0.37
ST-T abnormalities 0.40 (0.12–1.37) 0.15
clinically suspected acute myocarditis and positive
Peak of HS troponin 2.58 (0.74–8.99) 0.14
original Lake Louise criteria. In contrast, in the study
LVEDVi 0.77 (0.12–4.60) 0.77
by Gräni et al. (19), the diagnosis of myocarditis was
LVEF 0.26 (1.08–13.70) 0.038
made not by using the Lake Louise criteria, but by LV mass index 2.27 (0.85–6.04) 0.10
the exclusion of other conditions, and in many pa- RVEDVi 2.09 (0.44–10.00) 0.36
tients, CMR was performed in the subacute phase. RVEF 0.52 (0.12–2.20) 0.38
The prevalence of LGE and edema in the study of Segments with edema at CMR-I 1.35 (0.54–3.39) 0.51

Gräni et al. (19) was lower than the prevalence of LGE extent at CMR-I 1.50 (0.56–4.00) 0.40
Midwall septal LGE 2.68 (1.05–6.80) 0.039 2.84 (1.10–7.20) 0.028
LGE in our study at CMR-I, and in many of the pa-
LVEF at CMR-II 1.76 (0.68–4.60) 0.24
tients in that study, the Lake Louise criteria were
LVEDVi at CMR-II 1.07 (0.23–4.90) 0.93
negative. Segments with edema at CMR-II 1.60 (0.28–9.30) 0.58
In this study, we used original Lake Louise criteria LGE extent at CMR-II 1.07 (1.02–1.12) 0.005
to diagnose AM; recently, however, these criteria LGE without edema at CMR-II 5.58 (1.60–19.10) 0.007 4.50 (1.30–14.50) 0.008
were revised and included new mapping criteria (4). Increased LGE extent at CMR-II 2.60 (1.05–6.90) 0.037

New criteria were based on a 2-out-of-2 approach, and

the diagnosis is made when a T 2-based criterion
(edema at T 2-STIR or T 2 mapping) is summed with a
T 1 -based criterion (T 1 mapping, LGE, ECV). In our
study, 96% of patients had a combination of edema at
T 2-STIR and LGE, permitting confirmation of the
diagnosis of AM using these new criteria. Moreover,
in the revised Lake Louise document as well as in the
recent meta-analysis by Kotanidis et al. (20), the most
accurate combination of CMR parameters for diag-
nosis of AM was T1 mapping plus LGE, which, para-
doxically, was not included as a possible diagnostic
criterion in the revised criteria (4,20). Finally, the
original criteria had good accuracy in patients similar
to those included in our study (infarct-like presenta-
tion and preserved EF), whereas the diagnostic
accuracy of the new criteria is still unknown in a real-
life population.
STUDY LIMITATIONS. The main limitation of this
study is the absence of a T 1 and T 2 mapping evalua-
tion. Recent evidence suggests that native T1 and T 2
measurements may be used to evaluate the conva-
lescent phase of AM to detect complete healing from
inflammation, which can be confirmed when T1 and
T 2 values return to normal range (21). However, in
this multicenter prospective study, mapping tech-
niques were not available in all the scanners at the
time of enrollment.
In this study, we enrolled only patients with AM,
preserved EF, and an infarct-like presentation. We
cannot exclude the possibility that different results
could be found in patients with AM and LV dysfunc-
tion and in those with heart failure or arrhythmic
CI ¼ confidence interval; HR ¼ hazard ratio; HS ¼ high sensitivity; other abbreviations as in Tables 1 to 3.
presentation. However, the choice to enroll only pa-
tients with preserved EF and infarct-like presentation
was made because CMR is very accurate for diag-
nosing AM only in these type of patients (1,2). Further
studies should assess the prognostic role of CMR in
patients with heart failure and arrhythmic
presentation.
CONCLUSIONS
In the acute setting of AM, LGE does not mean defi-
nite fibrosis in the presence of edema, and it may
disappear during follow-up in a significant percent-
age of patients. CMR should be performed in all pa-
tients to evaluate the evolution of signs of
myocarditis at 6 months after symptom onset. The
presence of LGE without edema 6 months after
symptoms is associated with a worse prognosis,
particularly when it involves the midwall layer of the
septum. LGE without edema could represent definite
fibrosis, whereas the persistence of edema, found in
approximately one-third of patients, is a marker of
active inflammation and could be associated with a
residual chance of complete recovery.
ADDRESS FOR CORRESPONDENCE: Dr. Giovanni
Donato Aquaro, Fondazione Toscana G. Monasterio,
Via Giuseppe Moruzzi, 1, 56124 Pisa, Italy. E-mail:
aquaro@ftgm.it.
2448 Aquaro et al. JACC VOL. 74, NO. 20, 2019

Prognostic Role of 6-Month Follow-Up CMR in Myocarditis NOVEMBER 19, 2019:2439–48

PERSPECTIVES

COMPETENCY IN PATIENT CARE AND edema, particularly in the midwall of the interventricular
PROCEDURAL SKILLS: In patients with acute myocar- septum, is associated with adverse cardiac events.
ditis, comparing CMR images after 6 months with those
obtained during the first week after the onset of symp- TRANSLATIONAL OUTLOOK: Future studies should

toms allows assessment of the evolution of myocardial characterize the evolution and prognostic value of native

edema. In some patients, LGE disappears during this in- T1 and T2 extracellular volume mapping in patients with

terval, indicating that its early appearance does not imply acute myocarditis.

irreversible myocardial damage. Persistent LGE without

REFERENCES

1. Francone M, Chimenti C, Galea N, et al. CMR
sensitivity varies with clinical presentation and
extent of cell necrosis in biopsy-proven acute
myocarditis. J Am Coll Cardiol Img 2014;7:254–63.
2. Lurz P, Eitel I, Adam J, et al. Diagnostic per-
formance of CMR imaging compared with EMB in
patients with suspected myocarditis. J Am Coll
Cardiol Img 2012;5:513–24.
3. Friedrich MG, Sechtem U, Schulz-Menger J,
et al. Cardiovascular magnetic resonance in
myocarditis: A JACC White Paper. J Am Coll Car-
diol 2009;53:1475–87.
4. Ferreira VM, Schulz-Menger J, Holmvang G,
et al. Cardiovascular magnetic resonance in non-
ischemic myocardial inflammation: expert recom-
mendations. J Am Coll Cardiol 2018;72:3158–76.
5. Anzini M, Merlo M, Sabbadini G, et al. Long-
term evolution and prognostic stratification of
biopsy-proven active myocarditis. Circulation
2013;128:2384–94.
6. Aquaro GD, Perfetti M, Camastra G, et al., for
the Cardiac Magnetic Resonance Working Group of
the Italian Society of Cardiology. Cardiac MR with
late gadolinium enhancement in acute myocarditis
with preserved systolic function: ITAMY Study.
J Am Coll Cardiol 2017;70:1977–87.
7. Mahrholdt H, Wagner A, Deluigi CC, et al. Pre-
sentation, patterns of myocardial damage, and
clinical course of viral myocarditis. Circulation
2006;114:1581–90.
Pericardial Diseases. Eur Heart J 2013;34:
2636–48.
9. Abdel-Aty H, Boyé P, Zagrosek A, et al. Diag-
nostic performance of cardiovascular magnetic
resonance in patients with suspected acute
myocarditis: comparison of different approaches.
J Am Coll Cardiol 2005;45:1815–22.
10. Perfetti M, Malatesta G, Alvarez I, et al. A fast
and effective method to assess myocardial hy-
peremia in acute myocarditis by magnetic reso-
nance. Int J Cardiovasc Imaging 2014;30:629–37.
11. Hundley WG, Bluemke DA, Finn JP, et al. ACCF/
ACR/AHA/NASCI/SCMR 2010 expert consensus
document on cardiovascular magnetic resonance:
a report of the American College of Cardiology
Foundation Task Force on Expert Consensus Doc-
uments. J Am Coll Cardiol 2010;55:2614–62.
12. Martinez-Naharro A, Abdel-Gadir A, Treibel TA,
et al. CMR-verified regression of cardiac AL amy-
loid after chemotherapy. J Am Coll Cardiol Img
2018;11:152–4.
13. Kim RJ, Wu E, Rafael A, et al. The use of
contrast-enhanced magnetic resonance imaging to
identify reversible myocardial dysfunction. N Engl
J Med 2000;343:1445–53.
14. Kolman L, Welsh DG, Vigmond E, et al.
Abnormal lymphatic channels detected by T2-
weighted MR imaging as a substrate for ventricu-
lar arrhythmia in HCM. J Am Coll Cardiol Img 2016;
9:1354–6.
16. Kim RJ, Fieno DS, Parrish TB, et al. Rela-
tionship of MRI delayed contrast enhancement
to irreversible injury, infarct age, and contrac-
tile function. Circulation 1999;100:1992–2002.
17. von Knobelsdorff-Brenkenhoff F, Schüler J,
Dogangüzel S, et al. Detection and monitoring of
acute myocarditis applying quantitative cardio-
vascular magnetic resonance. Circ Cardiovasc Im-
aging 2017;10:e005242.
18. Vermes E, Childs H, Faris P, Friedrich GF. Pre-
dictive value of CMR criteria for LV functional
improvement in patients with acute myocarditis.
Eur Heart J Cardiovasc Imaging 2014;15:1140–4.
19. Gräni C, Eichhorn C, Bière L, et al. Prognostic
value of cardiac magnetic resonance tissue char-
acterization in risk stratifying patients with sus-
pected myocarditis. J Am Coll Cardiol 2017;70:
1964–76.
20. Kotanidis CP, Bazmpani MA, Haidich AB,
Karvounis C, Antoniades C, Karamitsos TD. Diag-
nostic accuracy of cardiovascular magnetic reso-
nance in acute myocarditis: a systematic review
and meta-analysis. J Am Coll Cardiol Img 2018;11:
1583–90.
21. Hinojar R, Foote L, Arroyo Ucar E, et al.
Native T1 in discrimination of acute and
convalescent stages in patients with clinical
diagnosis of myocarditis: a proposed diagnostic
algorithm using CMR. J Am Coll Cardiol Img
2015;8:37–46.

8. Caforio AL, Pankuweit S, Arbustini E, et al.
Current state of knowledge on aetiology, diag-
nosis, management, and therapy of myocarditis: a
position statement of the European Society of
Cardiology Working Group on Myocardial and
15. Friedrich MG, Strohm O, Schulz-Menger J,
Marciniak H, Luft FC, Dietz R. Contrast media-
enhanced magnetic resonance imaging visualizes
myocardial changes in the course of viral
myocarditis. Circulation 1998;97:1802–9.
KEY WORDS acute myocarditis, cardiac
magnetic resonance, late gadolinium
enhancement, myocardial edema, prognosis