Author + information
- Frederick L. Ruberg, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Frederick L. Ruberg, Section of Cardiovascular Medicine, Boston Medical Center, C-8, 88 East Newton Street, Boston, Massachusetts 02118
The accurate identification of cardiac involvement in systemic light-chain (AL) amyloidosis is a critical aspect of clinical care. Mortality from cardiovascular causes is the most common cause of death in patients with AL amyloidosis, and those with cardiac involvement, seen in approximately 50% of patients with AL disease, are at the highest risk (1). Recognition of cardiac involvement informs clinical decisions regarding antiplasma cell therapy and heart failure management (2).
In the latter half of the 20th century, cardiac involvement could not be reliably demonstrated without invasive endomyocardial biopsy (3). Contemporary diagnostic algorithms that incorporate cardiac specific biomarkers, such as troponin or B-type natriuretic peptide, and modern echocardiographic techniques have essentially obviated the need for cardiac biopsy in the vast majority of cases (4). However, even with these sensitive noninvasive tests, there can remain situations of clinical uncertainty as to whether cardiac involvement is present.
Cardiac magnetic resonance (CMR), unlike the tests described in the preceding text, can identify cardiac amyloidosis because of its capacity to afford tissue characterization. Myocardial amyloid deposition results in interstitial space expansion, which, in turn, can be visualized by late gadolinium enhancement (LGE) imaging (5). Furthermore, light chains also result in myocyte death and a degree of reactive fibrosis that contribute to interstitial space expansion and LGE. Multiple studies have validated LGE CMR as a means to identify cardiac involvement in AL disease with a combined sensitivity that approaches 85% to 90% (5–9). However, LGE imaging has a number of limitations in amyloid disease. First, gadolinium contrast is required and cannot be administered to patients with severe renal impairment. Second, the patterns of LGE seen in amyloidosis are by no means uniform and are quite variable, leading to confusion in interpretation (7). Finally, LGE can be difficult to quantitate in cardiac amyloidosis wherein the intensity of signal is variable, its distribution patchy, and frequently is noncontiguous.
In this issue of iJACC, Karamitsos et al. (10) circumvent the previously described shortcomings of LGE CMR by using noncontrast, quantitative, T1 mapping to identify cardiac amyloid involvement in systemic AL disease (10). Studies of T1 mapping have expanded over the past few years, with data reported for a variety of nonischemic processes, including amyloidosis (11,12).
In the current study, the investigators recruited 53 patients with biopsy-proven, systemic AL amyloidosis and grouped them into 3 categories by echocardiographic wall thickness measures, diastolic functional assessment, and biomarker data: no cardiac involvement (26% of the cohort), possible involvement (21%), and definite involvement (53%). Comparison cohorts of healthy controls and patients with aortic stenosis matched by wall thickness to the amyloid patients were also recruited. Cine CMR followed by pre-contrast T1 mapping and LGE post-contrast was performed at 1.5T with a global, mean T1 value calculated.
As predicted, patients with definite involvement had greater myocardial mass and higher estimated filling pressures. Karamitsos et al. (10) found that T1 values were highest among patients with definite cardiac amyloidosis (mean 1,140 ± 61 ms) as compared to either healthy controls (958 ± 20 ms) or those with aortic stenosis (979 ± 51 ms). Next, when stratified by the noninvasive pre-CMR group assignment, there was a stepwise increase in T1. Using receiver-operating characteristic (ROC) analysis, a threshold noncontrast T1 value of 1,020 ms yielded 92% accuracy for identification of either possible or definite cardiac involvement. Correlations between T1 values and indices of morphology, systolic, and diastolic function were observed.
Because endomyocardial biopsies were only performed in 8% of the patients, we do not know which patients definitely had cardiac amyloidosis, but the LGE data help elucidate the pre-test group assignments. In the definite group, nearly all (96%) had either LGE or an abnormal myocardium to blood pool signal suppression pattern indicative of diffuse cardiac infiltration. Conversely, in the no amyloid pre-test group, none of the patients had LGE or an abnormal suppression pattern. However, in the possible cardiac amyloid group, approximately 50% had LGE or an abnormal suppression pattern. Karamitsos et al. (10) suggest that some of those patients without LGE may also have cardiac amyloidosis, even with normal biomarkers and wall thickness, by noting that T1 values above the ROC threshold were observed in 4 of 6 patients without LGE. While interesting to consider, this observation is unlikely to be of clinical importance given the ample data that now exist relating normal biomarkers to favorable clinical outcomes (13).
The ROC analysis allocated positive cases from both the definite or possible groups, which, therefore, likely included some patients in the possible group without cardiac involvement. Thus, the reported accuracy of 92% may be an over-estimate of who truly has cardiac involvement. It is interesting to note that this accuracy is not that dissimilar from that reported for LGE CMR.
More careful review of the T1 data as classified by subgrouping reveals minimal overlap between the definite and no cardiac involvement groups, despite the finding that mean–T1 among patients with no cardiac involvement was nonetheless increased compared to normal controls.
What appears most intriguing is the observed difference in T1 between patients with definite cardiac amyloidosis and aortic stenosis, despite similar left ventricle mass, and LGE noted in 32% of the aortic stenosis patients. Histologically, we can presume that the aortic stenosis patients had some degree of interstitial fibrosis that resulted in the LGE, yet the LGE images alone may not have been sufficient to differentiate the 2 entities. Here is where pre-contrast T1 mapping might be most helpful—to differentiate between wall thickness increase due to amyloidosis versus hypertrophy and interstitial fibrosis. Karamitsos et al. (10) astutely comment upon this, but also caution that as wall thickness and fibrosis increase, noncontrast T1 values may also increase. It is more challenging to make a distinction between interstitial space expansion from amyloidosis versus fibrosis in post-contrast studies. In a recent study of cardiac amyloidosis, post-contrast T1 mapping was utilized to determine the gadolinium partition coefficient and thereby derive a measurement of the extracellular volume fraction (14). Extracellular volume fraction was markedly increased in amyloidosis, and thus the magnitude of increase may be important.
How is this technique useful to the clinician? Its greatest merit is that it is a noncontrast, quantitative measurement that is relatively simply to perform and analyze, has minimal subjectivity, and is highly reproducible. In general, T1 mapping is sensitive to global changes in myocardial tissue, and, therefore, hypothetically may confer an advantage over LGE when applied to diffuse infiltrative processes such as amyloidosis. Performed serially, it may afford a more accurate means to follow response to treatment and changes in myocardial amyloid burden.
It is important to note that these data should be considered only to apply to patients with systemic AL disease, and not to those with transthyretin cardiac amyloidosis. Patients with wild-type transthyretin disease tend to have considerably thicker hearts (15) and would predictably have even higher noncontrast T1 values. Hypertrophic cardiomyopathy would be another important comparator, as its imaging characteristics also overlap with cardiac amyloidosis. Notably, a report of noncontrast T1 mapping in hypertrophic cardiomyopathy by these investigators is unfortunately not comparable to the present study as it was performed at 3.0-T (16).
Finally, there are a number of pulse sequences that can be used to perform T1 mapping with different reproducibility and accuracy. The technique employed here, the shortened modified look-locker inversion recovery (shMOLLI) sequence, is less subject to artifacts from respiratory motion, but currently has limited availability (17).
In summary, T1 mapping techniques such as that reported by Karamitsos et al. (10) have great promise to shape our approach to the management of patients with cardiac amyloidosis. Future studies of T1 mapping protocols utilizing larger cohorts, that include other amyloid precursor protein types, and that correlate imaging characteristics to clinical outcomes, is warranted. However, for any quantitative CMR technique to achieve widespread clinical application, there must be standardization and simplification of sequences and post-processing products.
Dr. Ruberg 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
- Dubrey S.W.,
- Cha K.,
- Anderson J.,
- et al.
- Falk R.H.
- Maceira A.M.,
- Joshi J.,
- Prasad S.K.,
- et al.
- Vogelsberg H.,
- Mahrholdt H.,
- Deluigi C.C.,
- et al.
- Syed I.S.,
- Glockner J.F.,
- Feng D.,
- et al.
- Austin B.A.,
- Tang W.H.,
- Rodriguez E.R.,
- et al.
- Karamitsos T.D.,
- Piechnik S.K.,
- Banypersad S.M.,
- et al.
- Rapezzi C.,
- Merlini G.,
- Quarta C.C.,
- et al.
- Dass S.,
- Suttie J.J.,
- Piechnik S.K.,
- et al.