Value of classical criteria for diagnosis of cardiac amyloidosis re-evaluated with magnetic resonance

Abstract

The noninvasive diagnosis of cardiac amyloidosis (CA) is often challenging. ECG abnormalities are common but lack both sensitivity and specificity. While some features on transthoracic echocardiography have confirmed the classical picture of CA, they have generally been validated against endomyocardial biopsy and thus are indicative of advanced disease. The ability of cardiac magnetic resonance to characterize the myocardium has opened a new door in the noninvasive diagnosis of CA, representing an opportunity to revisit the role of traditional ECG and transthoracic echocardiography diagnostic crtieria in earlier disease stages.

Keywords::

• cardiac amyloidosis
• cardiac magnetic resonance
• diagnosis
• echocardiography
• electrocardiography

Classical diagnostic criteria of cardiac amyloidosis

Although cardiac amyloidosis (CA) is an infrequent disease, it constitutes the first cause of restrictive cardiomyopathy in the western countries [1]. Moreover, the aging of the population and the improved survival in hematological dyscrasias may increase the incidence of senile and amyloid light-chain CA, respectively, in coming years. CA is an infiltrative cardiomyopathy characterized by the deposition of insoluble amyloid fibrils in the extracellular matrix of the myocardium. Therefore, the definitive diagnosis of CA is established by the demonstration in endomyocardial biopsy specimens of homogeneous proteinaceous material with typical apple-green birefringence of Congo red staining under polarized light. However, the invasive character and limited availability of this technique early fueled the search for noninvasive diagnostic tests to assess myocardial involvement. Typical features of CA on transthoracic echocardiography (TTE) were initially described in the late 1970s and early 1980s. Child et al. [2] reported for the first time the presence of symmetrically increased left ventricular wall thickness on M-mode TTE studies of patients with a diagnosis of CA, a finding confirmed in later studies [3]. Although classical for CA, this feature is little specific since it is frequently found in other, more prevalent diseases such as hypertensive heart disease, hypertrophic cardiomyopathy or aortic stenosis. Furthermore, the presence of increased left ventricular wall thickness has been associated with advanced disease [4]. With the development of 2D TTE, a characteristic ‘granular sparkling’ appearance of thickened myocardium was described as a hallmark of amyloid infiltration [5]. Although considered highly specific for CA, it demonstrated low sensitivity in different studies [5,6]. Moreover, it is dependent on technical aspects of image acquisition resulting in reduced specificity if tissue harmonic imaging is used or sensitivity may be attenuated by new image processing tools. Other echocardiographic features such as small biventricular cavities, low-normal ejection fractions, bilaterally enlarged atria, thickened valves and interatrial septum or pericardial and pleural effusions conform, together with increased wall thickness, the archetypal picture of CA. However, they are nonspecific, vary in prevalence and often constitute markers of end-stage disease [5]. The introduction of Doppler TTE allowed the noninvasive estimation of ventricular filling pressures and the detection of the typical restrictive pattern in CA. Nonetheless, it was soon noted that severe diastolic dysfunction is exclusive of advanced disease, whereas early amyloid infiltration may show an abnormal relaxation pattern with no differences in pulmonary flow compared with healthy controls [7]. In addition, it needs to be acknowledged that restrictive filling may be observed in a wide variety of cardiomyopathies.

Regarding ECG, low voltage, defined as QRS amplitude ≤0.5 mV in all limb leads or ≤1 mV in all precordial leads, and pseudo-infarct patterns were found in approximately 50% of the patients with biopsy-proven CA [8]. Other findings, such as atrioventricular block or atrial arrhythmias are also common, but far from specific for myocardial amyloid infiltration. More interestingly, it was initially demonstrated that the combination of low ECG voltage and increased myocardial mass not only improved the diagnostic accuracy for the detection of CA but also correlated with symptoms and mortality [9]. Subsequent studies confirmed that low ECG voltage increases specificity [6] for the detection of CA, but still lacks sensitivity [7] to detect early-stage disease.

Tissue characterization of amyloid infiltration with cardiac magnetic resonance

The increasing role of cardiac magnetic resonance (CMR) for CA evaluation has been mainly a consequence of its ability to perform tissue characterization through different sequences, especially post-contrast late gadolinium enhancement (LGE). In the case of CA, there is a typical LGE pattern characterized by diffuse, predominantly subendocardial distribution (Figure 1). This feature has shown an outstanding diagnostic performance in biopsy-proven CA, with particularly high specificity (94%) and variable sensitivity (69–97%) [10,11], the latter due to the possibility of other LGE patterns [11]. The extension of hyperenhancement has been correlated with left ventricular end-systolic and end-diastolic volumes, segmental dysfunction, left atrial size and lower voltage on ECG [11]. Moreover, diffuse LGE has been associated with decreased survival [12]. Recently, differences in LGE patterns have been described for the CA subtypes. Transthyretin-related CA demonstrated more extensive LGE and more frequent transmural extent as well as right ventricular involvement compared to amyloid light-chain CA [13]. Despite these findings and worse hypertrophy in transthyretin patients, outcomes were better, as previously described also with TTE [14], confirming that CA is a heterogeneous disease.

Figure 1. Typical features of CA with CMR. Characteristic concentric increase in left ventricular wall thickness can be noted in a four-chamber cine view (Panel A) together with pleural effusion (asterisk). Post-contrast LGE sequences (Panels B to D) demonstrate diffuse, predominantly subendocardial enhancement (arrowheads). Panel D also shows hyperenhancement of the left atrial wall (arrow), suggestive of amyloid infiltration of atrial myocardium.

The presence of abnormal contrast kinetics between blood and myocardium is another distinctive feature of CA. It is reflective of increased intramyocardial gadolinium concentration due to amyloid infiltration as well as faster contrast washout from the blood pool, probably because of diffuse increase in systemic interstitial space. This abnormality can be demonstrated by myocardial and blood T1 time quantification/estimation on specific sequences [10].

Classical diagnostic criteria in comparison with CMR

Our group has recently revisited the role of classical diagnostic criteria in 130 patients, including 51 in whom the diagnosis of CA was reached with endomyocardial biopsy and/or typical LGE [15]. Regarding the presence of symmetric left ventricular thickening, we found that up to 40% of the patients with CA showed remodeling patterns different from concentric hypertrophy, such as concentric remodeling or eccentric hypertrophy. Furthermore, asymmetric wall thickening predominantly involving the septum was noted in over two-thirds of the patients, particularly if overall mass was within normal limits. We believe that the discordance of our findings with the classical picture of CA can be due to earlier disease stages in our patients (selected from a tertiary center and with relatively high pretest probability) and/or more precise characterization by CMR, the gold standard for ventricular volumes and mass quantification, in comparison with TTE. Interestingly, recent pathology data have also described different patterns of left ventricular wall thickening in CA, including a segmental form with a preference for the septum and inferolateral wall [16]. In addition, the presence of low voltage on the ECG and classical TTE criteria (concentric left ventricular thickening, severe left atrial enlargement and advanced diastolic dysfunction) demonstrated high specificity but poor sensitivity, ranging from 12 to 59%.

Time to move forward in the noninvasive evaluation of CA

The findings reported above highlight the potential of CMR for earlier diagnosis of CA in comparison with traditional ECG and TTE criteria, as confirmed also by others [12]. Thus, over-reliance in such criteria is likely to result in delays for reaching the diagnosis. We advocate for a more routine use of CMR screening in patients with increased risk for CA, such as those with plasma cell dyscrasias or known transthyretin mutations, even in the absence of signs or symptoms of CA. Beyond LGE, the presence of unexplained septal thickening may increase the suspicion for myocardial amyloid infiltration. There is also growing data on the potential of myocardial T1 time quantification with the so-called T1-mapping techniques both before [17,18] and after contrast administration [10] for the estimation of amyloid burden.

Importantly, other modalities besides CMR can play an important role in CA detection and characterization. Myocardial strain imaging is not only useful to differentiate restrictive cardiomyopathy from pericardial constriction, but also to detect subclinical levels of systolic dysfunction that carry prognostic implications [19]. Nuclear techniques also have a role in the evaluation of CA. ^99mTc-3,3-diphosphono-1,2-propanodicarboxylic acid scintigraphy can detect the presence of CA and is useful in differentiating amyloid subtype [20]. In addition, novel tracers for positron emission tomography such as ¹⁸F-florbetapir have recently shown promising results for CA detection in pilot studies [21].

Conclusion

Classical ECG and TTE criteria are specific for CA detection, particularly if present in combination, yet lack sensitivity when compared with CMR. CMR enables accurate depiction of left ventricular remodeling and has recently challenged the notion that CA is invariably accompanied by concentric wall thickening. The presence of typical LGE on CMR or myocardial uptake of specific radioisotopes on scintigraphy or positron emission tomography can detect myocardial amyloid deposition with high sensitivity and specificity. Thus, when compared with traditional ECG and TTE, approaches that combine novel imaging techniques such as strain TTE, CMR or nuclear techniques offer the potential for earlier diagnosis and, perhaps, improved outcome in this deadly disease.

Financial & competing interest disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Notes