Author + information
- Martin S. Maron, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Martin S. Maron, Tufts Medical Center, #70, 800 Washington Street, Boston, Massachusetts 02111
Over the past decade, cardiac magnetic resonance (CMR) has assumed an increasingly greater role in the evaluation of patients with hypertrophic cardiomyopathy (HCM) (1). By virtue of its high spatial resolution, CMR provides a precise morphological assessment of the diverse phenotypic expression of HCM (2). This has resulted in improved diagnosis by visualizing segmental left ventricular (LV) hypertrophy not well seen by echocardiography and identifying high-risk subgroups with scarred LV apical aneurysms, as well as characterizing structural abnormalities of the mitral valve and papillary muscle (1).
However, the capability of contrast-enhanced CMR to identify areas of abnormal myocardial substrate has perhaps generated the greatest interest in clinical practice. After intravenous injection of gadolinium, areas of hyperenhancement (i.e., late gadolinium enhancement [LGE]) within the myocardium can be identified and the amount quantified as a percentage of the total LV mass. Approximately half of HCM patients demonstrate LGE, with a diverse pattern and location, although most commonly involving hypertrophied segments of the LV wall (1).
This technique has now been applied to large populations of HCM patients to define the clinical significance of LGE. Initial cross-sectional studies confirmed that HCM patients with LGE were at greater risk of ventricular tachyarrhythmias on ambulatory monitoring compared with those without LGE (3). Subsequently, an international multicenter prospective study was recently completed in which almost 1,300 consecutive HCM patients were prospectively followed for >3 years after contrast-enhanced CMR. Extensive LGE occupying ≥15% of the LV myocardium was found to be an independent predictor of sudden death (4), with CMR providing the only opportunity to identify a subgroup of young asymptomatic HCM patients at increased risk of sudden death (by virtue of extensive LGE) and therefore potential candidates for implantable cardioverter-defibrillator therapy (4).
These findings underscore 2 important (and related) emerging principles. CMR with LGE can provide unique information that directly affects HCM patient management and that the amount of LGE is clinically relevant. This is in contrast to previous CMR studies on HCM that focused almost entirely on the association between a qualitative visual assessment of the presence of LGE and outcome (1,5), a strategy that is not practical for clinical decision making given how common LGE is in HCM. Therefore, if CMR is to be incorporated into the routine clinical evaluation of HCM patients for the purposes of risk stratification, a quantitative strategy for LGE will be required (1).
In this regard, numerous techniques have been used to assess LGE in HCM, including semiautomated algorithms that identify high signal intensity LGE pixels in the LV wall after applying a gray-scale threshold a number of SDs above the mean signal intensity within a remote region containing normal “nulled” myocardium (i.e., 2, 4, 5, or 6 SD) and the full width at half maximum (FWHM) method (pixels with a signal intensity >50% of the maximum intensity of the brightest region of hyperenhancement) (6). These quantification algorithms were initially validated in patients with coronary artery disease in whom lower signal intensity thresholds (i.e., 2 SD) correlate well with the spatial extent of infarct tissue (7). This is related to the fact that scar that results from a myocardial infarction is composed of focal, high signal intensity, homogeneous hyperenhancement surrounded by otherwise structurally normal myocardium.
However, the optimal approach to quantifying LGE in HCM is still not well defined (7). There are a number of reasons why this is the case, but perhaps the 3 most important issues are as follows: 1) as opposed to ischemic cardiomyopathy, HCM is characterized by diffuse histopathological abnormalities involving the entire LV myocardium, including replacement fibrosis or expanded extracellular matrix due to interstitial fibrosis and/or myocyte disarray (1). This results in a gradient of gadolinium deposition in the myocardium with focal areas of dense concentration producing high signal intensity hyperenhancement and diffuse regions of concentration resulting in lower signal intensity (7,8). 2) The limited spatial resolution of in vivo imaging can result in partial volume effects and, along with inappropriate nulling and background noise, can generate areas of increased signal intensity that appear visually as LGE but do not represent abnormal substrate (6,7). 3) Elucidating precisely the histological basis of LGE in HCM patients with normal LV function has not been possible due to the lack of a spontaneous HCM animal model (1).
As a result, it has not been possible to determine the optimal threshold technique that provides the closest representation of myocardial fibrosis or to clarify whether different gradients of LGE signal intensity correspond to different forms of fibrosis (interstitial vs. replacement). For this reason, substantial differences in the amount of LGE generated in an individual HCM patient can be present depending on which thresholding method is used (6–8), with no consensus with respect to the optimal method to use in HCM. However, higher gray-scale thresholds (6 SD) and FWHM have appeared to yield the closest approximation of the extent of LGE identified visually and the most reproducible (6,8).
In the article by Moravsky et al. (9), the authors provide the first systematic evaluation correlating histological findings with a variety of LGE quantification methods in patients with HCM and normal systolic function. The only other histopathological correlative studies on this disease have been 2 case reports from patients in the uncommon end-stage phase with systolic dysfunction (10). Myocardial tissue was obtained from 29 HCM patients at the time of surgical myectomy, and a histopathological analysis of these samples was undertaken to quantify areas of interstitial and replacement fibrosis. Portions of the contrast-enhanced CMR images at the myectomy site were analyzed using a variety of gray-scale thresholds and FWHM to determine which provided the optimal detection for fibrosis as derived from the histological samples.
A number of relevant observations were derived from these data including: 1) high gray-scale thresholds provide the best representation of total fibrosis burden compared with lower thresholds and the FWHM method; 2) fibrosis identified using high gray-scale thresholds is composed of interstitial and replacement fibrosis, although greater amounts of the former than latter; and 3) with increasing threshold levels (>5 and up to 10 SD), the specificity for identifying replacement fibrosis versus interstitial increases.
What are the implications of these data? First, these findings substantiate that high gray-scale thresholds are, in fact, identifying myocardial fibrosis and that 5 SD could be viewed as the most clinically relevant method for quantifying LGE in HCM. However, the actual difference in amount of LGE between 5 and 6 SD (or FWHM) is minimal with substantial overlap. Second, LGE quantified using lower gray-scale cutoffs appears to incorporate a larger proportion of voxels with increased signal intensity that are not representative of abnormal histopathology, but rather noise due to image artifact or suboptimal T1 inversion times, suggesting that these methods are not optimal for quantifying LGE in HCM (9).
One limitation of this study was the authors did not incorporate a manual method in which a gray-scale threshold is adjusted to define areas of visually identified LGE. This method was used to assess the extent of LGE in the aforementioned prospective multicenter LGE study (4), as it correlates well with the amount LGE quantified using high gray-scale threshold cutoffs, has excellent reproducibility, and is less time-consuming to perform, potentially making this method more attractive to incorporate into routine clinical practice.
Likewise, Moravsky et al. (9) also highlight that the current contrast-enhanced CMR techniques cannot reliably differentiate between interstitial and replacement fibrosis. However, even though the authors demonstrate that the specificity for detecting replacement fibrosis in HCM is greater with increasing gray-scale threshold cutoff values, even at the very high cutoff value of 10 SD, both types of myocardial fibrosis are present. Therefore, it is currently not possible to determine which type of fibrosis is more relevant with respect to generating ventricular tachyarrhythmias in HCM. Perhaps novel techniques such as T1 mapping will ultimately improve on these limitations of late gadolinium enhancement.
In summary, these data lend histopathological validation to the current clinical approach to quantifying LGE with high gray-scale thresholding methods in patients with HCM. As a result, this information brings us a step closer to establishing greater uniformity with respect to the most appropriate method to apply to HCM patients for the purpose of quantifying LGE in this disease. Ultimately, the final word will require prospective clinical trials in which a variety of quantification techniques can be directly compared to determine which represents the most robust method for identifying HCM patients at high risk.
Dr. Maron has reported that he has no relationships relevant to the contents of this paper to disclose.
↵⁎ Editorials published in JACC: Cardiovascular Imaging reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Imaging or the American College of Cardiology.
- American College of Cardiology Foundation
- Maron M.S.,
- Maron B.J.,
- Harrigan C.,
- et al.
- Adabag A.S.,
- Maron B.J.,
- Appelbaum E.,
- et al.
- Chan R.H.,
- Maron B.,
- Olivotto I.,
- et al.
- Bruder O.,
- Wagner A.,
- Jensen C.J.,
- et al.
- Kwong R.Y.,
- Farzaneh-Far A.
- Moravsky G.,
- Ofek E.,
- Rakowski H.,
- et al.
- Moon J.C.,
- Reed E.,
- Sheppard M.N.,
- et al.