Author + information
- Received September 13, 2017
- Revision received May 21, 2018
- Accepted May 31, 2018
- Published online January 7, 2019.
- Sung-Ji Park, MD, PhDa,∗∗∗ (, )
- Sung Woo Cho, MD, PhDb,∗,
- Sung Mok Kim, MD, PhDc,∗ (, )
- Joonghyun Ahn, MSd,
- Keumhee Carriere, PhDd,e,
- Dong Seop Jeong, MD, PhDf,
- Sang-Chol Lee, MD, PhDa,
- Seung Woo Park, MD, PhDa,
- Yeon Hyeon Choe, MD, PhDc,
- Pyo Won Park, MD, PhDf and
- Jae K. Oh, MDa,g
- aDepartment of Medicine, Division of Cardiology, Cardiovascular Imaging Center, Heart Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
- bDepartment of Medicine, Division of Cardiology, Inje University College of Medicine, Seoul Paik Hospital, Seoul, Republic of Korea
- cDepartment of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
- dBiostatistics and Clinical Epidemiology Center, Research Institute for Future Medicine, Samsung Medical Center, Seoul, Republic of Korea
- eDepartment of Mathematical and Statistical Sciences, University of Alberta, Edmonton, Alberta, Canada
- fDepartment of Thoracic Surgery, Heart Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Republic of Korea
- gDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- ↵∗Addresses for correspondence:
Dr. Sung Mok Kim, Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81, Irwon-ro, Gangnam-gu, Seoul 06351, Republic of Korea.
- ↵∗∗Dr. Sung-Ji Park, Division of Cardiology, Department of Internal Medicine, Cardiovascular Imaging Center, Heart Vascular Stroke Institute, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81, Irwon-ro, Gangnam-gu, Seoul 06351, Republic of Korea.
Objectives This study assessed diffuse myocardial fibrosis (MF) by cardiac magnetic resonance (CMR) imaging and speckle-tracking echocardiography (STE) in patients with severe aortic stenosis (AS) and validated findings by using histologic confirmation of MF.
Background MF is a concomitant pathologic finding related to hypertrophic response in severe AS. It would be beneficial to have reliable imaging methods to assess MF.
Methods CMR and STE were performed in 71 consecutive patients with severe AS before aortic valve replacement. The extracellular volume (ECV) and native T1 values obtained by CMR and global longitudinal strain (GLS) values by STE were measured. The degree of MF was quantified by using Masson trichrome stain in myocardial biopsy specimens obtained intraoperatively. The study population was divided into 3 groups according to the degree of MF on histology (mild, moderate, and severe MF).
Results The severe MF group had a higher incidence of heart failure (HF) and diastolic dysfunction than the mild and moderate MF groups. The ECV (r = 0.465; p < 0.0001), GLS (r = 0.421; p = 0.0003), and native T1 (r = 0.429; p = 0.0002) values were significantly correlated with the degree of MF. GLS was moderately correlated with ECV (r = 0.455; p = 0.0001) and less with the native T1 (r = 0.372; p = 0.0014) value. The model using ECV (R2 = 0.44; Akaike Information Criterion [AIC] = 55.8) was found to predict the degree of MF most accurately than that with GLS (R2 = 0.35; AIC = 66.84) and the native T1 (R2 = 0.36; AIC = 66.18) value. The secondary endpoint of interest was clinical outcome of a composite of total mortality, admission for HF, or development of HF symptoms. During follow-up (median: 4.6 years), and there were 16 clinical events. Although statistically insignificant, ECV is more closely related to prediction of the clinical outcome than native T1 or GLS.
Conclusions ECV as assessed by CMR could be an ideal surrogate marker for diffuse MF in patients with severe AS among all 3 models considered.
Severe aortic stenosis (AS) is a progressive disease which affects the myocardium and ultimately leads to myocyte death, myocardial fibrosis (MF), and heart failure (HF). Focal midwall MF has been demonstrated in patients with AS and is associated with impaired cardiac function and adverse clinical outcomes (1–3). For this reason, early detection of focal MF by an imaging modality such as late gadolinium enhancement (LGE) of cardiac magnetic resonance imaging (CMR) has been widely investigated (3,4). However, focal MF is believed to be irreversible after aortic valve replacement (AVR), whereas diffuse interstitial fibrosis of the myocardium has been reported to be reversible and has gained increasing interest as a potential treatment target (5). Therefore, reliable assessment of diffuse MF by noninvasive multimodality imaging is clinically important for managing patients with severe AS.
T1 mapping in CMR has emerged as a tool for quantifying the myocardial extracellular volume (ECV) fraction, which has been recently reported as a promising marker for diffuse interstitial fibrosis of the myocardium in AS patients (5–8). Native T1 mapping has been shown to provide a noninvasive estimation of diffuse MF and to correlate with subclinical myocardial dysfunction in asymptomatic patients with AS (9).
Furthermore, global longitudinal strain (GLS), which reflects intrinsic myocardial contractility and is assessed by speckle-tracking echocardiography (STE), was found to be associated with increased adverse events in patients with severe AS (10).
However, there are few data regarding the direct and simultaneous comparison and validation between noninvasive multimodality imaging for assessment of diffuse MF and histologically confirmed MF in severe AS. The present study sought to assess diffuse MF using the ECV and native T1 values obtained from CMR and GLS from STE and to compare these results with the degree of MF found from histology examination in patients with severe AS undergoing AVR.
Patients who had severe AS with preserved left ventricular ejection fraction (LVEF ≥50%) were enrolled prospectively at a single tertiary center, Samsung Medical Center in South Korea, between February 2012 and March 2015. Patients underwent comprehensive echocardiography, STE, and CMR on the same day. Severe AS was defined as aortic valve area index <0.6 cm2/m2, in accordance with previously published guidelines (11). AVR based on clinical indications were performed 1 day after noninvasive imaging.
Patients with any of the following criteria were excluded: age <18 years; LVEF <50%; presence of another concomitant valvular disease of at least moderate severity; previous AVR; obstructive epicardial coronary artery disease (>30% luminal stenosis in at least 1 coronary artery on coronary angiography); history of myocardial infarction or acute coronary syndrome; any absolute contraindication to CMR; or estimated glomerular filtration rate <30 ml/min/173 m2 (Figure 1). After CMR, the patients with subendocardial enhancement pattern on LGE images were excluded.
The Institutional Review Board of Samsung Medical Center approved this study, and all subjects gave written informed consent before the investigation.
Definition of symptoms
To classify patients’ symptoms, all patients’ initial symptoms assessed by their primary physicians and found in their medical records were carefully reviewed by 1 cardiologist (S.J.P). Dyspnea on exertion (DOE) was defined as greater than grade II in the New York Heart Association (NYHA) functional classification. Angina was defined as exertional chest pain. Those with chest heaviness or chest discomfort were included in the angina classification. Those with dizziness or presyncope were included in the syncope classification. In this study, HF was defined as HF with preserved ejection fraction (HFpEF), in accordance with a recently published guideline (12).
Two-dimensional (2D) echocardiography was performed using commercially available equipment (Vivid 9, GE Medical Systems, Milwaukee, Wisconsin). All parameters were measured using the current American Society of Echocardiography and European Association of Echocardiography guidelines (13). The mean transaortic pressure gradient and peak transaortic velocity were measured in all views possible, that is, apical 5- or 3-chamber and right parasternal views, and the highest values were used for analysis. The time-velocity integrals at the aortic valve and LV outflow tract levels were obtained by using continuous wave and pulse wave Doppler echocardiography, respectively, and the aortic valve area was calculated using the continuity equation using the parameters mentioned previously. The average of 3 consecutive Doppler signals was used.
Analysis of the 2D STE images was performed using software (EchoPAC, GE Ultrasound, Haifa, Israel). Loops of 3 consecutive cardiac cycles for 2D STE images were obtained. Two-dimensional data were analyzed using EchoPAC version 113.0.4 software (GE Vingmed Ultrasound AS, Horten, Norway) by an experienced investigator blinded to all clinical information for the patients with severe AS. Speckle-tracking analysis was performed using dedicated wall motion tracking software, Automated Function Imaging for 2D imaging (GE Vingmed). Segments with poor quality tracking or aberrant curves despite manual adjustment were removed from analysis. Longitudinal strain was computed from the 2D data set. GLS was acquired by average regional strain curves (16-segment model for 2D STE) (Figure 2A). For all strain components, peak systole, and time to peak strain were defined using regional strain curves (14).
CMR images were obtained using a 1.5-T CMR system (Syngo MR B17 version, Magnetom Avanto, Siemens Medical Solutions, Erlangen, Germany) equipped with a maximum strength gradient of 45 mT/m, a 200 mT/m per ms slew rate, and a 32-channel body array coil. CMR scans consisted of localizing images (axial, coronal, and sagittal), cine scans (2-, 3-, and 4-chamber views and a short-axis view), pre-T1 mapping, perfusion scans (both stress and rest scans with an intravenous infusion of 0.1 mmol/kg gadobutrol at an injection rate of 3 ml/s, followed by a 30-ml saline flush), post-T1 mapping, and LGE scans. All examinations were carried out by experienced technicians and supervised by an experienced radiologist.
The present study describes image acquisition parameters by focusing on T1 mapping images, which were acquired in short-axis before and at 15 min after contrast administration. A short-axis section at the mid-ventricular level was acquired using modified Look-Locker inversion-recovery (MOLLI). Pre-contrast MOLLI consisted of 5 images in the first Look-Locker segment and 3 images in the second segment (“5-3” protocol). Finally, 8 images acquired during 11 heartbeats were obtained, and in-line motion correction and map generation were applied. Post-contrast MOLLI was composed of 4 images in the first Look-Locker segment and 3 images in second and 2 images in the third segment (“4-3-2” protocol). Finally, 9 images acquired during 11 heartbeats were obtained, and in-line motion correction and map generation were applied. The following readout parameters were used: section thickness: 8 mm; flip angle: 35; field of view (FOV): 360 × 307; effective TI (TIeff): 120 ms; TIeff. increment: 80 ms; voxel size: 1.87 × 1.88 × 8 mm; TR/TE: 2.4/1.01 ms; partial fourier: 7/8; parallel imaging factor: 2.
Measurement of the ECV and native T1 values
The T1 maps were generated from the CMR workstation after in-line motion correction just after image acquisition. Two blinded reviewers (with 9 and 6 years of experience in CMR, respectively) independently reviewed all T1 mapping sequences. The first reviewer (9 years of experience) performed a second review of the sequences with 10 randomly selected patients after an interval of 2 weeks to assess intraobserver agreement. Regions of interest were drawn manually in the blood and septum, excluding the myocardium with LGE for T1 measurements (8,15). Regions of interest in the septum were chosen for myocardial T1 because this area corresponds to the site of myocardial biopsy. We drew the regions of interest on the compact myocardium and did not include the border of the myocardium because the border between the myocardium and the LV cavity showed gradual changes in T1 values due to both partial volume averaging artifacts and registration error, even after motion correction. To measure the T1 value of blood, a circular regions of interest was positioned in the LV cavity, avoiding papillary muscle (Figures 2A and 2B). The ECV of the myocardium was calculated as follows: ECV% = (ΔR1m/ΔR1b) × (1 − hematocrit level) × 100, where R1 is 1/T1, R1m is R1 in the myocardium, R1b is R1 in the blood, and ΔR1 is the change in relaxivity. The change in relaxivity (ΔR1) was determined using the following equation: ΔR1 = R1post − R1pre, where R1post and R1pre are the R1 after and before gadolinium chelate administration, respectively (16). Blood samples were taken 1 to 3 h prior to CMR to determine hematocrit and creatinine concentrations. Interobserver agreement between 2 different experienced observers was assessed on different days for 10 randomly selected patients.
Histologic analysis: Visual assessment and morphometry
In patients undergoing AVR based on clinical indications, 73 biopsy specimens from 73 patients with severe AS were obtained from the basal LV septum just after removal of the diseased aortic valve. Briefly, 1 intraoperative myocardial biopsy sample with a volume of at least 10 mm3 (range: 10 to 20 mm3) was obtained and processed using Masson’s trichrome staining for evaluating MF. Two samples with only endocardial tissue were excluded, resulting in a total of 71 biopsy samples suitable for analysis. Immunohistochemically stained specimens were digitized using the Aperio ScanScope slide scanner (Leica, Wetzlar, Germany). Quantification of the MF area was performed using a positive pixel count algorithm using Aperio ImageScope software (17). The Aperio ScanScope allowed scanning and quantification of the entire slide image, using a 20× objective lens with a numerical aperture of 0.75. The positive pixel count algorithm detected negative or white pixels (representing lipid vacuoles/steatosis) adequately after setting the appropriate parameters (hue: 0.1; hue width: 0.999; intensity threshold: 255). White spaces from frozen section artifacts created by chattering of the blade and lumens of blood vessels were excluded. Interobserver agreement between 2 experienced observers blinded to the T1 mapping and clinical data was assessed examining 30 different randomly selected samples per biopsy specimen.
All data were analyzed using R version 3.2.0. (R Foundation for Statistical Computing, Vienna, Austria) and SPSS version 20.0 (SPSS, Inc., Chicago, Illinois) software. Continuous variables are mean ± SD. After patients were divided into 3 groups of approximately equal numbers (n = 23 to 24 per group) according to the tertiles of MF (mild, moderate, severe), continuous variables were compared between these groups using the Kruskal-Wallis test. Categorical variables are numbers (percentages) and were compared using chi-square and Fisher exact tests. Bivariate correlation analysis of the ECV, GLS, and native T1 values and the degree of MF was performed to verify any concerns for collinearity. Pearson correlation coefficient was used as the measure of correlation. To identify independent predictors of the degree of MF, the authors performed multivariate stepwise regression analysis separately for the ECV, GLS, and native T1 values (Model 1 for the ECV, model 2 for GLS, model 3 for the native T1 values), along with those demographic and clinical variables that demonstrated a univariate p value <0.20. The final models were obtained by removing insignificant or redundant variables. To satisfy the model assumptions, a log-transformation of the MF value was performed. Diagnostic residual analysis was also carried out to ensure the model’s goodness of fit. The model’s performance scores, such as the R2, Akaike Information Criterion (AIC), F-test, and adjusted R2 scores, were used to determine the best marker among the ECV, GLS, and native T1 for the degree of MF from these models.
Clinical outcomes of total mortality, admission for HF, and development of HF symptoms were considered a composite event due to small sample size. The composite event was assessed from a simple Cox analysis using ECV, native T1, and GLS and compared their c-indexes of how well each marker predicted clinical outcome. In addition, a Kaplan-Meier survival curve was produced to delineate the survival experiences by ECV tertiles, along with a log-rank test result. Establishment of a trend to show an association between high ECV and high hazard was also attempted. The proportional assumption was verified by checking Schoenfeld residuals.
A 2-tailed p value of <0.05 was considered statistically significant.
Clinical, echocardiographic, and CMR characteristics
A total of 71 patients were divided into 3 groups according to the degree of MF found on histology (mild, moderate, or severe) (Figure 3). The mean value for the degree of MF was 14.16 ± 2.60 in the mild group, 21.68 ± 2.37 in the moderate group, and 37.75 ± 9.14 in the severe group. Hypertension was more prevalent in the moderate and severe MF groups than in the mild MF group. In addition, the severe MF group had more DOE, a higher incidence of HF, and a higher concentration of serum creatinine than the mild and moderate MF groups. There were no significant differences in LVMI among the 3 groups. The baseline clinical characteristics of the study population are summarized in Table 1. Among the echocardiographic parameters, the severe MF group had lower e′ and a higher E/e′ ratio, reflecting more severe diastolic dysfunction and a lower GLS than the mild and moderate MF groups (Table 2). The severe MF group also showed a higher ECV and native T1 value than the mild and moderate MF groups (Table 3).
Relationships between the degree of MF and the ECV, GLS, or native T1 value
Among the echocardiographic and CMR parameters, systolic left ventricular internal dimension (LVIDs) (r = 0.26; p = 0.03); e′ (r = −0.39; p < 0.001), E/e′ ratio (r = 0.33; p = 0.006), left ventricular end-systolic volume (LVESV) (r = 0.24; p = 0.04), LV mass (r = 0.33; p = 0.005), left ventricular mass index (LVMI) (r = 0.30; p = 0.01), GLS (r = 0.42; p = 0.0003), ECV (r = 0.47; p < 0.0001), and native T1 (r = 0.43; p = 0.0002) values were significantly correlated with the severity of MF (Supplemental Table 1, Figure 4). In addition, ECV, GLS, and native T1 values were correlated with each other, and GLS was moderately correlated with ECV (r = 0.46; p = 0.0001) and less with the native T1 value (r = 0.37; p = 0.0014) (Figure 5). Therefore, we proceeded with modeling MF separately for each indicator, ECV, GLS, and native T1, and no two are in the same model.
In multivariate stepwise linear regression analysis, the model with ECV (R2 = 0.44; p < 0.001) was the best in predicting the degree of MF compared with those with GLS (R2 = 0.35; p = 0.001) or native T1 (R2 = 0.36; p = 0.001) value (Table 4).
The intraclass correlation coefficient (ICC), along with its respective 95% confidence intervals (CIs), was used to determine interobserver and intraobserver reliability. Excellent interobserver agreement was found for histologic fibrosis (ICC = 0.937; 95% CI: 0.895 to 0.946). Excellent interobserver and intraobserver agreement was found for the T1 value (ICC = 0.968; 95% CI: 0.873 to 0.992 and ICC = 0.992; 95% CI: 0.969 to 0.998, respectively), and GLS (ICC = 0.927; 95% CI: 0.846 to 0.965 and ICC = 0.942; 95% CI: 0.879 to 0.973, respectively).
The secondary endpoint of interest was clinical outcomes of a composite of total mortality, admission for HF, or development of HF symptoms. Median follow-up was 4.6 years. The study population was divided into 3 groups according to ECV values (lowest ECV tertile, mid ECV tertile, and highest ECV tertile) for clinical outcomes. There were 16 clinical events (2 deaths and 14 admission for HF or development of HF symptoms). Two deaths occurred in those patients in the highest ECV tertile group. The total number of events also occurred more frequently in patients in the mid and highest ECV tertile groups than in the lowest ECV tertile group (3 years survival rate of 0.911 vs. 0.923 vs. 0.727 for lowest, mid-, and highest ECV tertile groups, respectively) (Figure 6). Although the data were too sparse due to small sample size to show statistical significance, it was evident that ECV is more closely related to prediction of the clinical outcome than native T1 or GLS values (hazard ratios [HR] were 1.13, 1.08, and 1.00, respectively, with p values of 0.18, 0.28, and 0.56, respectively for ECV, GLS, and native T1 values). The model with ECV had the highest c-index at 0.62. Furthermore, we considered the trend hypothesis that high ECV was associated with high hazard. The chi-square trend test statistic was 3.75 with a p value of 0.053 based on ECV tertiles and was marginally significant. Hence, we suggest that an HR of 1.13 (95% CI: 0.94 to 1.35) is statistically marginal but clinically significant and positive and that high ECV is associated with high hazard to clinical outcomes of mortality and HF (Figure 6).
Among various imaging parameters for the assessment of MF, we found that the ECV by CMR could be an ideal surrogate marker of diffuse MF in patients with severe AS. The principal findings of this study are as follows. 1) The ECV, GLS, and native T1 values were all significantly correlated with the histologically measured degree of MF. 2) The ECV was a strong independent predictor of the degree of MF and showed the highest accuracy to identify severe MF than GLS and native T1 values. 3) ECV and native T1 values were moderately correlated with GLS. 4) The degree of MF was associated with the clinical symptoms and imaging parameters of HFpEF in patients with severe AS. 5) ECV is more closely related to prediction of the clinical outcome than native T1 or GLS, although it was statistically insignificant.
Endomyocardial biopsy is the gold standard for the assessment of MF, but it is invasive and, hence, not suitable for serial follow-up. For accurate noninvasive assessment of diffuse MF, various imaging modalities have been investigated and validated by comparing results with the extent of MF seen on histology (16,18). Notably, the present study provides the first evidence that direct and simultaneous comparisons between CMR and strain imaging for assessment of diffuse MF and compared them to histologically confirmed MF in severe AS. To the authors' knowledge, this is the largest study performed so far to compare multimodality imaging-derived MF parameters with the histologic assessment of MF in patients with severe AS.
Multimodality imaging of myocardial fibrosis in severe AS
Myocardial T1 mapping has emerged as a novel CMR technique to assess diffuse MF (5,19). T1 mapping provides improved myocardial characterization through its ability to quantify signal intensity for each voxel in the myocardium, generating a parametric T1 map (5). The clinical utility of myocardial T1 mapping has been well reported in AS patients, with numerous studies demonstrating a correlation between this marker of diffuse MF and severity of AS, LV mass, and cardiac function (6,20). So far, 4 different T1 mapping techniques, including native T1, post-contrast T1, partition coefficient (λ), and ECV, have been investigated, with each technique presenting its unique advantages and limitations (5). Among them, ECV seems to be the most promising technique in assessing diffuse MF due to excellent scan-rescan reproducibility, which is necessary when assessing treatment-related responses and disease progression (5). Indeed, the present study shows that ECV is a powerful independent predictor of the degree of MF with highest accuracy for identifying severe MF compared with GLS and native T1 values.
Recently, myocardial deformation imaging, such as strain using STE, has been reported as a sensitive modality for assessing intrinsic myocardial contractility and the functional consequences of MF (21). Specifically, longitudinal deformation assessed by 2D STE has demonstrated incremental prognostic value in patients with AS (22). In a recent study, both 2D and 3D GLS were associated with increased adverse outcomes, but the latter provided a higher accuracy in predicting adverse outcomes in asymptomatic patients with severe AS and preserved left ventricular ejection fraction (10). However, there is a lack of evidence to show how this functional parameter relates to and might complement direct assessment of MF using CMR. Also, there has been no direct confirmation that strain truly reflects histologically confirmed fibrosis in patients with severe AS. In this study, we demonstrated that ECV, GLS, and native T1 values are correlated with each other and that ECV was more strongly correlated with GLS than with native T1 value. Therefore, the ECV obtained using CMR could be an ideal surrogate marker for predicting diffuse MF and its functional consequences in patients with severe AS.
Symptoms and degree of fibrosis in severe AS
LV systolic dysfunction is a late manifestation of ventricular decompensation and MF in patients with severe AS. However, a considerable portion of patients with preserved LVEF and mild MF in the present study complained of clinical symptoms such as DOE. In the current study, the severe MF group had a higher incidence of DOE and HF. However, there were no significant differences in angina symptoms among the different fibrosis groups. The authors' previous study (23) reported why specific symptoms develop and whether the severity of AS (within the range of severe AS) or the cardiac hemodynamic response to AS determined the occurrence of specific symptoms. In patients with severe AS who developed dyspnea, a markedly altered LV diastolic function with increased filling pressure occurred. Also, the present authors demonstrated that the development of angina is related to impaired coronary microvascular dysfunction with LV hypertrophy in severe AS without obstructive coronary artery disease (24). In the current study, the degree of MF was associated with the incidence of HFpEF and the imaging parameters of HFpEF such as e′, E/e′ ratio, LV mass, and LVMI. Furthermore, the ECV, GLS, and native T1 values were positively correlated with levels of N-terminal pro–B-type natriuretic peptide. Findings from the current study were concordant with those of the present authors' previous findings. The degree of MF is a key pathogenic component in patients with severe AS who experience DOE along with altered LV diastolic function and increased LV filling pressure. Therefore, early detection of MF and risk stratification in AS patients by CMR and/or strain echocardiographic imaging may be helpful in identifying an optimal time for AVR and improving clinical outcomes.
This study had several limitations. First, the study population was composed of a relatively small number of patients enrolled in a single center. Before CMR for assessment of MF can be used routinely in clinical practice, further large prospective studies are required. However, the present study enrolled the largest number of patients with severe AS that had myocardial biopsy specimens and simultaneously performed CMR and STE. Second, the present study enrolled relatively more patients with severe AS than previous studies (1,9). For obtaining evidence of the clinical utility of CMR for early detection of MF, future studies may be conducted in moderate, less severe and asymptomatic patients with severe AS. Third, patients with severe AS with LV systolic dysfunction were not included in the present study. Therefore, results are not applicable to patients with LV systolic dysfunction. Forth, how the presence of focal fibrosis (previously known as silent myocardial infarction) was differentiated from diffuse fibrosis (presumably related to AS) may not be clear. After CMR, those patients who had subendocardial enhancement pattern of LGE were excluded. Regions of interest were drawn manually in the blood and in the septum, excluding the myocardium with LGE for T1 measurements. Also, the presence of myocardial infarction was confirmed by histologic analysis, based on an experienced pathologist’s report that “there was no histologic evidence of myocardial infarction in 71 histologic samples.” Although focal fibrosis cannot be completely excluded in this study after CMR and histology analysis, we believe that focal fibrosis may not affect the findings. Finally, the event rate of the present study population was lower than others (1,2). This may reflect ethnic differences in the study population and the possibility that the clinical follow-up duration was not long enough. In the future, the prognostic value of MF imaging markers, including ECV, T1 value, and GLS, will be established in multicenter studies with larger populations.
In patients with severe AS, ECV showed the highest accuracy for identifying severe MF than GLS and native T1 values. Moreover, the degree of MF was associated with the clinical symptoms and imaging parameters of HFpEF in patients with severe AS. Although a larger clinical study is needed to confirm the findings of the present study, the results suggest that early detection of MF and risk stratification in AS patients by multimodality imaging are crucial for improving clinical outcomes.
COMPETENCY IN MEDICAL KNOWLEDGE: ECV is a powerful independent predictor of MF, and the degree of MF was associated with the clinical symptom and imaging parameters of HFpEF in patients with severe AS. Assessment of the myocardial fibrosis through multimodal imaging allows risk stratification in patients with severe AS before disease progression for improving clinical outcomes.
TRANSLATIONAL OUTLOOK: The present authors did not evaluate the improvement of clinical outcomes after multimodal imaging in patients with severe AS. Therefore, additional research is needed to determine whether multimodal imaging for myocardial fibrosis improves clinical outcomes in patients with severe AS with preserved LV systolic function.
The authors acknowledge MiJung Oh (Molecular Translational Research Center, Samsung Biomedical Research Institute [RDCS]) for histological assessment, and Hye Rim Yun, RDCS, for help in analyzing strain data.
↵∗ Drs. Park and Cho contributed equally to this work and are joint first authors.
All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic stenosis
- aortic valve replacement
- cardiac magnetic resonance imaging
- extracellular volume
- ejection fraction
- global longitudinal strain
- heart failure preserved ejection fraction
- late gadolinium enhancement
- myocardial fibrosis
- speckle tracking echocardiography
- Received September 13, 2017.
- Revision received May 21, 2018.
- Accepted May 31, 2018.
- 2019 American College of Cardiology Foundation
- Dweck M.R.,
- Joshi S.,
- Murigu T.,
- et al.
- Weidemann F.,
- Herrmann S.,
- Stork S.,
- et al.
- Azevedo C.F.,
- Nigri M.,
- Higuchi M.L.,
- et al.
- chin C.W.,
- Pawade T.A.,
- Newby D.E.,
- Dweck M.R.
- Singh A.,
- Horsfield M.A.,
- Bekele S.,
- Khan J.N.,
- Greiser A.,
- McCann G.P.
- Kockova R.,
- Kacer P.,
- Pirk J.,
- et al.
- Nagata Y.,
- Takeuchi M.,
- Wu V.C.,
- et al.
- Ponikowski P.,
- Voors A.A.,
- Anker S.D.,
- et al.
- Cho E.J.,
- Park S.J.,
- Yun H.R.,
- et al.
- Puntmann V.O.,
- Voigt T.,
- Chen Z.,
- et al.
- de Meester de Ravenstein C.,
- Bouzin C.,
- Lazam S.,
- et al.
- Moon J.C.,
- Messroghli D.R.,
- Kellman P.,
- et al.
- Hoffmann R.,
- Altiok E.,
- Friedman Z.,
- et al.
- Park S.J.,
- Enriquez-Sarano M.,
- Chang S.A.,
- et al.
- Ahn J.H.,
- Kim S.M.,
- Park S.J.,
- et al.