Background
Systemic sclerosis (SSc) is an autoimmune connective tissue disorder characterised by vascular dysfunction and multi-organ fibrosis. The heart is one of the major organs commonly involved in SSc, with an estimated clinical prevalence of 15-35% [
1]. Cardiovascular disease (CVD) in SSc may be direct (cardiac fibrosis, myocarditis, dilated cardiomyopathy, cardiac failure, premature coronary artery disease, conduction system abnormalities, valvular disease and pericardial disease) or indirect (pulmonary hypertension and renal crisis) [
2,
3]. In the majority of SSc patients, however, CVD often remains subclinical [
4]. SSc patients with apparent cardiovascular clinical features are at greater risk of deterioration and premature cardiovascular death [
5]. Therefore, early detection and monitoring of myocardial and vascular involvement is a crucial aspect of management [
6].
Diffuse myocardial fibrosis is the pathological hallmark of cardiovascular involvement in SSc, reported in up to 80% of cases in autopsy studies [
7], and is thought to represent the final common expression of ‘contraction band necrosis’, recurrent episodes of ischaemia-reperfusion injury, microvascular dysfunction and chronic myocardial inflammation [
4,
8,
9]. However, myocardial inflammation and diffuse fibrosis are difficult to detect clinically, and endomyocardial biopsy is limited by sampling error, low diagnostic sensitivity and its invasive nature [
10].
Cardiovascular magnetic resonance (CMR) can non-invasively detect myocardial inflammation and fibrosis. T2-weighted imaging can identify areas of myocardial oedema, and late gadolinium enhancement (LGE) imaging can identify areas of focal fibrosis in patients with SSc [
11,
12]. However, conventional T2-weighted imaging only has modest sensitivity in detecting myocardial oedema [
13,
14], especially in mild cases, and LGE is suboptimal as a technique to detect diffuse myocardial fibrosis [
15,
16]. Recently, T1 mapping and extracellular volume (ECV) quantification have emerged as novel methods that can overcome both of these limitations and are promising to detect subtle forms of myocardial inflammation and diffuse myocardial fibrosis. T1 mapping is highly sensitive to myocardial water and is superior to T2-weighted imaging in detecting myocardial oedema [
14,
17]. T1 mapping and ECV measurements can also act as surrogates for diffuse fibrosis [
18,
19] on the premise of detecting myocardial water in the expanded interstitial space, and have been shown to correlate well with histological indices of myocardial fibrosis in various clinical contexts [
20,
21].
We therefore hypothesised that T1 and ECV quantification would reveal subclinical myocardial involvement in asymptomatic SSc patients with no known cardiovascular involvement when compared to controls of similar age and sex. We also sought to investigate the relationship between myocardial findings on CMR and indices of SSc activity, severity and chronicity, and early signs of myocardial dysfunction.
Discussion
Our data demonstrated that subclinical myocardial changes are common in patients with SSc even with apparently normal hearts, which can be detected using multiparametric CMR. In addition to focal areas of fibrosis (as detected by LGE), there were also areas of focal myocardial oedema or inflammation (as detected by T2-weighted imaging). Further, using more sensitive techniques such as native T1 mapping and ECV quantification, we were able to demonstrate even more areas of myocardial involvement in SSc patients than conventional CMR techniques can reveal, with SSc patients showing significantly larger areas of T1 abnormality and expanded ECV, which likely represent a combination of low grade inflammation and diffuse myocardial fibrosis that are well-described disease processes in this cohort. Interestingly, T1-mapping and ECV quantification were sensitive enough to further stratify myocardial involvement in patients with diffuse SSc compared to patients with limited cutaneous SSc, with the former showing significantly larger areas of myocardial T1 abnormality and ECV expansion. Finally, T1 and ECV measures were associated with subtle myocardial systolic and diastolic dysfunction. The results of this study suggest that CMR, particularly T1 and ECV quantification, can be used for early detection of subclinical myocardial involvement in SSc patients, potentially serving as an early screening tool before overt LV dysfunction or irreversible myocardial damage occurs.
In historic autopsy studies, diffuse myocardial fibrosis has been reported as the pathological hallmark of cardiovascular involvement in SSc [
7,
8]. Chronic myocardial inflammation together with recurrent ischaemia-reperfusion injury and microvascular dysfunction are thought to play a crucial pathophysiological role in the development of diffuse myocardial fibrosis in SSc [
4,
8,
9] and ultimately lead to premature cardiovascular mortality, particularly in those SSc patients with obvious clinical features of heart disease [
5].
Myocardial fibrosis is characterised by excessive deposition of extracellular matrix proteins, rich in collagen [
31], which ultimately affects myocardial structure and function, and is associated with impaired systolic and diastolic function, cardiac chamber dilation and arrhythmias [
32,
33]. LGE CMR has been traditionally used to image regional fibrosis/scarring in the myocardium, based on the distribution difference of gadolinium in between healthy and diseased myocardium [
15]. The LGE technique has significant limitations in the assessment of diffuse myocardial fibrosis, where the entire myocardium may be affected more homogeneously, as occurs with SSc, where there may be little unaffected myocardium.
Native (precontrast) T1 and ECV mapping are novel CMR markers that provide a quantitative measure of tissue characterisation without relying on relative signal comparisons [
15]. T1 maps are quantitative, pixel-wise representations of the T1 relaxation time of the underlying myocardial tissue [
16]. Increase in native T1 values is non-specific and can be seen in acute myocardial oedema, infarction, myocarditis, amyloidosis and diffuse fibrosis [
14,
17,
34,
35]. Postcontrast T1 is dependent on renal function, body fat distribution, dose of contrast used, the time delay in measurement after contrast administration, but calculating the ECV may minimise these confounding effects by incorporating the pre- and post-contrast myocardial and blood T1, partition coefficient (λ), and adjusting for the haematocrit [
36]. In the absence of oedema or other causes of ECV expansion such as amyloid and prior myocardial infarction, ECV increases are considered the best non-invasive surrogate of diffuse fibrosis [
36]. ECV has been validated against histological collagen proportion [
37], and correlates with poor early cardiovascular outcomes [
38]. In this study, postcontrast T1 values were found to be poor measures of diffuse fibrosis compared to native T1 and ECV calculation. Due to the greater variability of this measurement, postcontrast T1 has not been considered as a preferred method of assessment of diffuse myocardial fibrosis in the consensus statement on Myocardial T1 mapping and extracellular volume quantification by the Society for Cardiovascular Magnetic Resonance (SCMR) and CMR Working Group of the European Society of Cardiology [
36].
We showed increased native T1 and expanded ECV in SSc patients without cardiovascular symptoms. It is difficult to separate how much of the increase in T1 was due to myocardial inflammation as opposed to diffuse myocardial fibrosis, as both would increase native T1 values (and hence also ECV). There was a degree of myocardial oedema as shown by results using T2-weighted imaging, but the areas of oedema detected are not directly subtractable from areas of abnormality detected by T1, since these are different techniques with different sensitivities and specificities for oedema [
14]. Native T1 is significantly more sensitive to myocardial water than conventional T2-weighted imaging, so areas of oedema detected by T1 are expected to be larger than those by T2-weighted imaging, even in just a single disease process. In this cohort, both myocardial inflammation and diffuse fibrosis likely co-exist, and thus, CMR findings (as for any other diagnostic imaging modality) must be interpreted within the clinical context. Trying to distinguish between myocardial inflammation and diffuse fibrosis based on imaging alone may be challenging; the chronicity and relapsing nature of this disease must be taken into account, which can result in active myocardial inflammation over existing diffuse fibrosis from a previous episode. This is an inherent limitation that lends a necessary uncertainty, but currently no non-invasive diagnostic test can achieve this goal and no other cardiac imaging modality can provide more information on myocardial tissue characteristics than multiparametric CMR at this time. In either case, this study shows that CMR can detect subclinical myocardial involvement in SSc patients whose hearts would appear otherwise normal using conventional measures; longitudinal studies following disease course and trials assessing response to treatment strategies may shed more light onto the clinical meaning of T1 and ECV abnormalities in this cohort.
Despite the absence of global functional impairment, we found impaired peak circumferential LV systolic strain in SSc patients, in keeping with previous reports [
39]. There was also evidence of diastolic dysfunction with impaired peak diastolic strain rate, elevated E/E’ and increased left atrial dimensions. Although there were no major clinical effects on the SSc patients in this study, the presence of low grade myocardial inflammation, T1 abnormalities and ECV expansion may not be benign [
16,
31‐
33] and may ultimately lead to focal or diffuse fibrosis. Certainly, an autopsy study has previously found focal and interstitial myocardial fibrosis in SSc patients who had sustained high prevalence of ventricular arrhythmias and conduction disturbances, intractable congestive cardiac failure and sudden cardiac death [
8].
T1 and ECV measures both correlated inversely with peak systolic strain and peak diastolic strain rate in SSc patients. Several reports (including assessment of focal myocardial fibrosis by LGE) have confirmed that myocardial fibrosis precedes strain abnormalities and that fibrosis is a significant contributor to the pathogenesis of myocardial relaxation abnormalities [
40,
41]. Notably, in hypertensive patients, diastolic function improves after treatment with aldosterone antagonists, likely reflecting an antifibrotic effect of these drugs [
42]. In diabetes, cardiac dysfunction, relating to loss of contractile reserve and abnormal myocardial stiffness, is proportional to the degree of extracellular matrix deposition [
43]. Our findings support the hypothesis that in SSc, adverse myocardial processes may lead to diffuse myocardial fibrosis and are associated with impairment in myocardial strain.
We have found that both ECV and native T1 are useful for the assessment of myocardial involvement in SSc. In patients who are unable to tolerate gadolinium, native T1 may be used as a surrogate biomarker for myocardial fibrosis [
15,
17], if other causes of increased T1 are unlikely. In this study, ECV and native T1 both correlated positively with indices of SSc activity and severity, indicating that CMR may be useful in assessing both myocardial disease severity and activity in SSc. Furthermore, ECV and native T1 are increased even in SSc patients without focal fibrosis on LGE imaging, suggesting that these novel CMR markers provide additional information on tissue characterisation beyond that achieved by LGE.
Limitations
Our study has several limitations. First, the number of SSc patients included in this study is small; nevertheless the control group was well chosen and large significant differences in the parameters measured have been observed. Second, native myocardial T1 values may increase with myocardial oedema, myocardial infarction, myocarditis or amyloidosis [
17,
34,
35] and are not specific for myocardial fibrosis as discussed. Third, T2 mapping was not performed in this cohort, which would have been an interesting comparison; however, T2 mapping at the present time seems to have a large inter-individual variability [
43‐
45], which may not have helped significantly in distinguishing between myocardial water from inflammation versus that in an expanded extracellular space. Fourth, besides T1 mapping there is no other serum or histological test performed to support the presence of diffuse myocardial fibrosis in our patients; in this study of early disease in asymptomatic patients, no myocardial biopsy for histological correlations could be justified. Finally, we used the VDAI as a measure of disease activity in SSc. However, several other scores do exist which could have been used, which all have their own limitations.
Acknowledgments
The authors acknowledge support from the National Institute for Health Research Oxford Biomedical Research Centre Programme. Prof. Stefan Neubauer also acknowledges support from the Oxford British Heart Foundation Centre for Research Excellence.
We gratefully acknowledge the role of rheumatologists and rheumatology specialist nurses in patient recruitment from the following hospitals: John Radcliffe Hospital, Oxford; Nuffield Orthopaedic Centre, Oxford; Great Western Hospital, Swindon; Royal Berkshire Hospital, Reading; and Stoke Mandeville Hospital, Aylesbury.
Funding sources
This study was funded by an investigator-led grant from GSK to Dr Theo Karamitsos.
Competing interests
Competing financial interests.
We have no financial conflicts of interests to declare.
Competing non-financial interests.
US patent pending 61/387,591: SKP, MDR. SYSTEMS AND METHODS FOR SHORTENED LOOK LOCKER INVERSION RECOVERY (Sh-MOLLI) CARDIAC GATED MAPPING OF T1. September 29, 2010. All rights sold exclusively to Siemens Medical.
US patent pending 61/689,067: SKP, MDR. COLOR MAP DESIGN METHOD FOR IMMEDIATE ASSESSMENT OF THE DEVIATION FROM ESTABLISHED NORMAL POPULATION STATISTICS AND ITS APPLICATION TO CARDIOVASCULAR T1 MAPPING IMAGES.
All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
Authors’ contributions
NABN, PMM, BPW, SN and TDK were involved in the study conception and design. NABN, JMF and TDK were involved in data acquisition. NABN, SKP, VMF, JMF, ABSR and TDK were involved in data analysis. NABN, SPK, VMF, SN and TDK were involved in data interpretation. All authors read and approved the final draft.