Identifying Cardiac Amyloid in Aortic Stenosis

Objectives The purpose of this study was to validate computed tomography measured ECV (ECVCT) as part of routine evaluation for the detection of cardiac amyloid in patients with aortic stenosis (AS)-amyloid. Background AS-amyloid affects 1 in 7 elderly patients referred for transcatheter aortic valve replacement (TAVR). Bone scintigraphy with exclusion of a plasma cell dyscrasia can diagnose transthyretin-related cardiac amyloid noninvasively, for which novel treatments are emerging. Amyloid interstitial expansion increases the myocardial extracellular volume (ECV). Methods Patients with severe AS underwent bone scintigraphy (Perugini grade 0, negative; Perugini grades 1 to 3, increasingly positive) and routine TAVR evaluation CT imaging with ECVCT using 3- and 5-min post-contrast acquisitions. Twenty non-AS control patients also had ECVCT performed using the 5-min post-contrast acquisition. Results A total of 109 patients (43% male; mean age 86 ± 5 years) with severe AS and 20 control subjects were recruited. Sixteen (15%) had AS-amyloid on bone scintigraphy (grade 1, n = 5; grade 2, n = 11). ECVCT was 32 ± 3%, 34 ± 4%, and 43 ± 6% in Perugini grades 0, 1, and 2, respectively (p < 0.001 for trend) with control subjects lower than lone AS (28 ± 2%; p < 0.001). ECVCT accuracy for AS-amyloid detection versus lone AS was 0.87 (0.95 for 99mTc-3,3-diphosphono-1,2-propanodicarboxylic acid Perugini grade 2 only), outperforming conventional electrocardiogram and echocardiography parameters. One composite parameter, the voltage/mass ratio, had utility (similar AUC of 0.87 for any cardiac amyloid detection), although in one-third of patients, this could not be calculated due to bundle branch block or ventricular paced rhythm. Conclusions ECVCT during routine CT TAVR evaluation can reliably detect AS-amyloid, and the measured ECVCT tracks the degree of infiltration. Another measure of interstitial expansion, the voltage/mass ratio, also performed well.

A ortic stenosis (AS) is the most common valve disease in the developed world (1). Its prevalence increases with age, with 2.8% to 4.8% of patients $75 years of age having at least moderate AS (2,3). Once symptomatic with severe AS, outcomes are poor without intervention (4), which can be either surgical or transcatheter aortic valve replacement (TAVR). TAVR numbers are increasing fast worldwide, in response to both an aging population and technological developments (5,6).
Another disease of aging is wild-type transthyretin-related cardiac amyloidosis (ATTR-CA); deposits are present within the myocardium at autopsy in up to 25% of patients $85 years of age (7). Recent work has shown a remarkably high prevalence (14% to 16%) of ATTR-CA in the elderly AS population being considered for TAVR (ASamyloid) (8,9). We do not yet fully understand the significance of this dual pathology, either for valve intervention or the role for specific amyloid therapies such as tafamidis (10), patisiran (11), and inotersen (12), but detection is likely to be important. Conventional first-line investigations for ATTR-CA, such as echocardiography, blood biomarkers, or electrocardiogram (ECG), are confounded by the dual pathology. ATTR-CA can now be diagnosed noninvasively by using bone scintigraphy, such as 99m Tc-3,3diphosphono-1,2-propanodicarboxylic acid (DPD), 99m Tc-pyrophosphate, and 99m Tc-hydroxymethylene diphosphonate, coupled with a negative search for a plasma cell dyscrasia (13). Although availability and awareness are increasing, it requires an extra test in elderly, often frail, patients.
As part of routine TAVR evaluation, patients typically undergo contrast computed tomography (CT) imaging to assess annulus dimensions, coronary artery height (and patency, where possible), and vascular access. Contrast CT imaging can also be used to measure the myocardial extracellular volume (ECV) in a manner similar to cardiovascular magnetic resonance (CMR) (14,15). The ECV increases moderately with diffuse fibrosis but massively with amyloidosis (16). Our group has previously validated ECV quantification by CT imaging (ECV CT ) against CMR and histology (endomyocardial biopsy) in severe AS (17,18) and against CMR in cardiac amyloid (18). Unlike recommended CMR acquisition, the ECV CT acquisition for cardiac amyloid can be performed earlier at 5 min rather than 10 min post-contrast (18).
In the current study, we hypothesized that ECV CT as part of routine TAVR evaluation CT imaging would be able to detect AS-amyloid. To improve workflow, we also sought to optimize the scanning protocol in terms of dose and timing (shortened scan delay).  These control patients were included to provide an estimate of "normal" ECV CT and were not used in the screening calculations.
ELECTROCARDIOGRAM. As we have described previously (19), Sokolow-Lyon criteria were calculated as the sum of the amplitude of the S-wave in lead V 1 and the R-wave in lead V 5 or V 6 (whichever was greater) (20). The voltage/mass ratio was defined as the Sokolow-Lyon total divided by the indexed left ventricular (LV) mass on echocardiography. Patients with bundle branch block or a ventricular paced rhythm were excluded from this analysis (21). Low limb lead voltages were defined as all limb leads with an amplitude #0.5 mV.
ECHOCARDIOGRAPHY. AS severity (aortic valve peak velocity, mean gradient, and valve area), biventricular systolic and left ventricular diastolic function were assessed using transthoracic echocardiography (22)(23)(24)(25)(26). As we have described previously (19), LV ejection fraction was calculated using Simpson's biplane if possible (otherwise visually) and the indexed stroke volume was calculated using the LV outflow tract velocity time integral and diameter, which was then indexed to body surface area. Relative wall thickness was defined as: (2 Â posterior wall diameter)/(LV internal diameter at end-diastole) (25). where IVSd is the interventricular septal diameter, LVIDd is the LV internal dimension at end-diastole, and PWd is the posterior wall diameter.
Relative  ECV ANALYSIS. We have briefly described this technique previously (29). Nonrigid registration software (Hepacare, Siemens Healthineers) allowed averaging and aligning of the axial shuttle mode datasets to improve image quality and reduce noise. The averaged baseline image was then subtracted from the averaged 3-and 5-min post-contrast images (providing a partition coefficient) and then registered with the CTCA image. A region of interest was placed in the LV blood pool on the CTCA image and the hematocrit (usually taken on the same day) inputted, generating a myocardial ECV CT map via the formula: DHU is the change in Hounsfield unit attenuation precontrast and post-contrast (i.e., HU post-contrast À HU pre-contrast ) (18,30,31). This information was loaded into prototype software (Cardiac Function, Siemens Healthineers), which allowed the ECV CT map to be superimposed on the CTCA image, the myocardial contours to be edited, and the results to be displayed as a 17-segment polar map (Figures 1 and 2). When calculating total ECV CT , focally elevated ECV CT (e.g., likely myocardial infarction) were not excluded, but American Heart Association segments with significant beam-hardening artifacts from adjacent pacing wires    In AS-amyloid, parameters reflecting LV thickness and mass were higher, whereas the MCF was lower.

Lone AS Grade 1 AS-Amyloid Grade 2 AS-Amyloid Control
Global longitudinal strain was impaired in both ASamyloid and lone AS but did not differ. Both hs-TnT and N-terminal pro-B-type natriuretic peptide levels were higher in AS-amyloid ( Table 1). ECV CT findings. ECV CT was feasible for measurement in all patients for whom data were obtained. ECV CT was 32 AE 3%, 34 AE 4%, and 43 AE 6% in those patients with Perugini grades 0, 1, and 2, respectively, using a 3-  as predictors of AS-amyloid (     Table 1. nonacademic and pressingly so, with the availability of 3 novel, potential, but costly medical therapies for cardiac amyloidosis (10)(11)(12) that have yet to be validated in patients with AS-amyloid. Clearly, an individualized treatment strategy is going to be needed, and answers will hopefully prove more forthcoming with the increasing availability of bone scintigraphy that will enable increased diagnostic rates and research activity.
The fact that pre-existing RBBB is associated with cardiac amyloidosis is intriguing and may prove relevant in the TAVR cohort given that we know RBBB is  associated with a higher likelihood of post-TAVR pacemaker implantation (38) and worse outcomes (38,39). Although the authors did not investigate for the presence of concomitant cardiac amyloidosis, it is possible that the presence of RBBB at baseline might be an ominous sign that deserves further investigation.
We propose CT imaging as a technique to increase AS-amyloid detection and present a diagnostic algorithm ( Figure 5). Because ECV CT is easy to implement, and the patient is already in the CT scanner, we think adoption of this technique could be high. This algo- We propose different thresholds for onward referral depending on how important grade 1 versus 2 is discovered to be, and whether specificity or sensitivity becomes the priority. A lower threshold of 29% using a 3-min post-contrast acquisition would never miss a case (sensitivity 100%) but would probably result in an unacceptably high referral rate for bone scintigraphy (specificity 19%). A threshold of 31.4% would have a sensitivity of 94% and not miss DPD grade 2 cases but would miss a proportion of DPD grade 1 cases (1 of 5 in our cohort); however, the trade-off is that fewer cases would be referred for an unnecessary DPD (specificity 48%).
Technological developments often result in new insights into established techniques. We were not surprised to find that AS-amyloid was hard to detect based on ECG (e.g., small voltages) or echocardiographic (e.g., reduced MCF) changes because both AS and amyloid can have widely different influences on heart muscle. RBBB being associated with AS-amyloid is interesting and may prove important given that we know it is both common in patients with TAVR and is associated with worse outcome (including higher likelihood of post-TAVR pacemaker insertion) (38).
Another interesting finding is that a combination parameter of both ECG and echocardiography, the voltage/mass ratio, performed exceptionally well for amyloid detection compared with parameters derived from just one technique. This is perhaps not surprising as ECV CT  although prevalence and other clinical information informs, this is not the primary focus of this paper.
Inline ECV CT software is not yet available, and the work presented here will need to be optimized for integration into the daily CT workflow. Although global longitudinal strain data were included in this study, unfortunately we did not have regional longitudinal strain data available at the time of submission, which may have proven additive in identifying cardiac amyloidosis. The relatively small number of patients with AS-amyloid in this study may also have affected our results.

CONCLUSIONS
Lone AS results in detectable increases in ECV CT compared with control subjects. ECV CT using a lowdose protocol, with a 3-min post-contrast acquisition, can detect AS-amyloid and grade its severity in the TAVR population, and it could be used as a screening tool in those patients already undergoing a clinically indicated CT scan.