Magnetic Resonance Imaging in Cardiac Amyloidosis: Unraveling the Stealth Entity

Abstract

Amyloidosis is a systemic disease involving many organs. Cardiac involvement is a significant cause of morbidity and mortality in these patients. Diagnosis of cardiac amyloidosis is based on endomyocardial biopsy which however is invasive and associated with complications. Noninvasive methods of diagnosis include magnetic resonance imaging (MRI) with various methods and sequences involved. Our study aims at describing MRI features of cardiac amyloidosis including new imaging sequences and to prognosticate the patients based on imaging features. We included 35 patients with suspected cardiac amyloidosis who underwent MRI at our center over 4 years. All images were retrieved from our archive and assessed by an experienced radiologist. Common morphological features in our patients included increased wall thickness of left ventricle (LV) (16. 1 ± 4.1 mm), right ventricle (RV) (6.3 ± 1.1 mm), and interatrial septum (6.2 ± 0.8 mm). Global late gadolinium enhancement (LGE) ( n  = 21 [65%]) including subendocardial or transmural was the most common pattern followed by patchy enhancement. Global transmural LGE was associated with worse prognosis. Four types of myocardial nulling patterns were observed on postcontrast time to invert (TI) scout imaging: normal nulling pattern (myocardium nulls after blood and coincident with spleen) and abnormal nulling pattern (ANP) which is further divided into three types: Type 1—myocardium nulls before blood pool but coincident with spleen, Type 2—myocardium nulling coincident with blood but not coincident with spleen, and Type 3—features of both Type 1 and Type 2. Type 3 ANP was the most common ( n  = 23) nulling pattern in our patients. Cardiac MRI is an essential in noninvasive diagnosis of cardiac amyloidosis. Transmural global LGE serves as a poor prognosticator in these patients. “Three-tier” TI scout imaging is essential to avoid false-negative enhancement results. Type 3 ANP is the most specific nulling pattern in cardiac amyloidosis.

Amyloidosis is a systemic disease characterized by deposition of proteinaceous fibrillar material in various tissues and organs of the body. There are several types of systemic amyloidosis; however, the most common ones affecting the heart include type amyloid light-chain (AL) amyloidosis (often associated with multiple myeloma or other monoclonal gammopathies) and amyloid transthyretin (ATTR) amyloidosis (including senile systemic amyloidosis), triggered by wild-type transthyretin (TTR) produced in the liver and variant ATTR amyloidosis, caused by point mutations in the TTR gene. ¹ Cardiac amyloidosis usually manifests itself as a restrictive cardiomyopathy. The basic pathology is the deposition of proteinaceous material throughout the myocardium that causes pressure atrophy of adjacent myocardial fibers. Deposition can also occur in the vessel walls causing coronary occlusion. These changes eventually result in ventricular and coronary wall remodeling with decreased compliance and eventual diastolic failure. ²

Although the pathology is quite different, there are other conditions that can mimic cardiac amyloidosis including other forms of restrictive cardiomyopathies, such as hypertrophic cardiomyopathy, sarcoidosis, or infiltrative lymphoma. The differentiation between these conditions is essential to guide treatment. Although endomyocardial biopsy can provide a definitive diagnosis, this is an invasive technique with its own complications. Also, the patchy involvement of the myocardium can give rise to a false-negative biopsy result. Therefore, a noninvasive method of diagnosing such conditions is essential. In the period preceding magnetic resonance imaging (MRI), cardiac involvement in amyloidosis was suspected based on electrocardiogram (ECG) abnormalities in a patient with known AL amyloidosis elsewhere, or increased wall thickness on echo and raised N-terminal pro b-type natriuretic peptide. However, these methods lacked sensitivity and specificity. MRI on the other hand can be game changer as far as cardiac involvement in amyloidosis is concerned. Restrictive cardiomyopathies as a group can have specific MRI features that can help differentiating them from hypertrophic variant. These include concentric thickening of the left ventricular wall, reduced systolic function with diminished ejection fraction, restriction of diastolic filling, and enlargement of atria without associated ventricular enlargement. A reduced ejection fraction seen in restrictive cardiomyopathies is in contradistinction to hypertrophic cardiomyopathies, which often are associated with a normal or even increased ejection fraction. ^{3 4 5 6 7} Although nuclear imaging (technetium 99m) has been used in the diagnosis of cardiac amyloidosis, its use has been found only in ATTR variant of the disease while the specificity and sensitivity in AL type is limited. However, it can also be used as an aid in the diagnosis of this stealth disease. ⁸

Several previous studies have attempted to describe specific features of cardiac amyloidosis on cardiac MRI including the decrease in signal intensity on T1- and T2-weighted imaging as well as specific pattern of postgadolinium enhancement patterns. In our part of the world with limited resources, such diagnosis may still be an enigma. Our study was aimed at elucidating the specific MRI features of cardiac amyloidosis in our set of population and assess their role in diagnosis and prognosis of cardiac amyloidosis.

Methods

Our study included 35 patients who were referred to us by the cardiology department as restrictive cardiomyopathy, diastolic dysfunction, and suspicion of amyloidosis based on ECG and echocardiography (ECHO) findings between 2016 and 2020. These patients were subsequently determined to have amyloidosis based on the presence of systemic amyloidosis confirmed by biopsy ( n  = 25) plus imaging features and a combined radiocardiac team discussion on the ECHO and MRI findings in the remainder ( n  = 10). The images were taken from the archive at the time of assessment making our study both a retrospective and partially prospective one. Cardiac biopsies were not performed in our patients. Patients with a history of myocardial infarction, myocarditis, known cardiomyopathies (dilated or hypertrophic), and those with contraindication for MRI contrast or claustrophobia were excluded.

Magnetic Resonance Imaging Technique

MRI was performed on a 1.5-T system (Magneton Avanto, Siemens medical system, Erlangen, Germany) with a four-element cardiac phased-array coil. We developed a specific protocol to be used in all cases of suspected amyloidosis. We started with Trufi multislice localizers to plan our imaging planes. This was followed by cine MR images using a balanced single shot free precession (b-SSFP) acquisition in the vertical long-axis, horizontal long-axis, and short-axis orientations with the following parameters: repetition time/echo time (TR/TE), 47.1/1.57; flip angle, 80 degrees; receiver bandwidth, ± 930 Hz/pixel; field of view, 34 cm; slice thickness, 5 to 6 mm; voxel size, 1.9 × 1.3 × 6 mm; and number of averages, 1 to 3. Retrospective ECG gating with k-space segmentation was used. However, in patients with arrhythmias or those unable to hold breath, prospective gating was employed.

Two-chamber time to invert (TI) scout images were taken to obtain the TI value for myocardial nulling taking care of not using the infiltrated myocardium as the reference. The following parameters were used: TR/TE, 23 to 49/1.12; flip angle, 30 degrees; receiver bandwidth, ± 965 Hz/pixel; field of view, 34 cm; slice thickness, 8 mm; voxel size, 3.5 × 1.8 × 8.0 mm; and number of averages, 1. We obtained the TI scout images before administration of contrast, at 5-minute postcontrast (TI ^5min ) and at 10-minute post contrast (TI ^10min ). The nulling of the myocardium in relation to the blood pool and the spleen was assessed. Normally, in postcontrast images, the blood pool nulls before the myocardium and myocardial nulling coincide with the spleen (normal nulling pattern [NNP]). The nulling of myocardium at a TI lower than that of blood pool or nulling not coincident with spleen was considered as abnormal nulling pattern (ANP). ANP was divided into three types: Type 1—myocardium nulls before blood pool but coincident with spleen, Type 2—myocardium nulling coincident with blood but not coincident with spleen, and Type 3—features of both Type 1 and Type 2 ( Figs. 1 2 3 ).

Fig. 1.

Postcontrast time to invert scout images with increasing inversion time from ( a ) to ( d ) showing Type 3 abnormal nulling pattern. Initially blood pool, myocardium, and spleen are all bright ( a ), myocardium nulls before blood ( b ), blood pool nulls next ( c ), and finally, it is the spleen ( d ).

Fig. 2.

Short axis time to invert (TI) scout images of the left ventricle showing Type 1 abnormal nulling pattern. Increasing TI images ( a – d ) show initial nulling of basal myocardium ( b ), followed by nulling of complete myocardium occurring before blood pool and coincident with spleen. Finally, nulling of blood pool is noted ( d ).

Fig. 3.

Postcontrast increasing time to invert scout images showing normal nulling pattern ( a – d ), Type 1 abnormal nulling pattern ( e – h ), and Type 3 abnormal nulling pattern ( i – l ).

First pass perfusion and late gadolinium enhancement (LGE) imaging were performed in the same slice locations using a segmented inversion recovery (phase sensitive inversion recovery [PSIR]) with the following parameters: TR/TE, 700/3.36; flip angle, 25 degrees; receiver bandwidth, ± 130 Hz/pixel; field of view, 34 cm; slice thickness, 8 mm; voxel size, 1.8 × 1.3 × 8.0 mm; and number of averages, 1. Images were acquired 5 to 25 minutes after administration of 0.2 mmol/kg of gadolinium (Omniscan), using an inversion time obtained on TI scout images taken at 5 and 10 minutes. As the time delay between contrast administration and imaging increased, the inversion time was changed according to the TI value obtained from TI scout images. All imaging sequences were acquired at end expiration.

Magnetic Resonance Imaging Analysis and Interpretation

Quantitative measures of left ventricular function were derived from the short-axis SSFP images using software (ARGUS) on a dedicated workstation. The standard technique of manually tracing the contours of the ventricular borders (epicardial and endocardial) was used to derive measures of left ventricular mass, end-diastolic volume (EDV), and end-systolic volume, from which stroke volume and ejection fraction were automatically calculated. The images were transferred to the workstation and were independently evaluated by two radiologists and after obtaining an opinion from a cardiologist, a final diagnosis was reached.

Patients with imaging features of amyloidosis on MRI underwent a profile of tests including testing for light chains and abdominal fat pad biopsy. The patients were followed up clinically over a period of 12 months.

Statistical Methods

The data were collected and evaluated using SPSS 21.0. Descriptive data were analyzed by frequencies and categorical data by percentages and continuous variables by means and standard deviations. Continuous variables were compared using Student's t -test. Correlation of MRI features with prognosis of patients was done by using Fisher's exact/Pearson's chi-square tests. For all comparisons, p -value of <0.05 was considered statistically significant.

Results

Patient Profile

We evaluated a total of 35 patients over a period of 4 years. The mean age of the patients in our study was 58.3 ± 5.6 years (range: 48–65 years). Our study included 24 (68.5%) males and 11 (31.5%) females.

Clinical Presentation

Heart failure and arrhythmias were the most common presenting features in our cases. Twenty-five (71%) of our patients presented with New York Heart Association (NYHA) grades 2 to 3 heart failure, while 16 (46%) had associated arrhythmias.

Biochemical/Pathological Evaluation

Biopsy of subcutaneous fatty tissue (BSFT) was performed in all patients and was found to be positive in 25 (71%) of our cases indicating systemic amyloidosis. Monoclonal light chain evaluation in serum and urine was also performed and showed presence of light chains (kappa and lambda) in 22 (63%) patients.

Echocardiography

All our patients had diastolic dysfunction on ECHO. Quantitative assessment of ventricular wall thickness was made. The mean left ventricle (LV) thickness in our patients was 14 ± 3.2 mm, while the mean right ventricle (RV) wall thickness was 5.8 ± 1.9 mm. Hyperechoic speckling appearance was seen in 30 (86%) of our patients. Pericardial effusion was noted in 25 0(71%) of our patients.

Magnetic Resonance Imaging

The various MRI parameters are shown in Table 1 . We also noted pericardial effusion in 21 (60%) of our patients. Pleural effusion was noted in 27 (77%) of our patients.

Table 1. Various quantitative MRI parameters seen in patients with cardiac amyloidosis.

Parameter	Mean value	Range
LV thickness (mm)	16.1 ± 4.1	11–20
RV thickness (mm)	6.3 ± 1.1	4–8
Interatrial septum thickness (mm)	6.2 ± 0.8	5.5–7
LVEF (%)	54 ± 7	49–61
LVEDV (mL)	123 ± 39	102–157
LVESV (mL)	58 ± 9	50–65

Abbreviations: LV, left ventricle; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; MRI, magnetic resonance imaging; RV, right ventricle.

Contrast-Enhanced Magnetic Resonance Imaging

LGE was seen in 32 (91%) of our patients. The most common pattern was global subendocardial or transmural seen in 21 (65%) ( Figs. 4 and 5 ) followed by focal patchy in 11 (35%). In majority of the patients with patchy enhancement (nine patients), there was an apicobasal gradient with predominant involvement of basal segments. The association of LGE with other imaging and clinical features are indicated in Table 2 .

Fig. 4.

Two-chamber ( a ) and four-chamber (b) diastolic balanced single shot free precession images showing mildly increased left ventricle wall thickness and interatrial septal thickness. Corresponding phase sensitive inversion recovery images (c and d) at the same levels shows global transmural enhancement in this 45-year-old patient with systemic amyloidosis.

Fig. 5.

Four-chamber balanced single shot free precession ( a ) image showing bilateral pleural effusion and mild pericardial effusion, increased interatrial septal and ventricular wall thicknesses are also noted. Four-chamber ( b ) and short axis ( c ) phase sensitive inversion recovery images show global predominantly subendocardial enhancement involving both left ventricle, right ventricle, and interatrial septum.

Table 2. Rrelationship of LGE with various other MRI features and clinical outcome of patients with cardiac amyloidosis.

MRI features	No LGE ( n  = 3)	Patchy LGE ( n  = 11)	Global LGE ( n  = 21)	p -Value
LV thickness (mm)	12 ± 1.5	13.6 ± 2.0	16.7 ± 3.2	<0.003
RV thickness (mm)	3.5 ± 0.8	4.8 ± 1.6	6.6 ± 1.7	<0.001
Interatrial septum thickness (mm)	2 0.5 ± 0.5	4.9 ± 2.0	6.6 ± 1.8	<0.005
LVEF (%)	60 ± 4	57 ± 8	50 ± 6	<0.001
Clinical course over 4 y
Progression to NYHA grades 3–4	None	4 patients (36%)	17 patients (81%)	<0.002
Death	None	1 patient (9%)	14 patients (66%)	<0.001

Abbreviations: LGE, late gadolinium enhancement; LV, left ventricle; LVEF, left ventricular ejection fraction; MRI, magnetic resonance imaging; NYHA, New York Heart Association; RV, right ventricle.

TI Scout Imaging (Look Locker)

The distribution of different nulling patterns observed in our patients on postcontrast TI ^5min and TI ^10min images are indicated in Table 3 .

Table 3. The pattern of nulling in the patients at TI ^5min and TI ^10min .

TI scout image	NNP	Type 1 ANP	Type 2 ANP	Type 3 ANP
TI 5min	3	7	2	23
TI 10min	0	10	2	23

Abbreviations: ANP, abnormal nulling pattern; NNP, normal nulling pattern, TI, time to invert.

Note: Type 3 ANP was the most common abnormal nulling pattern noted in our patients.

Follow-up

We found an overall poor prognosis in our set of patients with cardiac amyloidosis. Fifteen (42%) of our patients died during the course of the study, while 21 (60%) more progressed in their NYHA grading. Among the patients we lost, the most striking imaging feature was transmural LGE, severe diastolic dysfunction, and presence of pleuropericardial effusion seen in all of these patients at the time of presentation. We assessed the status of transmural LGE as a prognostic factor based on its effect on mortality by using the Fisher's exact/chi-square test which showed a significant relation ( p  < 0.001) ( Table 4 , Fig. 6 ).

Table 4. Statistically significant correlation of transmural LGE on MRI with mortality in patients with cardiac amyloidosis.

Chi-square tests
	Value	df	Asymp. Sig. (two-sided)	Exact Sig. (two-sided)	Exact Sig. (one-sided)
Pearson's chi-square	23.819 a	1	0.000
Continuity correction b	20.538	1	0.000
Likelihood ratio	27.390	1	0.000
Fisher's exact test				0.000	0.000
Linear-by-linear association	23.139	1	0.000
No. of valid cases b	35

Abbreviations: Asymp., asymptotic; LGE, late gadolinium enhancement; MRI, magnetic resonance imaging; Sig., significance.

Fig. 6.

Bar diagram showing the increased mortality in patients with transmural LGE on MRI in comparison to those with nontransmural enhancement or no enhancement. LGE, late gadolinium enhancement; MRI, magnetic resonance imaging.

Discussion

This study was conducted to ascertain the MRI features of cardiac amyloidosis and to look for any imaging features that can prognosticate these patients. We included a total of 35 patients in our study (mean age 58.4 years and a male-to-female ratio 24:11). Cardiac amyloidosis is more common in males and the mean age of presentation is in the fifth to sixth decades. These findings are consistent with previous studies conducted by Mahalingam et al ⁹ and vanden Driesen et al. ¹⁰

The most common clinical presentation in our study group was arrhythmias and unexplained heart failure. Many of our patients had pleuropericardial effusion at presentation and grouped into NYHA grades 2 to 3. BSFT was found to be positive in 25 (71%) of our cases (20 with AL type and 5 with TTR type) and monoclonal light chains (kappa and lambda) were present in 20 (57%) patients. Similar positivity of abdominal wall biopsies has been put forth by Ikeda et al ¹¹ and Paulsson Rokke et al. ¹²

ECHO has been used as a first-line modality for assessment of patients with explained heart failure. In patients with cardiac amyloidosis, we found increased wall thickness of both ventricles as the major finding in all the patients with mean LV thickness of 14 ± 3.2 mm and mean RV wall thickness of 5.8 ± 1.9 mm. The classical hyperechoic speckling was noticed in 80% of our patients. Some degree of valvular dysfunction in the form of mitral and tricuspid regurgitation was also noted in all the patients. These findings have been described by many previous studies including Cueto-Garcia et al, ¹³ Eriksson et al, ¹⁴ Hongo and Ikeda, ¹⁵ Cueto-Garcia et al, ¹⁶ and Siqueira-Filho et al. ¹⁷ These findings are however difficult to reproduce and good images can sometimes be difficult to obtain in patients who are obese and those with hyperinflated lungs. A recent advance in the field of ECHO in cardiac amyloidosis is speckle tracking for strain analysis and tissue Doppler imaging. Doppler imaging in amyloidosis shows impairment of longitudinal ventricular contraction before deterioration of the ejection fraction and onset of heart failure. This can be best demonstrated by strain imaging which shows severe impairment of longitudinal strain at the base of the LV, with relatively well-preserved apical strain. ¹⁸ Strain imaging however requires specific software which is not available on all ECHO machines. In our patients, speckle tracking was not performed due to the lack of appropriate software.

MRI in the recent years has emerged as the investigation of choice for noninvasive diagnosis of cardiac amyloidosis. The new sequences that have emerged including LGE on PSIR, T1 mapping, and TI scout with abnormal myocardial nulling pattern have further improved the diagnostic capabilities of MRI in diagnosing cardiac amyloid. Although not the gold standard for diagnosing cardiac amyloid, MRI can be an important supplement to clinical and laboratory findings especially when myocardial biopsy is not feasible. MRI however cannot be used in patients with claustrophobia, metallic foreign bodies or devices, renal failure, and has a limited availability in many parts of the world requiring expert acquisition.

In our study, we found that patients with cardiac amyloidosis have increased LV and RV free wall thickness with associated thickening of interatrial septum ( Table 1 ). An interatrial septal thickness of more than 6 mm has been considered as a good diagnostic clue and was found in 31 (88%) of our patients. The LV ejection fraction was in normal range in most of our patients at presentation; however, diastolic dysfunction was evident in the form of increased EDV ( Table 1 ). These structural quantitative measures have been previously described by many researchers including vanden Driesen et al, ¹⁰ Syed et al, ¹⁹ and aus dem Siepen et al. ²⁰ An increase in the ventricular wall thickness is not specific to amyloidosis as it can be found in other conditions such as hypertrophic cardiomyopathy, hypertension, aortic stenosis, etc. Interatrial thickness of more than 6 mm though is quite specific and should always raise the suspicion of amyloidosis.

LGE on contrast MRI not only helps in coronary disease for predicting viability of the myocardium but also helps in diagnosis and prognostication of various cardiomyopathies. In our study, we found LGE in most of the patients ( n  = 32) and the pattern was either patchy multifocal or global subendocardial or transmural. Only three (8%) of our cases had no LGE. We also correlated the type of LGE with imaging and clinical outcome of our patients. We found that patients with LGE had a worse prognosis compared with those without LGE. Among the LGE group, those with global transmural enhancement had the greatest abnormalities in their morphological as well as functional MRI features and an overall poor prognosis than patients with subendocardial or patchy enhancement ( Table 2 ). Thus, LGE in addition to helping in diagnosis of cardiac amyloid helps in prognostication of these patients. This is probably because LGE indicates cardiac amyloid load and the greater the LGE, the more amyloid deposit is expected within the myocardium. These presence and pattern of LGE in cardiac amyloid have been described by many researchers with results similar to ours, these include studies by Maceira et al, ²¹ Fontana et al, ²² and vanden Driesen et al. ¹⁰ LGE as a prognosticator in cardiac amyloid has also been suggested by authors including Syed et al ¹⁹ and Fontana et al ²² who in their studies found patients with transmural LGE to have the worse MRI characteristics and overall outcome. The absence of contrast enhancement in three of our patients may be due to either wrong selection of TI in diffusely infiltrated myocardium causing false-negative results or an early form of the disease with less threshold amyloid deposit. The former is less likely with PSIR images which are less dependent on operator selection of TI.

TI scout imaging has emerged as an important diagnostic tool in cardiac amyloid especially in areas where T1 mapping is not available yet. We performed TI scout imaging before and after contrast administration (5 and 10 minutes) helping us identify the nulling pattern as well as select an appropriate TI for PSIR imaging. NNP was observed in three patients at 5 minutes but in none at 10 minutes postcontrast PSIR images. Type 3 ANP was seen in most of the patients ( n  = 23 [66%]) followed by Type 1 ANP ( n  = 10 [28%]) and Type 3 ANP ( n  = 2 [6%]). These results are fairly consistent with a recent study of Ojha et al ²³ who found that Type 3 ANP was the most common and most specific nulling pattern in cardiac amyloid. Based on our observations, we recommend a three-tier TI scout imaging—before contrast, at 5 and 10 minutes after contrast administration. This protocol can help identify the temporal sequence of nulling in the patients. The reason for NNP at 5 minutes postcontrast images can probably be an altered hemodynamic or contrast kinetics as all these three patients showed Type 1 ANP at 10 minutes imaging. This temporal sequence of nulling was also observed by Mahalingam et al ⁹ in their study wherein temporal sequence of nulling was observed by the authors.

The limitations of our study were threefold. First is the number of patients which although comparable to other studies in our region are limited. Second, our center does not have the facilities of speckle tracking (ECHO) or T1 mapping which have now been added in cardiac amyloid protocol. Third, no endomyocardial biopsy was performed in our patients as such invasive procedures are not attempted at our center.

Conclusion

• Cardiac MRI is a one stop shop for assessment and prognostication of cardiac amyloid.

• Increased interatrial segment thickness is a good morphological indicator of cardiac amyloidosis.

• The need to perform multiple TI scout images to select the appropriate TI is critical to avoid false-negative results.

• LGE is a common feature of cardiac amyloidosis and transmural global LGE carries a worse prognosis.

• Three-tier TI scout imaging helps identify temporal nulling sequence in all patients. A Type 3 ANP is the most common and reliable indicator of amyloid deposits.

Funding Statement

Funding None.