Myocardial Stiffness by Intrinsic Cardiac Elastography in Patients with Amyloidosis: Comparison with Chamber Stiffness and Global Longitudinal Strain

Myocardial Stiffness by Intrinsic Cardiac Elastography in Patients with Amyloidosis: Comparison with Chamber Stiffness and Global Longitudinal Strain Cristina Pislaru, MD, Filip Ionescu, MD, Mahmoud Alashry, MBBCh, Ioana Petrescu, MD, Patricia A. Pellikka, MD, Martha Grogan, MD, Angela Dispenzieri, MD, and Sorin V. Pislaru, MD, PhD, Rochester, Minnesota

Background: The aim of this study was to test the hypothesis that intrinsic cardiac elastography can detect diastolic tissue abnormalities produced by cardiac amyloid infiltration and that measurements may have incremental value beyond traditional echocardiographic measures. The specific aims were (1) to evaluate the relationship between left ventricular myocardial stiffness (by elastography) and measures of diastolic chamber stiffness and systolic strain in patients with amyloidosis and (2) to compare their prognostic potential. Methods: We prospectively studied 67 patients with amyloidosis (cardiac amyloidosis, n = 48; noncardiac amyloidosis, n = 19) and 40 normal subjects. Patients underwent comprehensive echocardiography including measurement of left ventricular global longitudinal strain (GLS) by speckle-tracking. Intrinsic velocity propagation of myocardial stretch (iVP), a direct measure of myocardial elasticity, was quantified using intrinsic cardiac elastography. Chamber stiffness was evaluated from the end-diastolic pressure-volume relationships (P = aVb). The major end point at follow-up was the composite of death, cardiac hospitalization, worsening heart failure, and stroke. Results: The iVP of myocardial stretch was highest in patients with cardiac amyloidosis compared with those with noncardiac amyloidosis and normal subjects (3.2 6 1.0, 1.8 6 0.4, and 1.6 6 0.2 m/sec, respectively; P < .0001) and correlated with chamber stiffness, function, and structure (b coefficient, operating chamber stiffness, GLS, wall thickness; P # .001 for all). At follow-up (median, 2.6 years), measures of left ventricular and myocardial stiffness, GLS, diastolic dysfunction grade, and N-terminal pro–brain natriuretic peptide were associated with excess events. At multivariate analysis, iVP of myocardial stretch remained an independent predictor of adverse events, incremental to GLS and N-terminal pro–brain natriuretic peptide. Conclusions: Measurements by cardiac elastography correlate with functional and structural derangements produced by cardiac amyloid infiltration but provide unique information that is incremental to conventional echocardiography. (J Am Soc Echocardiogr 2019;-:---.) Keywords: Amyloidosis, Echocardiography, Diastolic function, Elasticity, Myocardial stiffness

Cardiac amyloidosis (CA) is a rare infiltrative heart disease caused by interstitial deposition of amorphous amyloid fibrils in the myocardial tissue1-5 and portends a poor prognosis.3-5 Amyloid deposition

From the Department of Cardiovascular Medicine (C.P., F.I., M.A., I.P., P.A.P., M.G., S.V.P.) and the Division of Hematology (A.D.), Mayo Clinic, Rochester, Minnesota. This work was supported in part by a Prospective Research Award (to Dr. Cristina Pislaru) from the Mayo Clinic. Conflicts of Interest: None. Reprint requests: Cristina Pislaru, MD, Mayo Clinic, 200 First Street SW, Rochester, MN 55905 (E-mail: [email protected]). 0894-7317/$36.00 Copyright 2019 by the American Society of Echocardiography. https://doi.org/10.1016/j.echo.2019.04.418

reduces systolic function6-10 and increases myocardial stiffness and filling pressures.11,12 However, there remains a need for early detection and improved risk stratification of these patients. Global longitudinal strain (GLS) has emerged as the most sensitive echocardiographic marker of left ventricular (LV) dysfunction and a strong predictor of outcome.13-15 To date, there are no established clinical techniques to quantify myocardial stiffness noninvasively, and the current gold standard relies on analysis of chamber stiffness (CS) from the end-diastolic pressure-volume relationship (EDPVR).16 Intrinsic cardiac elastography has been recently proposed as a novel method to quantify myocardial elasticity.17 The clinical potential of this method was reported in several conditions, such as aortic stenosis and mitral regurgitation,18 amyloidosis,19 pulmonary hypertension,20 and hypertrophic cardiomyopathy.21 However, several questions remain unanswered. 1

2 Pislaru et al

Abbreviations

AL = Amyloid light-chain ATTR = Amyloid transthyretin CA = Cardiac amyloidosis cMRI = Cardiac magnetic resonance imaging CS = Chamber stiffness CS1 = Chamber stiffness derived from the end-diastolic pressure-volume relationship CS2 = Chamber stiffness index represented by the quotient of E/e0 to enddiastolic volume

EDPVR = End-diastolic pressure-volume relationship

Journal of the American Society of Echocardiography - 2019

Herein we present the comprehensive results from a prospective study in patients with amyloidosis with the following specific aims (1) to evaluate the pathophysiological relationship between myocardial elasticity measured noninvasively by cardiac elastography with measures of diastolic CS and systolic strain measured by conventional echocardiography and (2) to compare the prognostic potential of these measures.

METHODS

The protocol was approved by the institutional review board at strain our institution and complied iVP = Intrinsic velocity with the Declaration of propagation of myocardial Helsinki; written informed constretch sent was obtained from each LV = Left ventricular subject. We studied consecutive patients with biopsy-proven sysLVEF = Left ventricular temic amyloidosis referred for ejection fraction echocardiographic evaluation in NCA = Noncardiac various stages of their disease. A amyloidosis control group was recruited from healthy volunteers or paNT-proBNP = N-terminal tients referred to the echocardipro–brain natriuretic peptide ography laboratory and found NYHA = New York Heart to have normal results on echoAssociation cardiography. Preliminary results V30 = Left ventricular volume from a smaller cohort have been at 30 mm Hg previously published.19 The diagnosis of amyloidosis was confirmed in all patients by tissue biopsy (heart, subcutaneous fat, kidney, liver, stomach, rectum, and/ or bone marrow) and typical Congo red staining for the detection of amyloid deposition. Cardiac involvement was diagnosed on the basis of cardiac biopsy findings or by positive results on noncardiac biopsy in conjunction with typical findings on cardiac magnetic resonance imaging (cMRI; e.g., difficult myocardial nulling and subendocardial or midmyocardial late gadolinium enhancement not related to a specific coronary territory) or echocardiography (e.g., characteristic granular ‘‘sparkling’’ appearance of the myocardium, relative apical sparing at strain imaging, increased wall thickness >12 mm unexplained by other causes, thickened interatrial septum, heart valves and right ventricular walls, and pericardial effusion) or typical findings at 99mTc-pyrophosphate scans.3,4 Exclusion criteria were atrial fibrillation or ventricular paced rhythm, severe valvular heart disease, history of cardiac surgery, congenital heart disease, and suboptimal echocardiographic image quality. Control subjects with histories of mild hypertension were not excluded if they had well-controlled blood pressure and no clinical or imaging signs of hypertensive heart disease. GLS = Global longitudinal

Echocardiography All patients underwent comprehensive transthoracic imaging as clinically indicated, and measurements were performed according to guidelines. Diastolic dysfunction was graded according to the latest (2016) recommendations.22 Normal subjects underwent similar transthoracic imaging to rule out significant cardiovascular, congenital, and significant valvular heart disease. LV ejection fraction (LVEF) > 53% was considered normal.23 Relative wall thickness was calculated as (2
posterior wall thickness)/LV diameter at end-diastole. Myocardial Strain by Two-Dimensional Speckle-Tracking LV myocardial strain was quantified by speckle-tracking from Bmode data acquired in three standard apical views (four-chamber, two-chamber, and long-axis views) and three short-axis views (basal, mid, and apical) using clinical scanners (E9 or E95 scanners, with EchoPAC software; GE Vingmed Ultrasound, Horten, Norway). Frame rates were between 50 and 70 frames/sec. GLS was obtained by averaging 18 segmental values from the apical views. Apical sparing (‘‘bull’s-eye’’ pattern) was quantified as relative apical strain = apical strain/(basal strain + mid strain)8; a ratio > 1.0 is considered abnormal.8 In a subset (n = 67), global circumferential strain was also measured from B-mode data acquired at three short-axis levels. Global circumferential strain was calculated by averaging all segmental values. Intrinsic Cardiac Elastography Additional ultrahigh–frame rate tissue Doppler data (250–465 frames/sec) were acquired using the same ultrasound scanners, from the same apical views, one wall at a time, while carefully aligning each LV wall with the Doppler beams. The base-to-apex propagation of the late diastolic myocardial stretch wave can be visualized only with ultrafast imaging data. The slope of the isovelocity wave front, called intrinsic velocity propagation of myocardial stretch (iVP; in meters per second), was measured offline in each wall, as previously described.17,18 A global value was generated by averaging values for three cardiac cycles and all six walls. Figure 1 shows two examples from a normal subject and a patient with CA. Reproducibility of this measurement was found to be very good (coefficient of variability = 5%–9%; intraclass correlation coefficient = 0.95 [95% CI, 0.91–0.98] and 0.91 [95% CI, 0.78–0.98] for intraobserver and interobserver agreement, respectively18). Elastic modulus (in pascals) was calculated from iVP and geometry14; due to the fact that modulus did not show an incremental value over iVP, the details for this parameter are presented in Appendix 1. Assessment of LV Chamber Stiffness LV CS was analyzed from the EDPVR (P = aVb) reconstructed noninvasively by the validated method proposed by Klotz et al.24 This method was shown to accurately predict invasively measured EDPVR using conductance catheters (R = 0.83–0.98).25 More details are given in Appendix 2. End-diastolic pressure was estimated from the E/e0 ratio as determined from invasive studies at our institution (1.96 + 0.596
E/e0 ).26 Using this method, b stiffness coefficient, LV capacitance (LV volume at 30 mm Hg [V30]), and operating CS (CS1; the slope of the tangent to EDPVR) were derived and directly compared with iVP and elastic modulus. CS at V30 was also calculated but did not show an advantage over CS1, so results are not

Pislaru et al 3

Journal of the American Society of Echocardiography Volume - Number -

HIGHLIGHTS  A novel ultrahigh frame rate, tissue Doppler-based cardiac elastography method was developed.  This method was applied in patients with cardiac amyloidosis.  Findings by elastography correlated with cardiac functional and structural derangements.  These measurements were incremental to known echocardiographic predictors of outcomes. presented. Another simplified index of CS (CS2) was calculated as the quotient of E/e0 to end-diastolic volume; this index was previously validated against invasive EDPVR.27 Other Clinical and Laboratory Data Serum N-terminal pro–brain natriuretic peptide (NT-proBNP) and cardiac troponin T levels measured within 7 days from the incident study were recorded. Cardiac staging was quantified using the Mayo Clinic three-stage scoring system on the basis of NT-proBNP ($332 pg/mL) and troponin T ($0.035 ng/mL) levels.28 Functional capacity was graded according to the New York Heart Association (NYHA) classification. P-R interval duration, QRS width, and incidence of bundle branch block and nonspecific intraventricular conduction delays were recorded from electrocardiograms obtained during the same patient visit. Follow-Up Follow-up data were prospectively collected as patients returned for routine care visits. The major end point was a composite of death, cardiac hospitalization or worsening heart failure symptoms (by one or more NYHA class), and stroke. Vital status was verified from the electronic medical records and the Social Security Death Index.

Statistical Analysis Analysis was performed using JMP version 12 (SAS Institute, Cary, NC). Data are presented as mean 6 SD, median (interquartile range), or absolute number (percentage). Normal distribution was verified using the Shapiro-Wilk statistic and the Kolmogorov-Smirnov test, and transformations were performed when needed (log NT-proBNP). Comparisons across groups were tested using one-way analysis of variance followed by pairwise comparisons using Tukey-Kramer and Wilcoxon rank-sum tests. Categorical data were compared using c2 or Fisher exact tests. Correlations between variables were evaluated using Spearman’s r rank correlation test. Receiver-operating characteristic analysis was used to determine the optimal cutoffs while maximizing sensitivity and specificity. Outcomes were analyzed with Kaplan-Meier and Cox proportional-hazards methods. Univariate and multivariate Cox proportional-hazards models were used to examine the associations of variables of interest to predict adverse events. Hazard ratios and their 95% CIs were calculated; for continuous variables, hazard ratios are presented per unit change in regressor (unless indicated differently). Multivariate regression analysis was performed by including variables with P values < .10 on univariate testing and using a backward stepwise selection method. To minimize overfitting, we constrained the number of variables in the final model according to the number of events. The incremental value of iVP over other variables or models including GLS and NT-proBNP to predict the outcomes was evaluated from the difference in log-likelihood and c2 testing. All P values were two sided, and the significance level was inferred for P < .05.

RESULTS A total of 120 subjects were enrolled for this study; of these, 13 were excluded (concomitant moderate to severe valvular heart disease in three, atrial fibrillation in two, amyloidosis other than amyloid light-

Figure 1 Examples of measurement of iVP in the left ventricle of a normal subject and a patient with CA. iVP was measured as the slope of the isovelocity wave front propagating from base to apex at the onset of late diastole (red arrows). This wave speed was 1.7 m/sec in the normal subject but much higher (5.1 m/sec) in the patient with CA.

4 Pislaru et al

Journal of the American Society of Echocardiography - 2019

Table 1 Baseline characteristics CA (n = 48)

Variable

AL CA (n = 40)

All CA

64 6 10*

Age (y) Male

28 (58)

Heart rate (beats/min) SBP (mm Hg) DBP (mm Hg)

114 6 21 67 6 10

ATTR CA (n = 8)

63 6 11 21 (53)

72 6 13*†

NCA (n = 19)

68 6 9

66 6 9*

7 (88)

72 6 13*† 116 6 21

Normal subjects (n = 40)

AL NCA

49 6 13

15 (79)

23 (58%)

70 6 12

64 6 10

65 6 10

106 6 18

121 6 21

117 6 13

68 6 11

71 6 9

67 6 11

65 6 6

Symptomatic

38 (79)

31 (78)

7 (88)

11 (58)

0

NYHA classes III and IV

14 (29)

10 (25)

4 (50)

0

0

History of HTN

13 (27)

11 (28)

2 (25)

9 (47)

Diabetes mellitus

1 (2)

1 (3)

1 (5)

0

13 (27)

15 (38)

3 (37)

10 (53)

3 (8%)

History of CAD

6 (13)

5 (13)

1 (13)

1 (5)

0

CKD

6 (13)*

5 (13)

1 (13)

3 (16)*

0

Dyslipidemia

0

10 (25%)

NT-proBNP (pg/mL)

1,413 (694–3,732)†

1,615 (654–3,862)†

1,413 (1,015–3,405)

Troponin T (ng/mL)

0.01 (0.01–0.07)

0.01 (0.01–0.06)

0.03 (0.01–0.05)

102 (73–295) 0.01 (0.01–0.01)

Diuretics

29 (60)†

24 (60)†

5 (63)

3 (16)

Newly diagnosed

12 (25)

5 (13)

7 (88)

4 (21)

Data are expressed as mean 6 SD, as number (percentage), or as median (interquartile range). CAD, Coronary artery disease; CKD, chronic kidney disease (estimated glomerular filtration rate < 30 mL/min/1.73 m2); DBP, diastolic blood pressure; HTN, hypertension; SBP, systolic blood pressure. *P < .05 versus normal subjects. † P < .05 versus patients with NCA.

Table 2 Echocardiographic measurements CA (n = 48)

Variable

All CA

NCA (n = 19)

AL CA (n = 40)

ATTR CA (n = 8)

AL

Normal subjects (n = 40)

Ejection fraction (%)

56 6 11*

58 6 11

51 6 12

64 6 5

63 6 4

CI (L/min/m2)

2.9 6 0.8

2.9 6 0.7

2.8 6 1.2

3.1 6 0.6

3.1 6 0.6

SVI (mL/m2)

40 6 11*†

†

40 6 9*†

42 6 17

50 6 9

48 6 9

13.6 6 2.7*†

16.8 6 3.6

10.6 6 2.1

9.7 6 1.2

MWT (mm)

14.1 6 3.1*

LVEDD (mm)

45.5 6 4.8*

45.6 6 4.9

44.8 6 4.4

48.9 6 4.5

48.4 6 4.1

LVMI (g/m2)

135 6 39*†

131 6 35*†

171 6 46

95 6 28

86 6 19

40 6 11*

40 6 11*

2

LAVI (mL/m )

†

†

GLS (%)

13.8 6 4.5*†

GCS (%)

16.3 6 5.2*†

43 6 4

†

14.4 6 4.4*† 16.9 6 5.0*†

32 6 7

30 6 6

11.0 6 4.3

20.1 6 1.8

20.2 6 2.0

12.6 6 5.7

21.1 6 4.3*†

22.4 6 2.7

Relative apical strain

0.84 6 0.27*

0.80 6 0.24*

1.08 6 0.26

0.60 6 0.09

0.54 6 0.09

E (m/sec)

0.87 6 0.24†

0.87 6 0.24†

0.88 6 0.24

0.65 6 0.18

0.69 6 0.17

A (m/sec)

0.53 6 0.27†

0.54 6 0.26†

0.48 6 0.30

0.76 6 0.23*

0.66 6 0.18

E/A ratio

2.0 6 1.1*†

2.0 6 1.0*†

2.4 6 1.2

0.9 6 0.4

1.1 6 0.4

E/A ratio > 2

22 (46)*†

18 (45)*†

e0 (average) (cm/sec)

5.0 6 1.7*†

5.1 6 1.8

4.4 6 1.3

7.5 6 2.2*

8.6 6 1.9

19.9 6 9.1*†

19.7 6 9.8*†

20.8 6 5.0

9.0 6 2.8

8.2 6 2.2

E/e0 (average)

†

†

4 (50)

0

1 (3)

Data are expressed as mean 6 SD or as number (percentage). CI, Cardiac index; GCS, global circumferential strain; LAVI, left atrial volume index; LVEDD, LV end-diastolic diameter; LVMI, LV mass index; MWT, mean wall thickness; SVI, stroke volume index. *P < .05 versus normal subjects. † P < .05 versus NCA.

Journal of the American Society of Echocardiography Volume - Number -

Pislaru et al 5

Figure 2 Cardiac elastographic results. (A) iVP was higher in patients with CA compared with those with NCA and normal subjects. (B) Significant differences were observed regardless of whether patients had cardiac biopsy– or cMRI-confirmed (+cMRI) CA. (C) There was a trend toward lower iVP in patients with AL versus ATTR CA.

Figure 3 Group average LV EDPVR (P = aVb). Patients with CA had steeper slopes and leftward shift of EDPVR compared with those with NCA and normal subjects, indicating chamber stiffening. This was reflected by the higher b stiffness coefficient and operating CS.

6 Pislaru et al

Journal of the American Society of Echocardiography - 2019

Figure 4 Correlation between iVP and measures of systolic function. The best correlation was found with GLS, followed by circumferential strain, and the lowest with ejection fraction. The X indicates patients with ATTR CA.

chain [AL] or amyloid transthyretin [ATTR] amyloidosis in three, incomplete iVP measurement in five). The final cohort included 107 subjects: 67 patients with systemic amyloidosis (88% with AL amyloidosis, 12% with ATTR amyloidosis) and 40 normal subjects. Of the 67 patients, 48 had cardiac involvement (CA group) and 19 had no evidence of cardiac involvement (noncardiac amyloidosis [NCA] group). Cardiac involvement was diagnosed on the basis of positive findings on endomyocardial biopsy (n = 11) or cMRI (n = 22; median time from recruitment, 2.7 months) or on the basis of positive noncardiac biopsy findings combined with characteristic features of cardiac amyloid infiltration on echocardiography (including strain imaging), electrocardiography, and laboratory findings. Baseline Characteristics Patients with amyloidosis were more frequently men, but they had similar age compared with normal subjects (Table 1). Most echocardiographic measurements were markedly impaired in patients with CA (Table 2). LVEF was preserved in most patients, but cardiac index was not different between groups. Myocardial Elasticity versus CS Figure 2A shows that iVP was markedly higher in patients with CA compared with those with NCA and normal subjects (3.2 6 1.0,

1.8 6 0.4, and 1.6 6 0.2 m/sec, respectively; P < .0001). Similar findings were seen for the elastic modulus (Appendix 1, Supplemental Figure 1; available at www.onlinejase.com). A similar strong difference was seen in the 33 patients with cardiac biopsy– or cMRI-confirmed CA compared with normal subjects (Figure 2B). The differences between patients with AL and those with ATTR amyloidosis were only borderline (Figure 2C). Patients with CA had more regional heterogeneity in iVP (higher SD between all LV walls) compared with normal subjects (P < .0001); however, both septal and nonseptal walls contributed to increased global iVP. Figure 3 shows the group average EDPVR. This relation was shifted leftward and was steeper (higher slope) in CA, indicating increased CS at all pressure levels. The load-independent b coefficient (P < .0001) and load-dependent CS (CS1, CS2; P < .0001 for both; Figure 3) were markedly higher in patients with CA compared with those with NCA and normal subjects, consistent with findings by iVP. These measures were, however, modestly correlated with iVP (b coefficient: R = 0.63, P < .0001; CS1: R = 0.43, P < .001; CS2: R = 0.51, P < .0001). LV capacitance (V30) tended to be smaller in patients with CA, but differences among groups did not reach statistical significance. Comparison of Myocardial Elasticity with Systolic Strain LVEF was normal in 69% of patients with CA and all patients with NCA. iVP was inversely correlated with LVEF (R = 0.46,

Journal of the American Society of Echocardiography Volume - Number -

Pislaru et al 7

Figure 5 Relationships between iVP and (A) LV relaxation, (B) estimated filling pressures, (C) severity of diastolic dysfunction, (D) NTproBNP level (log transformed), and (E) NYHA functional class. The X indicates patients with ATTR CA. P < .001) and global circumferential strain but most strongly with impairment in GLS (R = 0.76, P < .0001) and basal longitudinal strain (R = 0.79, P < .0001; Figure 4), which was characteristically more impaired than apical strain in patients with CA (‘‘apical sparing’’). There was no significant relation between iVP and LVEF, GLS, or global circumferential strain in normal subjects. Comparison of Myocardial Elasticity with Measures of Diastolic Function There was a progressive increase in iVP with worsening of LV relaxation (mitral annular e0 velocity: R = 0.56, P < .0001; Figure 5A), increasing filling pressures (E/e0 ratio: R = 0.64, P < .0001; Figure 5B), larger left atrial volumes, and more advanced diastolic dysfunction (P < .0001; Figure 5C). Similar relationships were found with NT-proBNP levels (R = 0.71, P < .0001; Figure 5D), Mayo Clinic staging score (for AL amyloidosis, P = .005), and NYHA functional class (P < .0001; Figure 5E). Relationship with Structural and Electrocardiographic Abnormalities, Age, Sex, and Hemodynamics Considering only patients with amyloidosis, there was a moderate correlation between iVP and indirect markers of amyloid burden such as increased septal and mean wall thickness (R = 0.62, P < .0001), relative wall thickness (R = 0.62, P < .0001), LV mass index (R = 0.58, P < .0001), and NT-proBNP (Figure 5D).

There was no significant relationship between iVP and P-R interval duration. iVP tended to be higher in patients with bundle branch block (n = 7; P < .05 vs rest) and with wider QRS intervals (P < .01). Age, systemic blood pressure, and heart rate did not show a consistent relationship with iVP, in either group. Male patients with amyloidosis tended to have higher iVP than women (P < .05), but this difference was not seen in normal subjects. Outcomes During follow-up (median, 2.6 years; interquartile range, 1.9– 3.2 years), 22 patients (33%) developed 26 adverse events (12 deaths, cardiac hospitalization or worsening heart failure symptoms in 12, and stroke in two). An iVP value of 2.7 m/sec was the optimal cutoff that best predicted the occurrence of adverse events at followup (area under the curve = 0.78; P < .0001). A higher iVP (>2.7 m/sec) identified patients at increased risk for adverse events (P < .0001; Figure 6). A graded effect was seen for tertiles of iVP (Appendix 1, Supplemental Figure 2: available at www. onlinejase.com). Similar trends were seen in patients with CA only or AL amyloidosis (Appendix 1, Supplemental Figures 3 and 4: available at www.onlinejase.com). Figure 6 shows the Kaplan-Meier curves for iVP, GLS, and measures of CS (b > 6.4, log-rank P = .002; CS1 > 1.8 mm Hg/mL, log-rank P = .02). The predictive strength of several variables of interest is compared in Figure 7, and their hazard ratios are presented in Table 3. Among all variables, iVP (hazard ratio, 2.4 for each 1 m/sec increase; P < .0001)

8 Pislaru et al

Journal of the American Society of Echocardiography - 2019

Figure 6 Kaplan-Meier analysis for predicting adverse events at follow-up. (A) iVP. (B) GLS. (C) CS. (D) Load-independent CS coefficient. Among all measures, iVP was the best predictor. was strongly associated with the composite end point and was superior to measures of CS, myocardial strain, NT-proBNP, and other measures of diastolic function (Figure 7A). The same conclusion was reached for patients with AL amyloidosis only (Appendix 1, Supplemetal Figure 5: available at www.onlinejase.com). On bivariate and multivariate analysis, iVP consistently remained a significant independent predictor of outcomes, even after adjusting for GLS, NT-proBNP, severity of diastolic function, NYHA class, and measures of CS. The addition of iVP (or elastic modulus) to other variables such as GLS, NT-proBNP, cardiac staging system, LVEF, stroke volume index, and a composite GLS and NT-proBNP model was incremental and improved model performance for outcome prediction (P < .05 for all; Figure 7B). Overall mortality was also higher in patients with an iVP > 3.8 m/s (P = .003), but this result is regarded as preliminary because of the small number of deaths in this cohort. DISCUSSION In this study we demonstrate that intrinsic cardiac elastography can detect presence of increased LV/myocardial stiffening in patients with CA as confirmed by an independent method (EDPVR), and measures of myocardial elasticity had strong predictive value incremental to functional assessment by conventional echocardiography including strain imaging.

Myocardial Stiffness in CA Cardiac involvement in amyloidosis predicts a poor prognosis, particularly in patients with AL amyloidosis.3,4 There is still a need for improved risk stratification of these patients.4 Newer measures have emerged based on the assessment of LV longitudinal strain, with apical-sparing pattern now being recognized as characteristic for CA while LVEF is still preserved.6,7 However, these parameters do not evaluate tissue elasticity but reflect the functional consequences of amyloid deposition disrupting cardiac architecture and impairing cardiomyocyte function.9,29,30 CA also causes profound diastolic disturbances, with increased LV and myocardial stiffness and restrictive filling physiology (‘‘stiff heart syndrome’’). Currently, there are no established clinical techniques to directly measure myocardial elasticity. Intrinsic cardiac elastography is a novel approach based on naturally occurring stretch waves during diastole: the faster the wave speed, the stiffer the myocardial tissue.17,18 This wave could be considered analogous to pulse-wave velocity in arteries, a classical measure of arterial stiffness.31 The potential of this measurement in CA was briefly reported.19 Herein we present for the first time that myocardial elasticity compares well with measures of chamber stiffening evaluated from EDPVR and with impairment in LV longitudinal and circumferential deformation by speckle-tracking echocardiography. A steeper nonlinear EDPVR relation in patients with CA compared with those with NCA and normal subjects indicates the presence of reduced

Pislaru et al 9

Journal of the American Society of Echocardiography Volume - Number -

Figure 7 (A) Univariate predictors of outcomes at Cox proportional-hazards regression analysis. A higher c2 value indicates a stronger predictor. (B) Incremental predictive value of myocardial elasticity in multivariate models.

Table 3 Univariate predictors of outcomes at Cox proportional hazards regression analysis Predictor

Hazard ratio (95% CI)

P

iVP

2.44 (1.72–3.44)

<.0001

Elastic modulus

1.59 (1.31–1.92)*

<.0001

Age

1.04 (0.99–1.09)

.062

Male

1.70 (0.66–5.18)

.283

NYHA class III or IV

4.18 (1.62–10.2)

.004

4.20 (2.15–8.69)

<.0001

Log(NT-proBNP) Cardiac staging system

12.51 (1.45–107.86)†

.006

Ejection fraction

0.92 (0.88–0.96)

<.0001

GLS

1.23 (1.12–1.36)

<.0001

Global circumferential strain

4.66 (1.34–16.24)

.006

Relative apical strain

8.51 (1.68–39.4)

.007

Mean wall thickness

1.23 (1.08–1.39)

.001

Left atrial volume index

1.06 (1.02–1.11)

.004

Restrictive filling (E/A ratio > 2)

5.72 (2.37–13.86)

.0001

E/e0 ratio

1.08 (1.04–1.13)

<.0001

b

1.37 (1.17–1.60)

<.0001

CS1

1.81 (1.18–2.70)

.004

CS2

3.68 (1.51–8.91)

.003

b, exponent of the EDPVR. *Hazard ratio expressed per 10-unit change in regressor. † Hazard ratio for stage I versus III.

chamber compliance at all filling pressures, supporting the findings by cardiac elastography. This reduced compliance limits the use of the Frank-Starling mechanism to enhance cardiac function without increasing filling pressures and likely contributes to pulmonary congestion and dyspnea. This chamber stiffening results mainly from the amyloid infiltration and its consequences, as reflected by the increased wall thickness despite normal chamber dimensions, higher LV mass,11 and elevated NT-proBNP32 and filling pressures,12 as confirmed in this study, and to some extent from the effect of comorbidities. The results clearly demonstrate that myocardial elasticity correlated well with all functional and structural derangements as evaluated by echocardiography and typically seen in patients with CA. The modest correlations however support the predictions that none of the traditional measures can fully reflect passive tissue elasticity. The abnormalities in myocardial elasticity were indirectly reflected by the reduction in myocardial strain, both longitudinal and circumferential, with the former being somewhat more sensitive, likely because of the higher vulnerability of the subendocardium to injury even in the setting of normal LVEF.33 Our results are consistent with findings from a smaller study using cardiac magnetic resonance elastography in 22 patients with CA.34 Of note, most shear-wave methods proposed need an external force and equipment to induce and track propagating waves and extract measures of elasticity,34,35 while the intrinsic method used here uses naturally induced waves and clinical scanners without the need for additional equipment. Other groups suggested the use of diastolic strain rates,36 diastolic wall strain, and LVEF/GLS ratio,37 but these parameters reflect more closely systolic function rather than passive stiffness.

10 Pislaru et al

The elastic modulus did not show an incremental value over the measured wave speed. Computer modeling studies predict that ventricular geometry has only a minor effect on the wave speed.38 Predictive Value of Myocardial Elasticity Compared with GLS and CS These results demonstrate for the first time that measures of myocardial elasticity by intrinsic cardiac elastography were incremental to other known predictors of outcomes such as GLS13-15 or NTproBNP.3,4 The finding that measures of chamber stiffening evaluated by EDPVR were also predictors of outcomes is also novel; nevertheless, their association with outcomes was weaker, possibly because of the large variability in chamber geometry seen in these patients. These results are sound considering that chamber properties are influenced not only by tissue properties but also by geometry and external factors16 and highlight the importance of directly measured tissue properties. Current results further reinforce our preliminary findings for predicting outcomes,19 despite the more restrictive composite end point used in this study. Further studies are needed to evaluate the potential of this method in larger patient populations, to characterize differences in myocardial elasticity between amyloid subtypes, and to evaluate the potential to monitor changes with treatment. Limitations We included patients at various stages of the disease, so future studies are needed to evaluate the prognostic impact at initial diagnosis; however, the focus of this report was to evaluate a novel measurement rather than its diagnostic capability. In this study, 69% of patients in the CA group had cardiac biopsy–confirmed (gold standard) or cMRI-confirmed CA, but statistical significance was achieved even in patients without cardiac biopsy or cMRI; moreover, all patients had typical features consistent with CA on echocardiography at initial diagnosis, including apical sparing on strain imaging, which is a powerful sign in the absence of other causes of LV hypertrophy.4,6-8 Invasive measurements of pressure and volume were not performed, but the noninvasive methods used here have all been validated against the invasive gold standard. The effect of amyloid deposition on myocardial stiffness cannot be separated from that due to chronic hypertension; however, in this cohort there were no differences in iVP among patients or normal subjects with versus without histories of hypertension. The methodology used here cannot be applied in patients with nonsinus rhythm. Presence of advanced restrictive physiology may limit wave propagation throughout the ventricle if filling during late diastole is too weak; however, the wave speed does not depend on the strength of atrial contraction.39 Doppler measurements are angle dependent, but measurements as performed here should be comparable with speckle-tracking while taking advantage of the ultrahigh temporal resolution needed. Data acquisition and offline analysis are time consuming, and customized semiautomatic analysis routines are currently being developed. CONCLUSION This study demonstrates that measurements by cardiac elastography correlate with functional and structural derangements produced by cardiac amyloid infiltration but provide unique information that is incremental to functional assessment by echocardiography including

Journal of the American Society of Echocardiography - 2019

speckle-tracking strain imaging. Further studies are needed to fully understand the role of this novel measurement for specific clinical applications.

SUPPLEMENTARY DATA Supplementary data to this article can be found online at https://doi. org/10.1016/j.echo.2019.04.418.

REFERENCES 1. Cueto-Garcia L, Reeder GS, Kyle RA, Wood DL, Seward JB, Naessens J, et al. Echocardiographic findings in systemic amyloidosis: spectrum of cardiac involvement and relation to survival. J Am Coll Cardiol 1985;6: 737-43. 2. Klein AL, Hatle LK, Burstow DJ, Seward JB, Kyle RA, Bailey KR, et al. Doppler characterization of left ventricular diastolic function in cardiac amyloidosis. J Am Coll Cardiol 1989;13:1017-26. 3. Gertz MA. Immunoglobulin light chain amyloidosis: 2016 update on diagnosis, prognosis, and treatment. Am J Hematol 2016;91:948-56. 4. Grogan M, Dispenzieri A, Gertz MA. Light-chain cardiac amyloidosis: strategies to promote early diagnosis and cardiac response. Heart 2017;103: 1065-72. 5. Ruberg FL, Berk JL. Transthyretin (TTR) cardiac amyloidosis. Circulation 2012;126:1286-300. 6. Koyama J, Ray-Sequin PA, Falk RH. Longitudinal myocardial function assessed by tissue velocity, strain, and strain rate tissue Doppler echocardiography in patients with AL (primary) cardiac amyloidosis. Circulation 2003; 107:2446-552. 7. Bellavia D, Abraham TP, Pellikka PA, Al-Zahrani GB, Dispenzieri A, Oh JK, et al. Detection of left ventricular systolic dysfunction in cardiac amyloidosis with strain rate echocardiography. J Am Soc Echocardiogr 2007;20: 1194-202. 8. Phelan D, Collier P, Thavendiranathan P, Popovic ZB, Hanna M, Plana JC, et al. Relative apical sparing of longitudinal strain using two-dimensional speckle-tracking echocardiography is both sensitive and specific for the diagnosis of cardiac amyloidosis. Heart 2012;98:1442-8. 9. Larsen B, Mereuta O, Dasari S, Fayyaz AU, Theis JD, Vrana JA, et al. Correlation of histomorphological pattern of cardiac amyloid deposition with amyloid type: a histological and proteomic analysis of 108 cases. Histopathology 2015;68:648-56. 10. Ishikawa Y, Ishii T, Masuda S, Asuwa N, Kiguchi H, Hirai S, et al. Myocardial ischemia due to vascular systemic amyloidosis: a quantitative analysis of autopsy findings on stenosis of the intramural coronary arteries. Pathol Int 1996;46:189-94. 11. Chew C, Ziady GM, Raphael MJ, Oakley CM. The functional defect in amyloid heart disease: the ‘‘stiff heart’’ syndrome. Am J Cardiol 1975;36: 438-44. 12. Russo C, Green P, Maurer M. The prognostic significance of central hemodynamics in patients with cardiac amyloidosis. Amyloid 2013;20: 199-203. 13. Buss SJ, Emami M, Mereles D, Korosoglou G, Kristen AV, Voss A, et al. Longitudinal left ventricular function for prediction of survival in systemic light-chain amyloidosis: incremental value compared with clinical and biochemical markers. J Am Coll Cardiol 2012;60:1067-76. 14. Barros-Gomes S, Williams B, Nhola LF, Grogan M, Maalouf JF, Dispenzieri A, et al. Prognosis of light chain amyloidosis with preserved LVEF: added value of 2D speckle-tracking echocardiography to the current prognostic staging system. JACC Cardiovasc Imaging 2017;10: 398-407. 15. Pun SC, Landau HJ, Riedel ER, Jordan J, Yu AF, Hassoun H, et al. Prognostic and added value of two-dimensional global longitudinal strain for prediction of survival in patients with light chain amyloidosis undergoing

Journal of the American Society of Echocardiography Volume - Number -

autologous hematopoietic cell transplantation. J Am Soc Echocardiogr 2018;31:64-70. 16. Burkhoff D, Mirsky I, Suga H. Assessment of systolic and diastolic ventricular properties via pressure-volume analysis: a guide for clinical, translational, and basic researchers. Am J Physiol 2005;289:H501-12. 17. Pislaru C, Pellikka PA, Pislaru SV. Wave propagation of myocardial stretch: correlation with myocardial stiffness. Basic Res Cardiol 2014;109:438. 18. Pislaru C, Alashry MM, Thaden JJ, Pellikka PA, Enriquez-Sarano M, Pislaru SV. Intrinsic wave propagation of myocardial stretch, a new tool to evaluate myocardial stiffness: a pilot study in patients with aortic stenosis and mitral regurgitation. J Am Soc Echocardiogr 2017;30:1070-80. 19. Pislaru C, Alashry MM, Ionescu F, Petrescu I, Pellikka PA, Grogan M, et al. Increased myocardial stiffness detected by intrinsic cardiac elastography in patients with amyloidosis: Impact on outcomes. JACC Cardiovasc Imaging 2019;12:375-7. 20. Tunhasiriwet A, Krittanawong C, Toparkngarm P, Alashry M, Pislaru S, Pellikka P, et al. Novel echocardiographic assessment of right ventricular myocardial stiffness in patients with pulmonary hypertension. J Am Coll Cardiol 2016;67(13 suppl):2050. 21. Anupraiwan O, Pislaru SV, Pellikka PA, Pislaru C. Noninvasive quantification of myocardial elasticity in patients with hypertrophic cardiomyopathy. Circulation 2018;138(suppl_1):A16921. 22. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF III, Dokainish H, Edwards T, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2016;29:277-314. 23. Lang R, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, et al. Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2015;28:1-39. 24. Klotz S, Hay I, Dickstein ML, Yi GH, Wang J, Maurer MS, et al. Single-beat estimation of end-diastolic pressure-volume relationship: a novel method with potential for noninvasive application. Am J Physiol 2006;291: H403-12. 25. Ten Brinke EA, Burkhoff D, Klautz RJ, Tsch€ ope C, Schalij MJ, Bax JJ, et al. Single-beat estimation of the left ventricular end-diastolic pressure-volume relationship in patients with heart failure. Heart 2010;96:213-9. 26. Ommen SR, Nishimura RA, Appleton CP, Miller FA, Oh JK, Redfield MM, et al. Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: a comparative simultaneous Doppler-catheterization study. Circulation 2000;102: 1788-94.

Pislaru et al 11

27. Kasner M, Sinning D, Burkhoff D, Tschope C. Diastolic pressure–volume quotient (DPVQ) as a novel echocardiographic index for estimation of LV stiffness in HFpEF. Clin Res Cardiol 2015;104:955-63. 28. Dispenzieri A, Gertz MA, Kyle RA, Lacy MQ, Burritt MF, Therneau TM, et al. Serum cardiac troponins and N-terminal pro-brain natriuretic peptide: a staging system for primary systemic amyloidosis. J Clin Oncol 2004;22:3751-7. 29. Maleszewski JJ. Cardiac amyloidosis: pathology, nomenclature, and typing. Cardiovasc Pathol 2015;24:343-50. 30. Brenner DA, Jain M, Pimentel DR, Wang B, Connors LH, Skinner M, et al. Human amyloidogenic light chains directly impair cardiomyocyte function through an increase in cellular oxidant stress. Circ Res 2004;94: 1008-10. 31. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, et al. Expert consensus document on arterial stiffness: methodological issues and clinical applications. Eur Heart J 2006;27: 2588-605. 32. Palladini G, Campana C, Klersy C, Balduini A, Vadacca G, Perfetti V, et al. Serum N-terminal pro-brain natriuretic peptide is a sensitive marker of myocardial dysfunction in AL amyloidosis. Circulation 2003;107:2440-5. 33. Kalam K, Otahal P, Marwick TH. Prognostic implications of global LV dysfunction: a systematic review and meta-analysis of global longitudinal strain and ejection fraction. Heart 2014;100:1673-80. 34. Arani A, Arunachalam SP, Chang ICY, Baffour F, Rossman PJ, Glasser KJ, et al. Cardiac MR elastography for quantitative assessment of elevated myocardial stiffness in cardiac amyloidosis. J Magn Reson Imaging 2017; 46:1361-7. 35. Vejdani-Jahromi M, Freedman J, Nagle M, Kim YJ, Trahey GE, Wolf PD. Quantifying myocardial contractility changes using ultrasound-based shear wave elastography. J Am Soc Echocardiogr 2017;30:90-6. 36. Liu D, Hu K, Stork S, Herrmann S, Kramer B, Cikes M, et al. Predictive value of assessing diastolic strain rate on survival in cardiac amyloidosis patients with preserved ejection fraction. PLoS ONE 2014;9:e115910. 37. Pagourelias ED, Mirea O, Duchenne J, Van Cleemput J, Delforge M, Bogaert J, et al. Echo parameters for differential diagnosis in cardiac amyloidosis. Circ Cardiovasc Imaging 2017;10:e005588. 38. Carlson K, Pislaru SV, Dragomir Daescu D, Pellikka PA, Pislaru C. Myocardial stiffness by longitudinal wave propagation using finite element modeling and clinical data. J Am Soc Echocardiogr 2018; 31:B104-5. 39. Alashry M, Luis SA, Tunhasiriwet A, Padang R, Ohler E, Nkomo V, et al. Left ventricular myocardial stiffness by intrinsic wave propagation method increases with severity of diastolic dysfunction. J Am Soc Echocardiogr 2016;29:B128-9.

11.e1 Pislaru et al

APPENDIX 1 Elastic Modulus Elastic modulus (in pascals) was calculated from the intrinsic wave speed (iVP) and ventricular geometry1 by assuming the left ventricle as an elastic tube and applying the Moens-Korteweg equation:
E ¼ rc2 D h;

where D is LV end-diastolic diameter, h is mean wall thickness at enddiastole, c is wave speed, and r is tissue density (1,060 kg/m3).

Journal of the American Society of Echocardiography - 2019

Findings on elastic modulus followed the same trend as for iVP, specifically for the markedly increased levels in patients with CA (Supplemental Figure 1) and the significant association with the cardiac adverse events at follow-up (Supplemental Figure 2).

REFERENCE 1. Pislaru C, Pellikka PA, Pislaru SV. Wave propagation of myocardial stretch: correlation with myocardial stiffness. Basic Res Cardiol 2014;109:438.

Supplemental Figure 1 Elastic modulus derived from wave speed. (A) Significant differences were seen among the three groups. (B) Differences were seen between patients with AL and ATTR CA but did not reach statistical significance.

Supplemental Figure 2 Kaplan-Meier analysis stratified by tertiles of intrinsic wave speed (iVP; left) and elastic modulus (right) for predicting adverse events at follow-up.

Journal of the American Society of Echocardiography Volume - Number -

Pislaru et al 11.e2

Supplemental Figure 3 Kaplan-Meier analysis for predicting adverse events at follow-up in patients with CA. (A) iVP. (B) GLS. (C) CS. (D) Load-independent CS coefficient. Among all measures, iVP was the best predictor.

11.e3 Pislaru et al

Journal of the American Society of Echocardiography - 2019

Supplemental Figure 4 Kaplan-Meier analysis for predicting adverse events at follow-up in patients with AL amyloidosis. (A) iVP. (B) GLS. (C) CS. (D) Load-independent CS coefficient. Among all measures, iVP was the strongest predictor.

Supplemental Figure 5 Univariate predictors of outcomes in patients with AL amyloidosis. Among all variables, myocardial stiffness (by iVP) was the strongest predictor.

Pislaru et al 11.e4

Journal of the American Society of Echocardiography Volume - Number -

APPENDIX 2 Reconstruction of EDPVR and Evaluation of Chamber Stiffness The nonlinear LV EDPVR was reconstructed noninvasively using the validated single-beat method proposed by Klotz et al.1 This method was shown to accurately predict invasively measured pressurevolume data in patients using conductance catheters (R = 0.83– 0.98).2 In this method, EDVPR can be described as an analytical expression between pressure and volume as

where Vm and Pm are the measured volume and pressure, V0 and V30 are theoretical volumes at 0 and 30 mm Hg, and An and Bn are empirical coefficients. Using data from 80 human hearts, Klotz et al. showed that V0 is a linear function of the operating volume and An and Bn to be 27.8 6 0.3 and 2.76 6 0.05, respectively. The curve-fitting coefficients can be calculated as follows: . logðPm=30Þ=logðVm=V30Þ a ¼ 30 V 30 ;

and b ¼ logðPm =30Þ=logðVm =V30 Þ:

P ¼ aVb;

where a and b are curve-fitting coefficients that describe EDPVR. A normalization procedure is introduced to account for the varying LV volumes among different subjects on the basis of the assumption of a common EDVPR shape as validated by Klotz et al.1:

In our study, we also calculated CS at operating pressure/volume (CS1) as well as at V30, as the slope of the tangent to the EDPVR curve: CS ¼ a  b  V ðb1Þ :

Vm;n ¼ ðVm  Vn Þ=ðV30  V0 Þ

REFERENCES Pm ¼ An 

V Bn m;n ;

i .h V30 ¼ V0 þ ðVm  V0 Þ ðPm =An Þð1=BnÞ ;

and V0 ¼ Vm ð0:6  0:006  Pm Þ;

1. Klotz S, Hay I, Dickstein ML, Yi GH, Wang J, Maurer MS, et al. Single-beat estimation of end-diastolic pressure-volume relationship: a novel method with potential for noninvasive application. Am J Physiol 2006;291: H403-12. 2. Ten Brinke EA, Burkhoff D, Klautz RJ, Tsch€ ope C, Schalij MJ, Bax JJ, et al. Single-beat estimation of the left ventricular end-diastolic pressure-volume relationship in patients with heart failure. Heart 2010;96:213-9.