Apical to Base Gradient of Technetium-99m Pyrophosphate Myocardial Counts in Cardiac Amyloidosis

An Insight Into the Mechanism of Myocardial Strain Gradient, or Merely “Clouds That Mimic Land Before the Sailor’s Eye?”

Corresponding Author

• Download figure
• Open in new tab
• Download powerpoint

• advanced cardiac imaging
• heart failure with preserved ejection fraction
• infiltrative cardiomyopathies
• nuclear cardiac imaging

There are 2 main types of amyloidosis that affect the heart—light chain (AL) amyloidosis, a hematologic disorder, and transthyretin (TTR) amyloidosis, in which the amyloid is derived either from wild-type TTR or from an amyloidogenic TTR with a point mutation. Untreated, AL has a far worse prognosis than TTR, characterized by rapidly progressive heart failure and death. Patients with TTR cardiac amyloidosis are often concerned about their prognosis; many have read the grim prognosis of untreated AL cardiac amyloidosis and either have incorrectly extrapolated this to their disease or have been misinformed by a physician who is unaware of the distinctly different prognoses (1,2). Even the cardiac amyloid expert needs to be cautious about prognostication, as survival is calculated differently in the limited published survival data—some studies count from the first cardiac symptom, and others from disease diagnosis. These may be very different time points in a disease in which diagnosis is often delayed. It would, therefore, be useful to have an objective measure by which prognosis could be assessed.

Technetium-99m pyrophosphate (TcPYP) scanning (or technetium-99m DPD, technetium-99m-3,3-diphosphono-1,2 propanodicarboxylic acid, in Europe) is increasingly used as a diagnostic test in patients with suspected amyloid cardiomyopathy. Pyrophosphate (PYP) was originally developed as an agent to image acutely injured necrotic myocardium (3), and was more recently repurposed for imaging TTR cardiac amyloidosis (4). Amyloid deposits have 3 major structural components: the precursor protein, heparin sulfate proteoglycan, and a calcium-dependent amyloid P-component that binds the fibrils together. The precise structural component of amyloid that PYP binds to in the heart in patients with cardiac amyloidosis is not known (Table 1); a calcium-dependent mechanism is accepted widely (5), but not universally (6–8). Although the precise mechanism of PYP uptake in myocardial amyloid deposits is not known, a strong body of evidence now supports its role in diagnosis of TTR cardiac amyloidosis.

A large, multicenter study has shown that a strongly positive scan, defined as cardiac uptake of isotope greater than bone uptake, is highly sensitive and specific for TTR amyloidosis, particularly in the absence of a monoclonal protein in the serum (9). A positive scan can diagnose TTR amyloidosis and distinguish this condition from AL cardiac amyloidosis, and TcPYP imaging is rapidly being adopted in the United States as a diagnostic test for the evaluation of suspected TTR cardiac amyloidosis, with guidelines for imaging already published (10). Amyloid cardiomyopathy, whether AL or TTR amyloidosis, has other unusual imaging features, namely a typical pattern of echocardiographically-visualized longitudinal strain. This abnormality, first described in a preliminary communication by our group (11) and later detailed by Phelan et al. (12), is characterized by well-preserved apical strain of the left ventricle (LV) and a gradient from apex to base, with the greatest impairment in strain in the basal segments. It has also been suggested that cardiac magnetic resonance (CMR) may, in some cases, show an apparent gradient of late gadolinium myocardial enhancement that preferentially affects the base (13,14), but this has not been confirmed in large studies.

The authors of the current study hypothesized that “given the data from echocardiographic and CMR studies, there would be regional variation in TcPYP uptake” (15). A total of 54 patients with TTR amyloidosis and a positive TcPYP scan were evaluated. The results showed not only that many of these patients had an apical to basal gradient of myocardial counts, but also that the absence of this gradient augured a worse prognosis. They therefore argue for the use of single-photon emission computed tomography (SPECT) imaging PYP not only as a diagnostic tool, but also as a pointer toward prognosis.

The question arises as to whether these data are sufficient to expand the use of PYP imaging from a diagnostic to prognostic tool, and if so, to what extent is it accurate? The data, although interesting, should be viewed with caution as the series is small, is from a single center, and relies on a single dichotomous value for determining outcome. The small number of subjects also limits detailed multivariable survival analyses. Furthermore, the finding of a lack of prognostic significance for apex-to-base longitudinal strain ratio, a functional parameter, may be related to its collinearity with apex-to-base PYP count ratio. Alternatively, it may be that the 2 parameters are unrelated phenomena, as discussed in the following text. In our opinion, therefore, clinical use of PYP to estimate survival in an individual patient presenting with TTR cardiac amyloidosis, although previously reported (16), should be considered premature; at least 1 prospective evaluation is needed not only to validate the data, but also to compare its prognostic value against recently published biomarker-based stratification of survival in this condition.

On reflection, perhaps the most interesting question arising from this study is not whether the validity of the results will hold up prospectively, but whether the SPECT-PYP gradient that is described has the same or a similar mechanism to the longitudinal strain gradient seen by echocardiography or is an unrelated, coincidental finding. The very existence of this unusual gradient seen by both SPECT and strain echocardiography suggests that a common mechanism may exist. Yet, longitudinal strain is primarily a function of the contraction of myocardial fibers oriented in the longitudinal direction, and is therefore a physiological phenomenon, whereas regional LV radiotracer counts are assumed to be related to amyloid burden, which is a pathoanatomical phenomenon. If regional strain differences were simply a function of different relative amounts of amyloid deposition, then the degree of strain impairment should be similar to SPECT-PYP gradient. However, not only did the authors fail to show such a correlation, but earlier work from their institution evaluating the prognostic value of the strain gradient by echocardiography showed the opposite finding, namely that the presence of a gradient on echocardiography in cardiac amyloidosis was associated with a worse prognosis than its absence (17).

Let us take the question further. If one assumes that regional TcPYP uptake is proportional to regional amyloid deposition, then why should there be a gradient of PYP uptake, with the least uptake at the apex? The most obvious answer is that there is a preferential deposition of amyloid at the base of the heart, something implied by the authors but not explicitly stated. It is plausible that loss of apical sparing of LV longitudinal contraction may represent more advanced cardiac amyloidosis, and would explain the worse prognosis. Small CMR studies have shown a greater degree of delayed gadolinium enhancement at the base than the apex, but among 3 explanted hearts that had also been studied during life by CMR (13), the heart with the greatest proportion of apical to basal gadolinium uptake on CMR did not demonstrate any corresponding difference in amyloid deposition. Perhaps a simpler explanation may relate to the anatomy of the LV. Physiologically, LV regional myocardial wall thickness varies in each of the 17 segments; the apical segments are thinner than the basal segments (18) (10.3 mm vs. 7.3 mm), with the true apex measuring 2.3 mm (18). The apical mass is further reduced by the smaller amount of tissue enclosing the smaller cavity size at the ventricular apex compared with the base. A uniform infiltration of the myocardium by amyloid would result in lower TcPYP counts at the apex, due to partial volume averaging from imaging the apex, a structure that is below the resolution of the SPECT imaging system. Although more quantitative than visual inspection of planar images, SPECT/PYP still has limited resolution at the apex. Our conclusion, therefore, is that it is more likely that the SPECT gradient on TcPYP scanning of TTR amyloid, although mimicking the pathophysiological gradient in longitudinal strain, is unrelated to it.

Whether or not the gradient in TcPYP myocardial counts is a function of uniform distribution of amyloid in segments with greater or lesser myocardial mass, the mechanism of apical sparing on strain echocardiography is also not clear. Although this has also been hypothesized to represent greater amyloid deposition at the base of the heart that the apex, the same objections to this hypothetical suggestion apply as for the gradient in counts. An alternative explanation, worthy of exploration, relates to the complexity of fiber orientation in the heart and the effect of wall stress on regional longitudinal strain (19,20). Wall stress is a function of wall thickness and cavity size, and regional wall stress differences exist in the LV, with several authors suggesting that wall stress is lowest at the apex. It is also apparent that changes in wall stress affect myocardial strain, and that the changes in myocardial architecture and geometry that occur in diseased ventricles result in changes not only in global strain (21), but also in relative regional LV strain that correspond, in part, to changes in LV wall stress (20). This is a topic that has not been studied intensely, and not at all in patients with amyloid heart disease. It is very possible that the changing geometry of the LV in cardiac amyloidosis might differentially affect LV strain, a process that could be further affected by the regional differences in LV fiber orientation (22). Remarkably, despite centuries of cardiac dissection, anatomic cardiac fiber orientation remains a topic of debate (23,24), and the interaction of pathological changes in the myocardium and the extracellular space have yet to be explored. Such exploration, through anatomy, imaging, or advanced computer modeling, might shed light on the effect that amyloid deposition has on regional myocardial function.

The authors refer to the apical-sparing uptake of TcPYP in amyloid cardiomyopathy as “mimicking” the pattern of impaired echocardiographic strain in basal and mid-segments. To mimic is defined as “to resemble or imitate” or “to bear a deceptive resemblance” (25). After thinking carefully about their data, we are indeed left with the question as to whether the nuclear images mimic the longitudinal strain abnormalities of echocardiography, or are truly related via a common pathological phenomenon. Although an apparently minor manifestation of an uncommon disease, these regional differences open vistas that encompass human anatomy and physiology in a much more expansive form. Although we favor mimicry, (captured so beautifully in William Wordsworth’s poetic phrase “clouds that mimic land before the sailor’s eye” [26]), it is only through broader studies of the anatomy and physiology of regional ventricular function in health and disease that the solution will be found.

• 2017 American College of Cardiology Foundation