Despite much progress over recent decades improving outcomes in patients with cardiovascular (CV) disease, it remains the leading cause of death in the US and worldwide.1 A major manifestation of CV disease is the clinical syndrome of heart failure (HF) that has become a global pandemic affecting more than 64 million people.2 An aging population, increasing cardiometabolic risk factors such as diabetes, obesity and chronic kidney disease that predispose to HF, and improved treatments that extend survival in patients with ventricular dysfunction are all expected to increase the medical and financial burden of the condition.

Establishment of comprehensive pharmacologic therapies, for many years diuretics, vasodilators, angiotensin-converting enzyme inhibitors (ACE-I), angiotensin II receptor blockers (ARB), beta blockers and mineralocorticoid receptor antagonists (MRAs), and more recently angiotensin receptor–neprilysin inhibitors (ARNi) and sodium glucose cotransporter-2 inhibitors (SGLT2i), are improving outcomes in people with HF.3 However, a key component to current HF management is the use of cardiac implantable electronic devices (CIED). These include an implantable cardioverter defibrillator (ICD), a left ventricular assist device (LVAD), and pacemaker cardiac resynchronization therapy (CRT). There are extensive societal guidelines describing when these advanced therapies should be undertaken, emphasizing the need for skillful balancing of potential clinical improvement versus device-induced morbidities, assuring that CIED therapies are directed to those likely to benefit.

Various clinical criteria and diagnostic testing modalities are used to guide initiation of CIED therapy. A key criterion has been cardiac systolic dysfunction measured by the LV ejection fraction (LVEF). However, the limitations of the somewhat crude LVEF are well known with large numbers of patients with heart failure with reduced EF (HFrEF) not benefiting from devices.4 In addition, it has been increasingly understood that many patients with HF have only a mildly reduced LVEF (HFmrEF) or preserved LVEF (HFpEF), with evidence that some of them may also benefit from devices. While more precise measurements of LV function, such as speckle tracking echocardiography to assess global longitudinal strain, may help guide the use of devices, techniques that more directly assess the underlying pathophysiologic and molecular processes of HF promise to offer better guidance.

Cardiac Radionuclide Imaging to Guide CIED Therapy

Radionuclide tracer techniques have long played an important role in assessing and guiding management of patients with HF. Equilibrium radionuclide angiography (ERNA) assessment of LV systolic and diastolic function has been available since the 1970s5 and continues to play a role in guiding ICD use. Phase analysis of ERNA to assess ventricular mechanical dyssynchrony was described in the 1980s 6 and has shown potential ability to help guide CRT.

Abnormalities of cardiac autonomic innervation are a major pathophysiologic component of HF. There is abundant literature on adrenergic cardiac radionuclide imaging for characterization and risk stratification of HF patients, as well as the use of the technique for guiding implantation of CIEDs.7 An important study in this regard is the international multicenter observational prospective ADMIRE-HF (AdreView Myocardial Imaging for Risk Evaluation in Heart Failure) trial8 in which 961 HFrEF patients underwent autonomic imaging with the sympathetic SPECT tracer iodine-123 metaiodobenzylguanidine (123I-mIBG). Satisfactory myocardial tracer uptake, as assessed by a heart to mediastinum ratio (HMR) ≥ 1.6, predicted an extremely low 17-month event rate of 3% for cardiac death versus 16.1% for a reduced HMR, and a 3.5% rate of potentially lethal ventricular arrhythmias versus 10.4% for a reduced HMR. The predictive power of 123I-mIBG imaging was independent of standard HF risk stratification parameters such as LVEF and B-type natriuretic peptide (BNP). Noteworthy is that for the 20% of patients in ADMIRE-HF who had a satisfactory HMR, many of whom did not have an ICD at the beginning of the study, there were only 2 cardiac deaths, thus a < 1% 17-month death rate. In a subsequent analysis extending ADMIRE-HF follow-up to 2 years (ADMIRE-HFX), patients with an HMR ≥ 1.8 had no fatal or potentially fatal ventricular arrhythmic events, thus unlikely to benefit from an ICD.4

Numerous other studies have supported the strength of adrenergic imaging to identify HF patients who meet guidelines criteria for an ICD but who may be unlikely to benefit. Among these is the cardiac PET Prediction of Arrhythmic Events with Positron Emission Tomography (PAREPET) study9 in which subjects with LVEF ≤ 35% underwent 11C-HED adrenergic imaging. The amount of denervated myocardium predicted the 4.1-year sudden cardiac arrest rate while LVEF or BNP did not. Importantly, PARAPET highlighted improving arrhythmic predictive value by combining radionuclide imaging with other criteria, showing that by adding LV end-diastolic volume index, creatinine, and the patient not taking an ACE-I identified subjects with a < 1% arrhythmic event rate.

Disappointingly, despite a wealth of data, the use of cardiac adrenergic imaging to guide ICD use has not received societal guidelines-approved recommendations. Unfortunately, a crucial prospective 123I-mIBG randomized trial effort was unsuccessful due to insufficient patient recruitment. In addition, high tracer cost relative to reimbursement coverage has precluded the use of such imaging in situations for which ICD use is uncertain that could provide clinical experience and accumulation of practical data to encourage and guide perspective studies more likely to succeed. There is ongoing development of practical adrenergic PET agents, such as 18F-Fluorbenguan and 18F meta-fluorobenzylguanidine, that can provide better cardiac molecular and physiologic detail for superior arrhythmic risk stratification, but much work remains.10

There have also been investigations using radionuclide adrenergic imaging to guide LVAD use, in particular hel** to identify patients who have received the device but could potentially be weaned off it. Very small (< 30 patient) studies have reported that HMR and tracer washout rates improve in HF patients with an LVAD and that those with better improvement are more likely to show reverse remodeling and may be candidates for device explantation rather than undergoing cardiac transplant or requiring the device permanently.7 More studies and data are needed in this regard.

Cardiac Radionuclide Imaging to Guide Cardiac Resynchronization Therapy (CRT)

In patients with HFrEF, a major contributor to disabling symptoms, adverse ventricular remodeling, exacerbation of functional mitral regurgitation, and the threat of lethal ventricular arrhythmias is myocardial mechanical dyssynchrony, mostly systolic but also in-part diastolic. While guidelines directed pharmacologic therapy can help ameliorate the dyssynchrony through unloading and sometimes reverse remodeling, for patients with advanced HF who meet various criteria, CRT pacing is recommended. By providing more synchronous electrical activation of the LV, or of the LV and RV together, numerous clinical trials have demonstrated that CRT can improve quality of life, increase LVEF, and decrease death and HF hospitalizations.3 While mechanical dyssynchrony is the pathophysiologic problem, CRT indications are based on various electrocardiographic (ECG) findings of electrical dyssynchrony, mostly left bundle branch block with QRS duration ≥ 150 ms, although sometimes patients with non-LBBB who have a QRS duration ≥ 150 ms, or lesser degrees of QRS widening with disabling symptoms or recurrent hospitalizations despite optimal medical therapy, may benefit.

However, 20 to 40% of patients who meet guideline indications do not respond to CRT.11 Potential reasons for clinically unsuccessful CRT include extensive myocardial scar blocking pacer stimulation of myocardium and malposition of the LV lead away from the late contracting myocardial segment.12 There have also been concerns that ECG electrical dyssynchrony may not always reliably reflect mechanical dyssynchrony. Several echocardiographic studies have explored whether mechanical dyssynchrony could better guide CRT use. In the “Results of the Predictor of Response to CRT” (PROSPECT) trial13, various echocardiographic parameters of mechanical dyssynchrony, including tissue Doppler-based methods, had only modest accuracy in predicting clinical response to CRT. However, a study limitation was high inter- and intraobserver variability in investigator determination of many of the echocardiographic dyssynchrony parameters. In a large clinical trial by the EchoCRT Study Group14 of patients with narrow QRS (< 130 ms) with echocardiographic evidence of LV mechanical dyssynchrony, there was no benefit from CRT, and evidence of increased mortality. Based largely on these echocardiographic studies, the consensus has been that mechanical dyssynchrony parameters should not be used over electrical dyssynchrony parameters for guidance of CRT.

In contrast to echocardiographic studies, studies investigating radionuclide assessment of mechanical dyssynchrony using phase analysis of ECG-gated SPECT imaging have frequently shown ability to predict response to CRT.11 In patients with HFrEF who underwent pacemaker implantation for CRT, Boogers et al.15 reported that ECG-gated SPECT mechanical dyssynchrony parameters, specifically a histogram bandwidth (HBW) > 72.5° and a phase standard deviation (PSD) > 19.6° had a > 80% accuracy for predicting CRT response. A study by Friehling et al.16 of HF patients who underwent ICD and biventricular pacemaker implantation found that a combination of the presence of ECG-gated SPECT mechanical dyssynchrony above a pre-determined threshold, myocardial scar < 40%, and pacer lead placement in a region of latest activation (or an adjacent segment) had a 96% positive predictive value for improved mechanical synchrony with pacing. Importantly, pacer-induced deterioration of ECG-gated SPECT dyssynchrony predicted an increased 7.5-month (median) occurrence of HF hospitalization, ICD discharge, or death.

Numerous studies have reported that cardiac adrenergic imaging with 123I-mIBG can also predict response to CRT. Among these is a report by D’Orio Nishioka et al.17 in which for HFrEF patients with a QRS duration > 120 ms, an HMR > 1.36 had a > 70% sensitivity and specificity for predicting symptomatic and LV improvement with CRT, with HMR independent of QRS width, LVEF, or LV size. The authors proposed that cardiac sympathetic innervation abnormalities reflect myocardial damage too severe to allow functional improvement with CRT. A later study by Tanaka et al.18 showed that adding echocardiographic evidence of mechanical dyssynchrony improved the ability of a higher HMR to predict LV improvement and 3-year survival free from HF progression. However, it is noteworthy that the study by D’Orio Nishioka et al.,17 and another by Cha et al.19 observed that patients who responded to CRT were more likely to have a non-ischemic cardiomyopathy.

Moving forward

Given the promising but generally under-recognized potential ability of radionuclide imaging assessment of mechanical dyssynchrony and cardiac autonomic innervation to help guide implementation of CRT, in the manuscript by Mishkina et al.20 in the current issue of the journal, the authors undertake a systematic analysis of the matter using a solid-state cadmium-zinc telluride (CZT) camera that promises better image quality and detail, with focused attention on potential differences for cardiomyopathies of presumed ischemic versus non-ischemic origin. The study included 58 HFrEF patients, 26 with ischemic heart failure (IHF), and 32 with non-ischemic heart failure (NIHF) who had guidelines indications for CRT. The authors reported that while, for IHF, a higher mechanical dyssynchrony HBW, but not the 123I-mIBG HMR, significantly and independently predicted a favorable CRT response (≥ 15% decrease in LVESV and/or ≥ 5% increase in EF within 12 months), for NIHF , a higher 123I-mIBG HMR (via derivation of a “pseudoplanar” HMR from CZT acquired tomographic images), but less mechanical dyssynchrony, significantly and independently predicted a favorable CRT response.

While the study results and conclusions support previous literature findings that there are differences between IHF and NIHF regarding the meaning of 123I-mIBG image findings and the ability of mechanical dyssynchrony to predict response to CRT,21,22 the small number of patients (all four subgroups had 20 or fewer subjects) is problematic. For example, while, for IHF patients, higher 123I-mIBG cardiac uptake did not achieve statistical significance for predicting a favorable CRT response, supplementary Table 1 shows a trend towards doing so, with an early HMR of 1.77 in responders versus 1.58 in non-responders (p = 0.15), and a late HMR of 1.53 in responders versus 1.49 in non-responders (p = 0.74). Thus, with more subjects, statistical significance might have been achieved. Regarding mechanical dyssynchrony in IHF patients, while a higher HBW predicted a favorable CRT response, the PSD paradoxically went in the opposite direction with a lower PSD predicting a favorable CRT response (60° vs 65°, p = 0.3). For NIHF patients, study results indicate that responders paradoxically had significantly lower PSD and HBW medians, i.e., less mechanical dyssynchrony, which contradicts an understanding of how CRT works. As the authors state, larger studies are needed to confirm and expand on their results.

As a key aspect of the study was differentiating IHF from NIHF, the matter needs to be considered further. The authors defined IHF by results of coronary angiography (as per cited reference #21, an epicardial stenosis > 50%) within 3 months prior to CRT or if LV dysfunction was associated with prior revascularization or myocardial infarction. However, as is currently increasingly appreciated, patients without coronary artery disease by conventional angiographic epicardial stenotic criteria frequently have ischemia and LV dysfunction from non-obstructive epicardial and/or microvascular disease. At the same time, patients with LV dysfunction conditions rightly classified as NIHF, such as hypertrophic cardiomyopathy, cardiac sarcoidosis, or even valvular heart disease, often have a myocardial ischemic contribution to ventricular dysfunction.

To be able to effectively use radionuclide imaging to predict the utility of implanting various CIEDs, it will be necessary to better understand and characterize the histopathologic tissue damage present in the various forms and combinations of ischemic and non-ischemic cardiomyopathies, and how imaging can properly assess these.22 These would include assessment of myocyte damage, distortion and dysfunction, replacement and interstitial fibrosis, changes in extracellular matrices, ongoing insults and injury such as ischemia and inflammation, and regional and global myocardial geometric abnormalities resulting from these processes. Radionuclide imaging has shown potential to image and quantify all these processes.

An important contention of the manuscript is that superior technical aspects of radionuclide imaging, such as the use of a CZT camera that showed good reproducibility of HMR values, or potentially PET, are necessary to obtain reliably accurate images and measurements. Continued improvements in imaging hardware and software, as well as further development of radiotracers that image underlying molecular and structural disease processes, will be crucial.

Final Thoughts

Small, observational studies such as the one presented here are important hypothesis generating investigations. However, these need to encourage subsequent undertaking of well-designed prospective studies that are likely to succeed. Only then will radionuclide imaging techniques be accepted as methods with unique ability to direct use of cardiac devices in patients with heart failure to improve life quality and longevity.

Disclosures

The author declares no conflict of interest.