Author + information
- Received February 23, 2015
- Revision received October 16, 2015
- Accepted October 21, 2015
- Published online June 1, 2016.
- Samuel J. Boynton, MDa,
- Jeffrey B. Geske, MDb,
- Angela Dispenzieri, MDc,
- Imran S. Syed, MDb,
- Theodore J. Hanson, MDa,
- Martha Grogan, MDb and
- Philip A. Araoz, MDa,∗ ()
- aDepartment of Radiology, Mayo Clinic Rochester, Rochester, Minnesota
- bDepartment of Medicine, Division of Cardiovascular Diseases, Mayo Clinic Rochester, Rochester, Minnesota
- cDepartment of Medicine, Division of Hematology, Mayo Clinic Rochester, Rochester, Minnesota
- ↵∗Reprint requests and correspondence:
Dr. Philip A. Araoz, Department of Radiology, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Objectives This study sought to determine the prognostic value of cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) in amyloid light chain (AL) cardiac amyloidosis.
Background Cardiac involvement is the major determinant of mortality in AL amyloidosis. CMR LGE is a marker of amyloid infiltration of the myocardium. The purpose of this study was to evaluate retrospectively the prognostic value of CMR LGE for determining all-cause mortality in AL amyloidosis and to compare the prognostic power with the biomarker stage.
Methods Seventy-six patients with histologically proven AL amyloidosis underwent CMR LGE imaging. LGE was categorized as global, focal patchy, or none. Global LGE was considered present if it was visualized on LGE images or if the myocardium nulled before the blood pool on a cine multiple inversion time (TI) sequence. CMR morphologic and functional evaluation, echocardiographic diastolic evaluation, and cardiac biomarker staging were also performed. Subjects’ charts were reviewed for all-cause mortality. Cox proportional hazards analysis was used to evaluate survival in univariate and multivariate analysis.
Results There were 40 deaths, and the median study follow-up period was 34.4 months. Global LGE was associated with all-cause mortality in univariate analysis (hazard ratio = 2.93; p < 0.001). In multivariate modeling with biomarker stage, global LGE remained prognostic (hazard ratio = 2.43; p = 0.01).
Conclusions Diffuse LGE provides incremental prognosis over cardiac biomarker stage in patients with AL cardiac amyloidosis.
Amyloid light chain (AL)–type amyloidosis is a rare systemic disease with an incidence of approximately 1 per 100,000 person-years (1). It is characterized by tissue deposition of insoluble fibrils made up of plasma cell–derived immunoglobulin light chain precursor proteins (2). Cardiac involvement, found in up to 60% of patients with AL type amyloidosis, is a major determinant of morbidity and mortality (3).
Cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) has been shown to be associated with myocardial amyloid deposition (4), and it has been associated with all-cause mortality (5–8). However, with the advent of staging systems using serum biomarkers as their basis (9), it has become important that CMR findings be shown to be incrementally prognostic for serum biomarkers. Therefore, the purpose of this study was to evaluate retrospectively the prognostic value of CMR LGE for determining all-cause mortality in AL amyloidosis and to evaluate the prognostic value with serum biomarker stage.
Institutional Review Board approval was obtained. Written informed consent was not required for retrospective medical review. This study is a follow-up investigation of a cohort of patients with amyloidosis originally described by Syed et al. (4), who compared CMR LGE patterns with clinical and echocardiographic findings in cardiac amyloidosis.
Inclusion criteria for this study were as follows: 1) CMR ordered for evaluation of cardiac amyloidosis between January 1, 2006 and December 31, 2007; 2) age ≥18 years; 3) histologically proven AL-type amyloidosis; 4) the presence of a monoclonal protein in the urine or serum or a monoclonal population of plasma cells in the bone marrow, or both; and 5) CMR performed within 3 months of diagnosis.
Exclusion criteria were as follows: 1) history of myocardial infarction or myocarditis; 2) previous peripheral blood stem cell transplant; or 3) history of previous heart transplant. Of 151 patients with documented amyloidosis who underwent CMR at our institution (Mayo Clinic Rochester, Rochester, Minnesota) during this time frame, 76 were included in this study.
CMR myocardial late gadolinium enhancement patterns
CMR image acquisition was performed as described by Syed et al. (4). Electrocardiogram (ECG)-gated CMR was performed with a 1.5-T system (Twin speed EXCITE, GE Healthcare, Waukesha, Wisconsin). LGE images covering the left ventricle in multiple short-axis and long-axis views were obtained between 7 and 12 min after an intravenous bolus of 0.2 mmol/kg gadodiamide (Omniscan, GE Healthcare, Princeton, New Jersey) with segmented inversion recovery fast gradient echo sequences with the following parameters: echo time, 1.6 ms; repetition time, 3.7 ms; flip angle, 20°; matrix, 256 × 160; and field of view, 320 mm. Multiple inversion time (TI) cine fast gradient echo sequences were obtained to determine the TI with maximal myocardial nulling for LGE images. Forty images were obtained on a single slice with varying TI, with the best time selected for LGE sequences.
LGE pattern was evaluated by 2 observers by consensus who categorized delayed enhancement into 1 of 3 categories:
1. “Global,” in which the inversion recovery images showed circumferential, diffuse LGE extending from the endocardium to the epicardium or where the myocardium was unable to be nulled adequately after multiple varying TI sequences or if, on visual inspection of the multiple TI cine fast gradient echo sequence, myocardial tissue crossed the null point (became black) before the blood pool;
2. “Focal patchy,” in which there were nondiffuse, discrete areas of LGE, including circumferential LGE confined to the endocardium;
3. “None,” in which there were no areas of LGE, and myocardial tissue did not cross the null point before the blood pool on the multiple TI cine fast gradient echo sequence (Figure 1).
CMR-derived morphology and functional data
Short-axis cine steady-state free precession images were obtained from the midatria to the ventricular apex. The sequence parameters for the steady-state free precession images were as follows: echo time, 1.7 ms; repetition time, 3.4 ms; flip angle, 45°; matrix, 256 × 192; and field of view, 320 to 440 mm, with phase field of view, 0.75 to 1.0, and 8-mm slice thickness with a 1-mm interslice gap.
Myocardial mass, left ventricular (LV) volume, and right ventricular volume were evaluated by manually tracing epicardial and endocardial borders on the short-axis steady-state free precession cine sequences on commercially available post-processing software (MASS Analysis 6+, Medis, Leiden, the Netherlands). Indexed values were obtained by dividing each value by the patient’s body surface area.
Additional measurements evaluated were LV myocardial thickness, right ventricular myocardial thickness, and the presence or absence of pericardial and pleural effusions. Myocardial thickness was measured on short-axis images during end diastole. Pericardial effusion and pleural effusion were documented as present or absent.
Serum cardiac biomarkers were used to stage patients with the Mayo staging system, developed by Dispenzieri et al. (9), which has cardiac troponin T (cTnT) levels and N-terminal probrain natriuretic peptide (NT-proBNP) levels as its basis. The staging system uses a cTnT threshold level ≥0.035 μg/l and an NT-proBNP threshold level ≥332 ng/l. If both biomarkers were below their respective threshold levels, patients were designated as stage I; if either cTnT or NT-proBNP levels were elevated, patients were designated as stage II, and if both cTnT and NT-proBNP levels were elevated, patients were designated as stage III. Serum biomarkers obtained within 3 months of CMR were used.
All patients received standard transthoracic echocardiograms as part of their work-up. Diastology data were obtained at the apical 4-chamber acoustic window using 2-dimensional and Doppler color flow techniques. All patients received a standard 12-lead resting ECG during their initial assessment. Measured height and weight were used to calculate the body mass index, defined as the weight divided by the body surface area. Subjective functional status was assigned according to the New York Heart Association functional classification criteria and was obtained from chart review.
The outcome variable for this study was death, from any cause. The medical records at our institution were reviewed for documentation of death. Documentation included a system notification of death, an electronic copy of a death certificate, a medical provider’s note of death, or documentation of correspondence by a relative that the patient had died. Patients were censored if they did not have a documented date of death, were documented to be alive at the end of the study, or were lost to follow-up, in which case their last clinic visit or correspondence to the institution was used, whichever came later. Abstraction of the medical records was done by a physician blinded to the interpretation of magnetic resonance imaging variables.
LGE images were available in all 76 subjects. A multiple TI cine sequence was not available for 1 patient, so 75 patients had evaluation of myocardial nulling compared with blood pool nulling. Continuous variable analysis of ventricular mass and volume was available for 75 patients because 1 patient’s study was not adequate for accurate measurements. Biochemical biomarker analysis was limited to the 69 subjects who had both cTnT and NT-proBNP data.
All data analysis was performed using commercially available software (JMP version 9, SAS Institute, Cary, North Carolina). Normally distributed data were presented as mean ± SD for continuous variables and counts or percentages for categorical variables; non-normally distributed data were given as median and interquartile range (IQR). The 2-tailed Student t test or 1-way analysis of variance was used when appropriate for continuous variables, and the chi-square or Fisher exact test was used for categorical variables as appropriate.
Kaplan-Meier curves were used to estimate overall survival. Differences between the curves were assessed using the log rank method, and p values of <0.05 were considered significant. Survival was calculated from the date of diagnosis. For univariate analysis, Cox proportional hazards analysis was used to evaluate survival. For the 3 LGE patterns (global, focal patchy, none), Cox proportional hazards analysis was used to determine whether there was any significance among any of the 3 patterns, and for this an overall p value for LGE was reported. A similar procedure was used for the 3 biochemical biomarker stages. For multivariate analysis, Cox proportional hazards analysis was used. Two multivariate models were created. In the first, the visual LGE pattern with the worst prognosis was combined with the biomarker stage. In a second, separate model, myocardial nulling before the blood pool was included with the biomarker stage.
Demographic and clinical features of the patient cohort are summarized in Table 1. Median study follow-up was 34.4 months. There were 40 deaths (53%), with survival probability at 1, 3, and 5 years equaling 60%, 53%, and 48% respectively (Figure 2).
Of the 76 patients, 32 patients (42%) had the global pattern of LGE; 24 patients (32%) had focal patchy LGE; 20 (26%) had none. A summary of patients’ characteristics by LGE pattern and myocardial nulling is presented in Table 2.
LGE pattern was prognostic for all-cause mortality, with p = 0.002 for detecting any difference among the 3 patterns. The global pattern was associated with an adverse prognosis (hazard ratio [HR]: 2.93; p < 0.001) and the none pattern was associated with an improved prognosis (HR: 0.37; p < 0.01). Survival curves divided by visual LGE pattern are shown in Figure 3. Log rank for detecting a difference among groups was p < 0.001.
Biochemical biomarker stage was prognostic, with p < 0.001 for detecting any difference among stages. Stage III was associated with decreased survival (HR: 3.12; p < 0.001), whereas stage I was associated with increased survival (HR: 0.19; p < 0.01). The overlap of visual LGE pattern with biomarker stage is shown in Figure 4. At multivariate analysis, global LGE remained associated with all-cause mortality when it was combined with biomarker stage (HR: 2.03; p = 0.05) (Table 3).
This study showed that global LGE was associated with a worse prognosis in univariate analysis and in multivariate analysis combined with biomarker stage in patients with AL amyloidosis.
The current investigation is the second study to show a CMR LGE parameter to be incrementally prognostic over biomarkers in cardiac amyloidosis. Fontana et al. (8) studied 119 patients with AL amyloidosis and 122 patients with transthyretin-related amyloidosis and found transmural LGE to be prognostic for all-cause mortality in univariate analysis (HR: 5.38; p < 0.0001) and in a multivariate model including NT-proBNP, stroke volume, LV ejection fraction, LV mass, and echocardiographic E/e′ ratio (HR: 4.13; p < 0.05). This present study confirms the findings of Fontana et al. (8) and supports the incremental prognostic power of CMR LGE.
The findings of this study also support several earlier studies showing that CMR LGE provides incremental prognostic power in patients with amyloidosis. White et al. (7), in a study combining 46 patients with cardiac amyloidosis with a larger cohort of patients, found that examination of myocardial nulling compared with the blood pool provided incremental prognostic data over LV ejection fraction, ECG low-voltage pattern, and LV mass (p = 0.0002). In 29 patients with AL amyloidosis, Migrino et al. (6) showed that LGE (defined as present or absent) added prognostic power over New York Heart Association functional class heart failure (p = 0.04). This study, along with those of Fontana et al. (8) and White et al. (7), suggest that conventional LGE images alone should not be used to detect diffuse amyloid infiltration of the myocardium.
LGE requires the user to select the TI that makes normal myocardium dark (nulled) and abnormal myocardium bright. If the myocardium is diffusely involved, without normal myocardium for reference, diffuse amyloid infiltration can have a variety of appearances depending on the severity of infiltration and the user’s choice of TI. If the myocardium has a concentration of contrast agent comparable to that of blood, resulting in a similar T1 value as blood, then the myocardium may appear diffusely enhancing with a bright blood pool or diffusely poorly nulled with a dark blood pool (Figure 1B), depending on the TI selected. If the myocardium has a shorter T1 than the blood pool, the myocardium may appear diffusely enhancing with a dark blood pool (Figure 1A); however, it is possible for a user to select a very short TI and create a normal-appearing LGE image with dark myocardium and a bright blood pool.
Thus it is significant that in the current investigation, in addition to inspecting the LGE images, we also defined global LGE as being present if the myocardium nulled before the blood pool on a multiple TI cine fast gradient echo sequence. Fontana et al. (8) used post-contrast T1 maps to select the optimal LGE images or else used a phase-sensitive inversion recovery sequence, which does not allow short T1 myocardium to appear dark compared with a longer T1 blood pool. White et al. (7) used only a cine TI sequence and found prognostic power in comparing myocardial nulling with blood pool without reference to LGE images.
Conflicting results of early studies of LGE prognosis in cardiac amyloid appear to be related to an inability to detect diffuse enhancement. In 1 of 2 studies with negative results, Maceira et al. (10) found that visual assessment of LGE (defined as “present” or “absent”) was not associated with survival, but in the same patients these investigators found that comparisons of blood pool and myocardial T1 values were associated with survival. Thus this “negative” study did support the proposition that abnormal contrast material accumulation in the myocardium is associated with decreased survival, but it suggested that visual assessment was not the preferred method. In the other study with negative results, Ruberg et al. (11) used a semiautomated technique in which LGE was defined as signal intensity 6 standard deviations higher than the myocardium of lowest signal intensity, a method that would not detect LGE if the entire myocardium was involved.
First, the number of events was small, even though this was one of the larger studies of LGE in AL amyloidosis. Although the event rate was high (53%), with 76 patients and only 40 deaths, the absolute number of deaths limits the number of variables for multivariate modeling. This situation is unfortunate given the large number of parameters available from univariate analysis (Table 1). Second, this study did not incorporate T1 mapping, which has been shown to be associated with all-cause mortality in univariate analysis (12). Unfortunately, the image acquisition (obtained in 2006 and 2007) did not allow for retrospective T1 mapping in these patients. Finally, this study did not consider patients’ treatment and response to treatment. In a tertiary referral institution, many patients undergo diagnostic evaluation but then return for treatment to their home institution. Although mortality data are available, data on response to treatment are incomplete and were therefore not included in this study.
This study showed that the LGE global pattern is associated with all-cause mortality in patients with AL cardiac amyloidosis in univariate analysis and provides incremental prognostic information over biomarker stage.
COMPETENCY IN MEDICAL KNOWLEDGE: Among patients with systemic AL amyloidosis, global LGE in the myocardium predicts all-cause mortality incrementally over biomarker stage.
TRANSLATIONAL OUTLOOK: Many variables are prognostic in patients with AL amyloidosis. Future studies will be needed to determine the optimal prognostic strategy in these patients.
This work was made possible by Clinical and Translational Science Awards (CTSA) Program grant UL1 TR000135 from the National Center for Advancing Translational Sciences (NCATS), a component of the National Institutes of Health (NIH). The contents of this paper are solely the responsibility of the authors and do not necessarily represent the official view of the NIH. All authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- amyloid light chain
- cardiac magnetic resonance
- cardiac troponin T
- hazard ratio
- late gadolinium enhancement
- left ventricular
- N-terminal probrain natriuretic peptide
- inversion time
- Received February 23, 2015.
- Revision received October 16, 2015.
- Accepted October 21, 2015.
- American College of Cardiology Foundation
- Kyle R.A.,
- Linos A.,
- Beard C.M.,
- et al.
- Syed I.S.,
- Glockner J.F.,
- Feng D.,
- et al.
- Austin B.A.,
- Tang H.,
- Rodriguez E.R.,
- et al.
- White J.A.,
- Kim H.W.,
- Shah D.,
- et al.
- Fontana M.,
- Pica S.,
- Reant P.,
- et al.
- Dispenzieri A.,
- Gertz M.A.,
- Kyle R.A.,
- et al.
- Banypersad S.M.,
- Fontana M.,
- Maestrini V.,
- et al.