2017 Volume 81 Issue 7 Pages 1014-1021
Background: We explored the usefulness of myocardial strain analysis on cardiac magnetic resonance imaging (CMR) scans for the identification of cardiac amyloidosis.
Methods and Results: The 61 patients with systemic amyloidosis underwent 3.0-T CMR, including CMR tagging and late-gadolinium enhanced (LGE) imaging. The circumferential strain (CS) of LGE-positive and LGE-negative patients was measured on midventricular short-axis images and compared. Logistic regression modeling of CMR parameters was performed to detect patients with LGE-positive cardiac amyloidosis. Of the 61 patients with systemic amyloidosis 48 were LGE-positive and 13 were LGE-negative. The peak CS was significantly lower in the LGE-positive than in the LGE-negative patients (−9.5±2.3 vs. −13.3±1.4%, P<0.01). The variability in the peak CS time was significantly greater in the LGE-positive than in the LGE-negative patients (46.1±24.5 vs. 21.2±20.1 ms, P<0.01). The peak CS significantly correlated with clinical biomarkers. The sensitivity, specificity, and accuracy of the diagnostic model using CS parameters for the identification of LGE-positive amyloidosis were 93.8%, 76.9%, and 90.2%, respectively.
Conclusions: Myocardial strain analysis by CMR helped detect LGE-positive amyloidosis without the need for contrast medium. The peak CS and variability in the peak CS time may correlate with the severity of cardiac amyloid deposition and may be more sensitive than LGE imaging for the detection of early cardiac disease in patients with amyloidosis.
Systemic amyloidosis consists of a group of uncommon diseases characterized by extracellular deposition of fibrillar proteins that results in loss of the normal tissue architecture and function.1 Cardiac amyloidosis is attributable to intramyocardial amyloid infiltration, which leads to progressive increases in ventricular wall thickness and stiffness.2 The 3 main types of systemic amyloidosis associated with clinically relevant cardiac involvement are amyloid light-chain (AL) amyloidosis attributable to clonal plasma cell dyscrasia that produces the immunoglobulin light-chain precursors of fibrillary deposits; the hereditary, transthyretin-related mutant form (ATTRm) that can be elicited by >100 mutations of transthyretin (TTR), a transport protein mainly synthesized by the liver; and the wild-type (nonmutant) TTR-related amyloidosis (ATTRwt) that primarily affects the heart in elderly men.1,3 Progressive amyloid deposition can induce restrictive cardiomyopathy, episodes of arrhythmia, and severe conduction disorders, including atrioventricular block with faintness, syncope, or result in sudden death.1,4
Cardiac magnetic resonance imaging (CMR) using late-gadolinium enhancement (LGE) assists the diagnosis of cardiac involvement in patients with systemic amyloidosis.5 The assessment by LGE is important because LGE is strongly associated with clinical, morphological, functional and biochemical markers that help to evaluate the prognosis of patients with cardiac amyloidosis.6 However, in patients with suspected cardiac amyloidosis and whose renal function may be significantly impaired, the use of a gadolinium-based contrast medium is inadvisable. Furthermore, LGE is thought to manifest in the later stages of the disease and the LGE pattern may be atypical and patchy, even in patients with life-threatening disease.2,6 Therefore, a non-contrast CMR technique that accurately identifies cardiac amyloidosis and, ideally, facilitates the quantitative assessment of the myocardial amyloid load would be of great value. CMR tagging has been established as an essential technique for measuring regional myocardial function.7 It facilitates quantification of local intramyocardial motion and strain.
Myocardial strain analysis by 2-dimensional (2D) speckle-tracking echocardiography can effectively detect subtle systolic function impairment in a variety of diseases and is useful for assessing the prognosis of patients with cardiac amyloidosis.8 However, the value of myocardial strain analysis by CMR in patients with cardiac amyloidosis remains to be elucidated and was the aim of this study.
This retrospective study was approved by the institutional review board; patient informed consent was waived. Between July 2011 and May 2016, 61 consecutive patients with systemic amyloidosis (39 men and 22 women, aged 58.5±18.5 years, range, 26–86 years) and no contraindications for CMR were enrolled in this study; 41 were affected by ATTRm, 10 by ATTRwt, and 10 by AL amyloidosis. Systemic amyloidosis was diagnosed based on clinical findings, amyloid deposition, and genetic testing; it was confirmed histologically by Congo red and immunohistochemical staining of bone marrow, soft tissue, fat, upper gastrointestinal tract, rectal, and liver specimens. Our inclusion criteria included an estimated glomerular filtration rate (eGFR) >30 mL/min/1.73 m2. The patient characteristics are summarized in Table 1.
LGE-negative (n=13) |
LGE-positive (n=48) |
P value | |
---|---|---|---|
Sex (M/F) | 7/6 | 32/16 | 0.39 |
Age (years) | 43.1±10.8 | 62.7±14.0 | <0.01 |
Cardiac symptoms (no/yes) | 12/1 | 17/31 | <0.01 |
BSA (m2) | 1.65±0.17 | 1.60±0.14 | 0.36 |
eGFR (mL/min/1.73 m2) | 77.7±12.8 | 66.8±15.5 | 0.02 |
BNP (pg/mL) | 34.3±49.4 | 209.4±275.1 | <0.01 |
CMR findings | |||
IVS wall thickness (mm) | 11.3±1.9 | 16.9±4.0 | <0.01 |
LVM (g/m2) | 52.4±13.3 | 79.3±27.1 | <0.01 |
EDV (mL) | 67.1±16.3 | 73.7±23.3 | 0.34 |
ESV (mL) | 21.1±9.4 | 29.6±15.8 | 0.25 |
LVEF (%) | 68.8±8.3 | 61.1±14.1 | 0.07 |
Echocardiography findings | |||
E/e’ ratio | 9.7±3.9 | 17.9±6.3 | <0.01 |
E-deceleration time (ms) | 209.3±59.7 | 199.0±78.1 | 0.61 |
Data are mean±standard deviation. BNP, plasma B-type natriuretic peptide; BSA, body surface area; CMR, cardiac magnetic resonance; e’, tissue Doppler-derived early diastolic peak velocity at the lateral mitral annulus; E, early diastolic transmitral flow velocity; EDV, end-diastolic volume; ESV, endsystolic volume; eGFR, estimated glomerular filtration rate; IVS, interventricular septum; LGE, late-gadolinium enhancement; LVEF, left ventricular ejection fraction; LVM, left ventricular mass.
All patients underwent standard echocardiographic studies using a commercial ultrasound machine (Vivid 7; GE Healthcare, Milwaukee, WI, USA) equipped with a 3.5-MHz transducer. The examinations were performed by experienced investigators. Morphological and functional evaluations of the cardiac chambers were according to the recommendations of the American Society of Echocardiography.9 The pulsed Doppler sample volume was positioned at the opened leaflet tips of the mitral valve. The early and late diastolic peak flow velocities and the E-wave deceleration time were measured by transmitral Doppler imaging. The ratio of the transmitral Doppler early filling velocity (E/e’) was also recorded.
CMR ProtocolAll CMR studies were performed on a 3.0-T MR scanner (Achieva 3.0T X-series TX; Philips Medical Systems, Best, The Netherlands). The patients were scanned in the supine position using a dedicated 32-channel cardiac torso coil. Images were acquired using T2-weighted black blood with STIR; cine imaging obtained with a segmented steady-state free-precession sequence; CMR tagging and LGE imaging was performed. LGE imaging was in the short-axis and 3 long-axis views (4-, 2- and 3-chamber views) with a mid-diastolic inversion-prepared 2D gradient echo (2D-IR) and a 3D phase-sensitive inversion recovery (3D-PSIR) sequence. It was started 8 min after the injection of 0.2 mmol/kg of a gadolinium-based contrast medium (Magnevist or Gadovist; Bayer Yakuhin, Osaka, Japan).
CMR tagging was performed in the short-axis image plane of the left ventricle using a complementary spatial modulation of magnetization (cSPAMM) method;10 this creates a tag grid on the images using a 2D turbo field echo sequence with ECG-triggered breath holding. Typical parameters were: repetition time 4.3 ms, echo time 2.7 ms, flip angle 10°, shot duration 39.2 ms, bandwidth 775 Hz, section thickness 8.0 mm, field of view 380 mm, matrix size 256×204, SENSE factor 2, tag grid 8.0 mm, 14–22 cardiac phases/R-R interval on the ECG. The midventricular level of the CMR-tagging image was used to analyze myocardial strain.
LGE InterpretationTwo blinded board-certified cardiovascular radiologists with 17 and 10 years of CMR experience, respectively, interpreted and manually traced the LGE lesions and myocardium on the 2D-IR and 3D-PSIR images. The calculated LGE volume was expressed as the %LGE of the myocardial volume of the left ventricle [%LGE=(LGE volume/myocardial volume)×100]. The average %LGE value recorded by the 2 observers was adopted as the final %LGE. The LGE pattern was classified into 3 types according to the degree of transmurality: no LGE, subendocardial LGE (global subendocardial involvement but no transmural LGE), and transmural LGE (LGE with transmural extension). Patients with basal transmural LGE but apical subendocardial LGE were classified as transmural LGE.
Myocardial Strain AnalysisTo measure circumferential strain (CS), CMR-tagging images were analyzed on a post-processing workstation (Ziostation 2, Ziosoft, Tokyo, Japan). We manually contoured the endocardium and epicardium, then the CS of each segment (anterior, septal, lateral, and inferior) was automatically calculated and strain curves and a superimposed color parametric map in the ventricular short-axis plane were generated. The peak CS (%) and peak CS time (ms) of each segment were recorded for each patient. The LGE images were visually inspected for the presence or absence of LGE by a reader blinded to the myocardial strain results. The myocardial strain results of amyloidosis patients with and without LGE were compared. We also evaluated the relationship between the clinical parameters and the myocardial strain results.
Statistical AnalysisAll numeric values are reported as the mean±standard deviation. Differences in the mean values between the 2 groups with normally and non-normally distributed data were determined with the 2-tailed independent t-test and the Mann-Whitney U-test, respectively. The χ2 test was used to compare the 2 groups. Correlations were assessed by Pearson correlation or Spearman coefficient. We built 2 types of logistic regression models using backward elimination based on clinical parameters [patient age, plasma B-type natriuretic peptide (BNP) level, left ventricular ejection fraction (LVEF), E/e’, and the E-deceleration time] and CMR parameters [wall thickness of the interventricular septum (IVS), LV mass (LVM), end-diastolic and endsystolic volumes, and strain parameters] to detect LGE-positive cardiac amyloidosis patients. We then developed a combined logistic regression model including all variables associated with significant P values in the previous models. Receiver-operating characteristic analysis was performed to determine the diagnostic cutoff values of the CMR strain parameters for the identification of LGE-positive amyloidosis patients. Because of the limited number of patients with this rare disease in our study, instead of cross-validation for which a second dataset is needed to evaluate the generalizability of the model, we assessed the predictive performance by calculating the sensitivity, specificity, and accuracy using patient data from training data. Differences of P<0.05 were considered statistically significant. We used JMP software for statistical analysis (ver. 9.0.2; SAS Institute, Cary, NC, USA).
The baseline characteristics of our patients are shown in Table 1. Of the 61 patients with systemic amyloidosis 48 were LGE-positive and 13 were LGE-negative. On CMR scans the IVS wall thickness and LVM were greater in the LGE-positive than in the LGE-negative patients (IVS wall thickness: 16.9±4.0 vs. 11.3±1.9 mm, P<0.01; LVM: 79.3±27.1 vs. 52.4±13.3 g/m2, P<0.01). There was no significant difference in the end-diastolic and endsystolic volumes, or LVEF. The LGE-positive patients were significantly older, their BNP level was significantly higher, and the E/e’ ratio on echocardiographs was also significantly higher (all, P<0.01) than in the LGE-negative patients. The peak CS was significantly lower in the LGE-positive than in the LGE-negative patients (−9.5±2.3 vs. −13.3±1.4%, P<0.01) (Table 2). Although the mean peak CS time was longer in the LGE-positive than in the LGE-negative patients (393.8±80.3 vs. 361.2±50.5 ms, P= 0.08), the difference was not significant. Variability in the peak CS time was significantly greater in the LGE-positive than in the LGE-negative patients (46.1±24.5 vs. 21.2±20.1 ms, P<0.01) (Table 3). As shown in Table 4, peak CS significantly correlated with the IVS wall thickness (r=0.58, P<0.01), LVM (r=0.63, P<0.01), LVEF (r=−0.59, P<0.01), %LGE (r=0.72, P<0.01), BNP level (r=0.60, P<0.01), eGFR (r=−0.39, P<0.01), and the E/e’ ratio (r=0.51, P<0.01). Variability in the peak CS time also significantly correlated with the IVS wall thickness (r=0.58, P<0.01), %LGE (r=0.36, P<0.03), and the E/e’ ratio (r=0.51, P<0.01). Significant CMR parameters in the logistic regression model associated with LGE-positive amyloidosis were the peak CS [odds ratio (OR) 3.57, 95% confidence interval (CI) 1.57–8.10, P<0.01) and variability in the peak CS time (OR 1.48, 95% CI 1.10–1.98, P<0.01) (Table 5). The sensitivity, specificity, and accuracy of the diagnostic models to identify LGE-positive amyloidosis patients were 93.8%, 69.2%, and 88.5%, respectively, for the clinical model; 93.8%, 76.9%, and 90.2%, respectively, for the CMR model, and 93.8%, 84.6%, and 91.8%, respectively, for the combined model. The cutoff values for the mean peak CS and the variability in time to peak CS for identifying LGE-positive amyloidosis patients were −11.9% and 33.5 ms. Representative cases are shown in Figures 1–3.
LGE-negative (n=13) |
LGE-positive (n=48) |
P value | |
---|---|---|---|
Anterior segment (%) | −12.3±1.8 | −9.0±2.4 | <0.01 |
Septal segment (%) | −12.0±1.4 | −7.2±2.7 | <0.01 |
Lateral segment (%) | −14.8±2.9 | −11.4±3.0 | <0.01 |
Inferior segment (%) | −14.1±1.8 | −10.5±2.9 | <0.01 |
All segments (%) | −13.3±1.4 | −9.5±2.3 | <0.01 |
Data are mean±standard deviation. CS, circumferential strain; LGE, late-gadolinium enhancement.
LGE-negative (n=13) |
LGE-positive (n=48) |
P value | |
---|---|---|---|
Anterior segment (ms) | 359.8±58.3 | 389.4±100.9 | 0.18 |
Septal segment (ms) | 350.7±51.4 | 396.7±95.2 | 0.02 |
Lateral segment (ms) | 373.5±52.5 | 391.4±83.6 | 0.47 |
Inferior segment (ms) | 361.5±61.7 | 398.3±88.4 | 0.10 |
All segments (ms) | 361.2±50.5 | 393.8±80.3 | 0.08 |
Δpeak CS time (ms) | 46.0±45.2 | 96.7±53.7 | <0.01 |
Variability in peak CS time (ms) | 21.2±20.1 | 46.1±24.5 | <0.01 |
Cardiac cycle (ms) | 894.8±157.0 | 916.2±184.4 | 0.70 |
Data are mean±standard deviation. Abbreviations as in Table 2.
Peak CS | Variability in peak CS time | |||
---|---|---|---|---|
r value | P value | r value | P value | |
BNP | 0.60 | <0.01 | 0.26 | 0.04 |
eGFR | −0.39 | <0.01 | −0.20 | 0.11 |
CMR findings | ||||
IVS wall thickness | 0.58 | <0.01 | 0.58 | <0.01 |
LVM | 0.63 | <0.01 | 0.28 | 0.03 |
EDV | 0.15 | 0.26 | 0.12 | 0.35 |
ESV | 0.50 | <0.01 | 0.34 | 0.02 |
LVEF | −0.59 | <0.01 | −0.20 | 0.13 |
%LGE | 0.72 | <0.01 | 0.36 | 0.03 |
Echocardiography findings | ||||
E/e’ ratio | 0.51 | <0.01 | 0.51 | <0.01 |
E-deceleration time | −0.26 | 0.04 | −0.27 | 0.04 |
Data are mean±standard deviation. A, late diastolic transmitral flow velocity. Other abbreviations as in Tables 1,2.
Explanatory variable | OR | 95% CI | P value |
---|---|---|---|
Clinical model | |||
Age | 1.08 | 1.01–1.16 | 0.02 |
E/e’ ratio | 1.39 | 1.08–1.79 | 0.01 |
CMR model | |||
Mean peak CS (per %) | 3.57 | 1.57–8.10 | <0.01 |
Variability in peak CS time (per ms) | 1.48 | 1.10–1.98 | <0.01 |
Combined model | |||
E/e’ ratio | 1.30 | 0.98–1.74 | 0.04 |
Mean peak CS (per %) | 3.81 | 1.41–10.26 | <0.01 |
Variability in peak CS time (per ms) | 1.32 | 0.97–1.80 | 0.04 |
CI, confidence interval; OR, odds ratio. Other abbreviations as in Tables 1,2,4.
A 31-year-old man with familial amyloid polyneuropathy. (A) Left ventricular short-axis late-gadolinium enhancement (LGE) magnetic resonance imaging (MRI) scan. (B) Superimposed color-coded strain map (ventricular short-axis plane). (C) Circumferential strain (CS) time-curves for the anterior (purple), septal (light blue), lateral (red), and inferior segments (light green). The interventricular septal wall thickness was normal (10.6 mm); the ejection fraction was 71.8%. There was no LGE on the cardiac magnetic resonance imaging scans. The plasma B-type natriuretic peptide level (5.0 pg/mL) and the E/e’ ratio (6.4) on echocardiographs were also normal. Strain analysis showed a normal pattern; the mean peak CS and the variability in the peak CS time were −14.9% and 22.0 ms, respectively.
A 55-year-old woman with familial amyloid polyneuropathy. (A) Left ventricular short-axis LGE MRI scan. (B) Superimposed color-coded strain map (ventricular short-axis plane). (C) CS time-curves for the anterior (purple), septal (light blue), lateral (red), and inferior segments (light green). The thickness of the interventricular septal wall was increased (17.7 mm). The ejection fraction was normal (51.2%). LGE MRI (A) revealed global subendocardial LGE and a dark blood pool, typical of cardiac amyloidosis. Both the plasma B-type natriuretic peptide level (725.3 pg/ml) and the E/e’ ratio (23.3) on echocardiographs were elevated. Strain analysis showed reduced negative peak CS in all segments. The peak CS time was delayed and highly variable. Mean peak CS and the variability in the peak CS time were −6.1% and 55.6 ms, respectively. Abbreviations as in Figure 1.
A 36-year-old man with familial amyloid polyneuropathy. (A) Left ventricular short-axis LGE MRI scan. (B) Superimposed color-coded strain map (ventricular short-axis plane). (C) CS time-curves for the anterior (purple), septal (light blue), lateral (red), and inferior segments (light green). The thickness of the interventricular septal wall was slightly increased (14.2 mm); the ejection fraction was normal (60.7%), and there was no LGE on cardiac magnetic resonance imaging scans. The plasma B-type natriuretic peptide level was slightly elevated (29.9 pg/ml) and the E/e’ ratio (8.8) was at the lower limit of normal on echocardiographs. Strain analysis showed reduced negative peak CS strain especially in the anterior (purple), septal (light blue), and lateral (red) segments. The peak CS time was slightly delayed in the anterior (purple), septal (light blue), and lateral (red) segments. The mean peak CS and the variability in the peak CS time were −10.4% and 25.1 ms, respectively. These observations suggest that strain analysis may be more sensitive than LGE imaging for the detection of early cardiac amyloidosis. Abbreviations as in Figure 1.
Cardiac amyloidosis is a manifestation of one of several systemic amyloidoses. In patients with this disease, multiple organs and tissues show extracellular deposition of pathologic, insoluble fibrillar proteins. Cardiac involvement may be the predominant feature; its relative predominance varies with the type of amyloidosis. The main forms of amyloidosis involving the heart are AL, ATTRm and ATTRwt amyloidosis.
AL amyloidosis is attributable to the immunoglobulin light-chain usually associated with plasma cell dyscrasia. It affects the heart, kidney, liver, peripheral/autonomic nerves, soft tissue, and the gastrointestinal system. One-third to one-half of AL amyloidosis patients present with heart disease; their heart failure tends to progress rapidly once the heart is affected and their prognosis is very poor. AL amyloidosis patients without cardiac involvement have a median survival of around 4 years. The prognosis among affected patients with markedly elevated BNP and cardiac troponin levels is approximately 8 months.11 The major treatment for this disease is chemotherapy.2,3,12 ATTRm, also known as “familial type of amyloid polyneuropathy (FAP)”, is one form of hereditary systemic amyloidosis. It is caused by the production of mutant TTR in the liver and mainly affects the peripheral/autonomic nerves and the heart. Although FAP was thought to be restricted to an endemic presence in specific areas, especially Portugal, Sweden, and Japan, progress in biochemical and molecular genetic analyses has shown that it is present throughout the world and more than 130 mutations in the TTR gene have been reported.4 The treatments are liver transplantation and new drugs (diflunisal and tafamidis) to stabilize TTR,13 which are usually effective. The median survival exceeds 20 years;14 the reported 5-year survival rate is 92%.15 Cardiac involvement, observed in approximately 80% of FAP patients, affects their prognosis negatively. Another type of ATTR amyloidosis is ATTRwt, also known as senile systemic amyloidosis. The main involved organ is the heart. ATTRwt is almost exclusively found in elderly men (male to female ratio=20:1–50:1); its symptoms are slowly progressive.16 Amyloid deposits are found at autopsy in approximately 25% of individuals older than 80 years; their clinical significance remains unclear.16 The natural history of ATTRwt remains poorly understood, and a median survival of approximately 7 years from presentation has been suggested.17 ATTRwt in patients older than 60 years is often misdiagnosed as hypertensive heart disease.18 The true incidence of ATTRwt is probably underestimated, but recent advances in CMR have greatly improved the detection of cardiac amyloid during life, which suggests that ATTRwt is more common than previously thought.
We documented that the myocardial peak CS was significantly lower, and the variability in the peak CS time significantly greater, in LGE-positive than in LGE-negative patients. These CMR strain parameters had high diagnostic accuracy for the detection of LGE-positive amyloidosis. They correlated well with the clinical markers of systolic and diastolic dysfunction (e.g., IVS wall thickness, LGE volume, BNP level, and E/e’ ratio). In addition, they may be more sensitive than LGE imaging for the detection of early disease, and decreased myocardial strain detected on CMR scans may be a direct marker of cardiac amyloid load (Figure 3).
The diagnosis of cardiac amyloidosis can be difficult because its clinical, electrocardiographic, and echocardiographic signs can be related to a number of different conditions. In advanced but not early disease there are characteristic echocardiographic findings. Typically, there is concentric ventricular thickening with right ventricular involvement, poor biventricular long-axis function with a normal/near normal EF, and valvular thickening.19 Diastolic dysfunction is the earliest echocardiographic abnormality and may precede cardiac symptoms.20 The LV wall becomes thicker, the E/e’ ratio higher, and the E-deceleration time shorter as cardiac involvement progresses in patients with amyloidosis.21 In our series the myocardial peak CS and the variability in the peak CS time were significantly correlated with the thickness of the IVS wall (r=0.47 and 0.58, P<0.01), the E/e’ ratio (r=0.51 and 0.51, P<0.01), and the E-deceleration time (r=−0.26 and −0.27, P=0.04). A speckled or granular myocardial appearance, although characteristic of amyloid deposition, is an inexact finding that may reflect machine-gain settings. Bi-atrial dilation in the presence of biventricular, valvular, and interatrial septal thickening is a useful diagnostic clue.19 Cardiac biomarkers such as BNP N-terminal fragments (NT-proBNP) and cardiac troponins are highly informative in patients with cardiac amyloidosis. Abnormal NT-proBNP is predictive of future clinically significant cardiac involvement.22 In our patients there was a significant correlation between the myocardial peak CS and the BNP level (r=0.60, P<0.01). Endomyocardial biopsy is the gold standard for demonstrating cardiac amyloid deposition, but its invasive nature prevents its routine use. Instead, clinical and imaging findings supported by the presence of amyloid at extracardiac sites have been used to obtain a diagnosis. Also, sampling errors render quantification unreliable and may give rise to false-negative results. On the other hand, CMR yields additive information on cardiac amyloid deposits by characterization of myocardial tissue. The appearance of global, subendocardial LGE and an associated dark blood pool is characteristic and correlates with the prognosis.5,6 The LGE pattern can be atypical and patchy, especially in early disease23 and transmural-, especially TTR-related, amyloidosis.24 LGE imaging requires gadolinium-based contrast medium, which may be contraindicated because many patients present with renal failure.25 In addition, CMR is suboptimal in patients with amyloidosis because amyloid infiltration in the interstitium of the heart reduces the difference in signal intensity between the blood and the myocardium and the 2 compartments may null together or even be reversed.26 We found that myocardial strain analysis of CMR scans facilitated the detection of cardiac involvement without gadolinium-based contrast medium. Thus, it may be a clinically useful diagnostic test in patients with amyloidosis. As quantification of the peak CS and variability in the peak CS time helps to determine the cardiac amyloid burden it may help to evaluate the prognosis and to assess the therapeutic response of patients with cardiac amyloidosis.
Several studies have documented that myocardial strain assessment by 2D speckle-tracking echocardiography is effective for detecting subtle systolic function impairments in a variety of diseases and for evaluating the prognosis of patients with cardiac amyloidosis.8,20 To assess myocardial strain by echocardiography, longitudinal shortening using the apical windows has been used widely because the ultrasound beam angle is favorable.27 According to Phelan et al,28 on 2D speckle-tracking echocardiographs, cardiac amyloidosis is characterized by regional variations in the longitudinal strain from the base to the apex called “apical sparing” or the “basal-to-apical gradient”. This pattern of longitudinal strain is an easily recognizable, accurate and reproducible method of differentiating cardiac amyloidosis from other causes of LV hypertrophy. Ternacle et al8 observed basal-to-apical longitudinal strain abnormalities in approximately half of their patients with cardiac amyloidosis. Longitudinal strain correlated with the LGE area and the amyloid burden, and apical longitudinal strain was an independent predictor of major adverse cardiac events. Huang et al observed systolic and diastolic dysfunction in patients with systemic amyloidosis whose EF was preserved; they detected early impairment in LV function by myocardial strain analysis using echocardiography.29 Although their findings are concordant with ours, to our knowledge ours is the first study of the ability of myocardial strain analysis by CMR to detect and quantify cardiac involvement in patients with systemic amyloidosis. The myocardial segment contraction peak time is widely used to quantify myocardial dyssynchrony on echocardiographs; the greater variability in the time to peak strain indicates asynchronous contraction. We found that variability in the peak CS time significantly correlated with clinical severity parameters such as IVS wall thickness, LGE volume, BNP level, and the E/e’ ratio; it was a significant CMR parameter for the detection of LGE-positive amyloidosis.
Currently, CMR remains the reference standard for the assessment of regional function30 and the most widely validated reproducible tool for myocardial strain quantification.31 Several CMR-based myocardial strain imaging techniques using different acquisition schemes [e.g., cSPAMM, strain-encoded imaging (SENC), displacement-encoded with stimulated-echo (DENSE), and feature tracking] are now available; these techniques are more extensive, more sophisticated, and better than they were 20 years ago.7,32 Although most of these techniques were developed separately, they are in fact closely related. With the cSPAMM method we used, black lines or grids (special radiofrequency prepulses called tags) are superimposed on and embedded in the myocardium at the beginning of a cine sequence [just before the start of contraction (at the R wave of the ECG)], and a subsequent deformation of these lines throughout the cardiac cycle is observed. The fading effect from T1 relaxation (mainly during diastole) is seen as time passes after the application of tags. Because the myocardial T1 relaxation time is longer at 3.0 T than at 1.5 T (1,100 vs. 900 ms), the fading effect is less pronounced at higher field strengths. Thus, CMR-based myocardial strain imaging using 3.0-T MRI scanners can be used to evaluate myocardial strain.
Study LimitationsThe sample size was small. However, our series was relatively large for this rare disease and our patients were registered consecutively over the course of 5 years. Second, while our patients had histologically proven amyloid deposits, few had undergone endomyocardial biopsy, the standard clinical practice at our institute. Consequently, we did not examine the relationship between histologic findings and myocardial strain. However, with the current diagnostic consensus strategies for cardiac amyloidosis,2 cardiac involvement can reasonably be inferred in patients with proven systemic amyloidosis by the combination of clinical features, ECG, echocardiography, and biomarkers. Endomyocardial biopsy is only required when suspected cardiac amyloidosis is an isolated feature or when the cardiac amyloid fibril type cannot be identified by other means. Third, we evaluated only CS in the midventricular plane. We did not assess longitudinal strain because with our software we were unable to determine the longitudinal strain that corresponds with motion towards the apex (base-to-apex shortening with an accordion-like movement). Longitudinal strain has been studied extensively and has emerged as the most useful parameter for speckle-tracking echocardiography. “Apical sparing”, a useful finding on echocardiographs, may be reflected by longitudinal strain on CMR images. Fourth, we did not compare myocardial strain findings between CMR and echocardiography because echocardiographic strain data were not available in our institute. Additional studies are needed to evaluate the myocardial strain findings in patients with cardiac amyloidosis by comparing CMR and echocardiography findings. Lastly, we did not compare the CMR strain findings in patients with AL, ATTRm, and ATTRwt amyloidosis because the sample size was too small for meaningful comparative analysis. Large-scale comparative studies are needed to define the CMR strain features of each type of amyloidosis.
In conclusion, myocardial strain analysis by unenhanced CMR helped to detect LGE-positive patients with cardiac amyloidosis whose prognosis is thought to be poor. The peak CS and variability in the peak CS time may correlate with the severity of cardiac amyloid deposition and may be more sensitive than LGE imaging for the detection of early stage disease in patients with cardiac amyloidosis.