Author + information
Section Editor: Christopher M. Kramer, MD
THE PARADOX OF MOLECULAR IMAGING is one that is on the minds of many in the imaging community today. The promise of agents that would identify local biological processes such as inflammation, remodeling, angiogenesis, and metabolism is what attracts imaging researchers into this exciting realm. This excitement is tempered by the reality that there is a long and steep uphill climb to clinical application of any of these imaging approaches. This conundrum is highlighted in this iForum piece in iJACC by Drs. Sinusas, Thomas, and Mills.
Dr. Sinusas highlights the promise of single-photon emission computed tomography (SPECT) and positron emission tomography (PET) approaches to imaging of metabolism and neuroreceptors, given their high sensitivity and lower threshold to clinical applicability as compared with ultrasound and magnetic resonance imaging approaches. He also reviews the state of molecular imaging of atherosclerosis and angiogenesis. He notes that hybrid imaging of multiple modalities offers significant promise for these types of application. He points out that there are significant barriers to clinical translation for many of these modalities, including cost as well as radiation (in the case of the nuclear techniques and CT).
The piece by Drs. Thomas and Mills presents more of the day-to-day clinical reality in regard to molecular imaging. Conventional imaging as performed in 2011 has already revolutionized the way diagnoses are made in clinical cardiovascular medicine. Thus, the bar is set very high for advances in molecular diagnostics. In addition, the regulatory (Federal Drug Administration [FDA]) and reimbursement barriers may be so high as to prevent the ultimate clinical application of any of these novel developments. This is an important topic for debate, and we are glad to bring it to our readers.
Molecular Imaging: In Utilization, on the Horizon, or in the Distant Future
the concept of molecular imaging originated within the oncology community and has become part of routine clinical care of patients with cancer, but still remains primarily on the immediate horizon for the cardiovascular community. Molecular imaging is defined as the noninvasive visualization, characterization, and measurement of biological processes at the molecular and cellular level in humans and other living systems. Molecular imaging already plays a critical role for individually tailoring pharmacological and cell- or genetic-based therapeutic interventions for cancer, and has become an integral part of clinical trials (1). Therefore, the molecular imaging approach goes beyond providing only diagnostic and prognostic information based on the early identification of the molecular events associated with physiological or pathological processes. The successful application of a molecular imaging strategy requires: widespread education of the medical and lay communities, the availability of appropriate imaging probes with sufficient sensitivity and specificity, and imaging instruments that enable the visualization and quantification of these probes with adequate spatial resolution and accuracy. This interrogation of the molecular processes must be performed in combination with evaluation of anatomic or physiological changes. This generally requires application of multimodality imaging with fusion of highly sensitive molecular-targeted approaches with high-resolution anatomic approaches. To be clinically applicable, these targeted imaging approaches need to complement or provide incremental value over existing imaging approaches or measurement of circulating biomarkers. The goal is to provide a more global view of a disease process in balance, rather than focusing on isolated molecular or cellular events. The topic of cardiovascular molecular imaging has been previously reviewed in great detail in a recent review (2). This paper will try to provide a prospective on how cardiovascular molecular imaging could play a central role in day-to-day management of patients with cardiovascular disease.
Molecular imaging: in current cardiovascular practice and on the horizon
Two obvious examples of the application of molecular imaging in clinical cardiovascular medicine include metabolic and neuroreceptor imaging.
The primary emphasis of metabolic imaging in clinical practice has been in the study of myocardial intermediary metabolism, using either PET or SPECT imaging.
Several different radiolabeling strategies have been employed for clinical PET imaging of metabolism. The most widely used PET imaging approach is to radiolabel a substrate analog with fluorine-18 (18F). The best example in this category is 2-[18F] fluoro-2-deoxy-D-glucose (FDG). Evaluation of the myocardial kinetics of FDG provides an in vivo approach for imaging glucose uptake by the myocardium, but can also track myocardial or vascular inflammation. Several 18F-labeled PET agents have been proposed for evaluation of fatty acid metabolism that would have the potential for widespread metabolic imaging of the heart. These agents have already entered into clinical trials. The evaluation of different patterns of myocardial glucose and fatty acid metabolism with these 18F-labeled agents may have important diagnostic, prognostic, and therapeutic implications for management of patients with cardiovascular disease.
Alternatively, naturally occurring metabolic substrates can be radiolabeled in specific carbon locations with carbon-11, such as various fatty acids (1-11C-palmitate or 1-11C-acetate), glucose (1-11C-glucose), and lactate (l-3-11C-lactate). An advantage of this approach is that the metabolism of the radiolabeled substrate is identical to the unlabeled substrate. With the application of appropriate mathematical modeling schemes, the myocardial uptake and downstream metabolism of these substrates can be assessed in vivo. Disadvantages of this approach relate to the limitation of radiolabeling with 11C, including the requirement for an onsite cyclotron and advanced radiochemistry capabilities. Both of these issues have significantly limited the widespread clinical utilization of the 11C PET imaging approaches.
SPECT radiolabeled tracers have also been utilized to assess myocardial glucose and fatty acid metabolism, and offer practical clinical advantages. One of the first successful SPECT approaches for evaluation of oxidative fatty acid metabolism involved the use of 15-(p-iodophenyl)-pentadecanoic acid (IPPA), which contains an aromatic ring at the omega position radiolabeled with radioiodine. Unfortunately, conventional SPECT imaging systems did not have the sensitivity or temporal resolution to effectively evaluate the rapid kinetics of IPPA, and this approach was never widely applied. This limitation may disappear with the recent introduction of solid-state multidetector SPECT systems that provide high sensitivity and the capability for dynamic “PET-like” imaging.
An alternative solution to the challenges associated with rapid clearance kinetics of IPPA was the development of branched-chain analogs of IPPA, such as 123I-beta-methyl-p-iodophenylpentadecanoic acid (BMIPP). Alkyl branching inhibited beta-oxidation, thereby increasing radiotracer retention and improving SPECT image quality. Kontos et al. (3) recently reported the results of a large multicenter clinical trial that applied BMIPP SPECT imaging for the detection of acute myocardial ischemia in patients presenting to the emergency department with chest pain. This molecular-based metabolic imaging approach provided comparable sensitivity to other imaging approaches, while providing incremental value in the early detection of acute coronary syndrome and a retained performance even after resolution of symptoms. Thus, metabolic imaging may offer unique advantages over more conventional imaging of regional myocardial function or perfusion. The existing technical limitations of SPECT image quantification may be overcome with the availability of hybrid SPECT/CT imaging systems that facilitate correction of attenuation, scatter, and partial volume errors, allowing for determination of absolute radiotracer retention. Metabolic imaging is likely to play an important role in the evaluation and management of ischemic myocardial injury, hypertrophic heart disease, and heart diseases associated with diabetes and obesity, and complement conventional imaging of perfusion and function.
Another classic targeted molecular imaging approach that has entered multicenter clinical trials involves the imaging of cardiac neuroreceptors in the heart. This work has primarily focused on in vivo imaging of sympathetic function in the myocardium. Important alterations in pre- and post-synaptic cardiac sympathetic function occur in several cardiovascular diseases, including ischemic heart disease and heart failure, and may be predictive of risk for sudden cardiac death or death from heart failure, as well as for response to therapeutic interventions. Pre-synaptic function can be measured using radiolabeled norepinephrine analogs such as 11C-meta-hydroxyephedrine (11C-HED), a PET radiotracer, or 123I-meta-iodobenzylguanidine (123I-MIBG), a SPECT radiotracer. Post-synaptic function can be assessed with 11C-CGP12177, a radiolabeled beta-blocker for PET imaging.
Many clinical studies have demonstrated that 123I-MIBG SPECT imaging provides powerful diagnostic and prognostic information in patients with heart failure. In these patients with heart failure, 123I-MIBG scans typically show a reduced heart–mediastinum uptake ratio, heterogeneous distribution within the myocardium, and increased 123I-MIBG washout from the heart. A large, prospective, industry-sponsored trial, the ADMIRE-HF (AdreView Myocardial Imaging for Risk Evaluation in Heart Failure) study of 123I-MIBG imaging for risk stratification of patients with symptomatic heart failure on contemporary therapy was recently published (4), and demonstrated a highly significant relationship between the time to heart failure–related events and the heart-to-mediastinal ratio, which was independent of other commonly measured parameters such as left ventricular ejection fraction (LVEF) and B-type natriuretic peptide (BNP). This clinical study also showed a clear association between severity of myocardial sympathetic neuronal dysfunction and risk for subsequent cardiac death (4).
If the value of cardiac autonomic assessment using 123I-MIBG or 11C-HED imaging in combination with other conventional or molecular imaging indexes is confirmed, this neuroreceptor imaging approach may help in the selection of patients who would benefit the most from an implantable cardioverter-defibrillator by means of identification of those at increased risk for potentially fatal arrhythmias, leading to more cost-effective implementation of this life-saving device.
Molecular imaging in the future with expanded clinical impact
Molecular imaging is moving forward in many other areas relevant to the cardiovascular system. Some of these approaches involve targeted imaging of critical biological processes associated with cardiovascular disease, including inflammation, thrombosis, angiogenesis, apoptosis, necrosis, fibrosis, atherosclerosis, and remodeling. Figure 1 summarizes the complex cascade of events associated with the progression from early to advanced atherosclerosis, ischemic injury, and subsequent post-infarction remodeling, along with some of the potential molecular targets for imaging these molecular processes in temporal relation to conventional imaging approaches that focus on evaluation of physiology and changes in anatomic structure. Many probes targeted at these processes remain under development, although some have already reached clinical testing, or have even completed clinical testing and have received FDA approval. Several examples of novel molecular imaging approaches that have entered clinical testing are briefly outlined below.
Imaging of Atherosclerosis
Atherosclerosis is a chronic inflammatory disease, and several molecular approaches targeted at imaging vascular inflammation have been proposed for detection of unstable atherosclerotic plaque. Many clinical studies have evaluated the potential of 18F-FDG PET imaging of inflammation as an index for detection of unstable plaque (5). Other investigators have targeted the proliferation of vascular smooth muscle cells associated with vascular remodeling. Investigators have also targeted components of the clotting cascade because plaque rupture or erosion can lead to activation of circulating platelets and clotting cascade proteins.
Imaging of Angiogenesis
Angiogenesis is a complex process that involves many cell types and molecular signals. Potential targets for imaging of the angiogenic process include targeting endothelial cell markers associated with proliferation and/or migration, inflammatory cell markers, and markers related to alterations of the extracellular matrix. Many laboratories, including my own, have been focused on imaging of integrin activation as a way to track in vivo the angiogenic response to ischemic injury or stimulated angiogenesis. We have also developed and validated semiautomated quantitative approaches for accurately estimating in vivo radiotracer retention in order to track changes in the angiogenic response in relationship to changes in tissue perfusion. The recent imaging of αvβ3 integrin by the PET imaging tracer 18F-Galakto-RGD demonstrates the feasibility of detecting angiogenesis in the myocardium in humans. This molecular imaging approach may provide a more sensitive means of evaluating angiogenesis and optimizing therapy.
Imaging Myocardial Infarction and Ventricular Remodeling
Evaluation of myocardial infarction and post-infarct remodeling with imaging has traditionally been focused on evaluation of regional and global ventricular function and changes in global geometry. Specifically targeting the molecular events associated with this complex process may offer substantial advantages over conventional approaches that evaluate the gross pathophysiological changes that occur late in the process.
We have demonstrated that in vivo targeted SPECT imaging of activation of matrix metalloproteinases (MMPs) is feasible, and can provide new understanding regarding the role that MMPs play in post-infarct remodeling (6). The application of targeted MMP SPECT imaging for evaluation of remodeling requires absolute radiotracer quantification, which is now possible with hybrid imaging systems. The rennin-angiotensin system is also locally activated during post-infarct remodeling and contributes to the progression to heart failure. In evaluation of the rennin-angiotensin system, a number of angiotensin-converting enzyme inhibitors and angiotensin 1 antagonists have been radiolabeled for molecular imaging of the heart. These pre-clinical studies suggest that the imaging of the signaling events that take place within an infarct may be utilized to track critical molecular processes associated with remodeling. To gain a better understanding of post-infarct remodeling, the molecular images will need to be related to changes in regional mechanics as assessed by conventional imaging approaches. Much more work is still needed to assess the feasibility of applying molecular imaging in clinical trials for evaluation of post-infarction remodeling.
Barriers to clinical translation of molecular imaging
The lower sensitivity of ultrasound or magnetic resonance–based molecular imaging approaches limits their clinical translation. SPECT and PET imaging approaches provide high sensitivity, relatively low cost, and a minimal potential for adverse biological effects, and therefore provide the quickest means for translation of molecular imaging into patient care. However, the limited resolution of nuclear imaging requires anatomic colocalization of nuclear images with higher-resolution anatomic x-ray CT images. The application of hybrid SPECT-CT and PET-CT imaging systems will clearly improve the quantification of nuclear-based molecular imaging approaches, although the advantages of hybrid imaging must be weighed against the potential additive exposure to ionizing radiation associated with the additional CT imaging. This issue of radiation exposure is a growing concern within both the medical and lay communities.
The utilization of molecular imaging should expand with appropriate education of the cardiovascular community and the increased availability of various hybrid imaging systems (SPECT-CT, PET-CT, PET–magnetic resonance imaging) that will facilitate quantification of molecular imaging agents. However, the current hybrid imaging systems have not been optimized for cardiac applications, and do not provide adequate correction for cardiac and respiratory motion require for absolute quantification of targeted imaging agents.
In Defense of Structure, Function, and Perfusion: The Case Against Molecular Imaging
“The future… will be exactly like the past, only far more expensive”
—John Sladek (7)
molecular imaging. the very words conjure up visions of “Bones” McCoy with his medical Tricorder scanning a patient (human or alien) and divining all manner of diagnoses—past, present, and future. And how could one not see the future of medicine in the promise of molecular imaging? Integrin-targeted paramagnetic nanoparticles identify areas of vascular injury inapparent in standard magnetic resonance angiograms. P-selectin targeted ultrasound microbubbles identify areas of myocardium that have suffered as little as 5 min of ischemia hours earlier (8). Radioactive tracers targeted to MMP hold promise to identify vulnerable plaques. But before we get too carried away with these future wonders of imaging, we should pause to ponder a few questions. How many of these techniques are available today for general clinical use? How likely is it that a pharmaceutical company will invest the hundreds of millions of dollars to push these interesting but investigational imaging tools through an uncertain FDA process to confirm efficacy and safety needed for approval and marketing? And, finally, just how critical is the need for these investigational agents in current cardiology practice? In other words, how limited are we in current patient care in cardiology with our current imaging methodologies? When it comes to addressing the critical questions in clinical cardiovascular medicine—structure, function, and perfusion—current techniques do an outstanding—and improving—job. Indeed, it is far more likely that better diagnoses will result from the steady evolution of our conventional imaging modalities, ultrasound, nuclear cardiology, and cardiovascular computed tomography (CCT), and cardiac magnetic resonance imaging (CMR), not from the promised revolutionary changes in molecular imaging.
Despite all its cellular and molecular intricacies, the assessment of cardiovascular health involves few concepts outside the realm of a plumber or auto mechanic: the heart should be anatomically accurate in its construction; it should accept blood at low pressure and pump it out at high pressure through valves that neither leak nor obstruct; and it requires a steady delivery of fuel at rest and, more importantly, with exercise. In all of these arenas, contemporary cardiovascular imaging does an outstanding job.
Consider cardiovascular anatomy. Simple 2-dimensional echocardiography, available almost anywhere in the world at a cost under $8,000, can quickly and reliably diagnose congenital abnormalities, ventricular dilation, and aortic aneurysms. More comprehensive assessment is possible with 3-dimensional echocardiography, now available in single-beat acquisitions of 90° × 90° involving the processing of over 160,000,000 voxels per second. Even more detailed anatomy can be discerned with CCT and CMR, which provide isotropic high-resolution 3-dimensional imaging while avoiding the pitfalls of poor imaging windows that often plague echocardiography (9).
Perhaps the most common task in cardiovascular imaging is to assess regional and global ventricular function. Despite all the known pitfalls of the ejection fraction, it is a simple concept that immediately conveys useful information to clinicians in any setting. Simple eyeball assessment of an echocardiogram can quickly characterize left ventricular function as normal or mild, moderate, or severe dysfunction. A normal resting echocardiogram during chest pain virtually rules out acute ischemia as the cause of that chest pain. Beyond this qualitative assessment, new quantitative methods are available for echocardiography. By tracking the “speckles” generated by interference patterns in the myocardium, it is possible to derive regional strain, perhaps the purest measure of ventricular contraction. By combining data from the 3 standard apical views, one can generate a bullet plot that allows one to appreciate regional strain patterns at a glance. Critical parameters of stroke volume and cardiac output can be measured either by the difference in end-diastolic and end-systolic volumes by 2-dimensional or 3-dimensional echo or by use of the continuity equation with pulsed Doppler tracings or color Doppler from the left ventricular outflow tract. Similar measurements with even greater precision can be made with CMR and CCT. By placing a grid of magnetic tagging on the myocardium, it is possible to measure regional strain by CMR in a fashion analogous to echo speckle tracking (10). Equally important to systolic function is diastolic function—the ability of the heart to fill quickly and efficiently at low pressure—and here conventional modalities again offer superior assessment to anything proposed by molecular imaging. By carefully combining observations of transmitral and pulmonary venous flow, annular motion, and ventricular wall thickness and left atrial size, it is possible to discern the rate of left ventricular filling, estimate end-diastolic pressure, and characterize diastolic ventricular stiffness. Assessment of ventricular torsion provides a mechanistic link between systole and diastole (11). That these measurements can be made quickly and repeatedly explains the widespread integration of traditional cardiac imaging into heart failure management.
Coronary artery disease
One of the most common tasks of a cardiologist is to assess coronary artery disease: its detection and assessment of myocardial ischemia and viability. Despite all the promise of molecular imaging, the conventional tools we have today are superb for the task. Subclinical atherosclerosis can be detected by carotid ultrasound or magnetic resonance angiography; flow-limiting epicardial coronary disease can be assessed with high sensitivity and specificity by SPECT perfusion imaging or exercise echocardiography. CT angiography provides compelling evidence of coronary anatomy, obstructive lesions, and plaque characteristics (Fig. 2). Finally, coronary angiography provides the definitive roadmap to guide surgical intervention or provide access for coronary intervention. The technique is largely unchanged from the days of Mason Sones, yet nothing has displaced it after 50 years. But what do we do in the case of a chronic ischemic cardiomyopathy? How can we assess the viability of dysfunctional myocardium to guide the utility of revascularization? Fortunately traditional cardiovascular imaging brings an abundance of tools to this task. One of the most reliable signs of myocardial viability is evidence of contractile reserve. For this, inotropic stimulation (usually with dobutamine) can be used with echocardiographic or magnetic resonance imaging to provide evidence of increased wall thickening. If a territory improves at low-dose dobutamine, then degrades at high dose (the so-called biphasic response), this is a particularly strong indicator of the need to revascularize. More sensitive than contractile reserve is metabolic viability, evidence that the myocardium is still maintaining membrane integrity despite little or no mechanical contraction. Such membrane integrity can be shown metabolically by positron imaging for FDG uptake. More recently, delayed CMR imaging of gadolinium has been shown to be helpful in judging viability. Where there is gadolinium, there is scar, and scar will not improve with revascularization. With coronary occlusion, necrosis spreads from the endocardium to the epicardium; so the more transmural the late gadolinium enhancement, the less contractile enhancement with revascularization. Thus, for the vast majority of clinical questions in ischemic heart disease, contemporary conventional imaging has the answer.
The same story plays out again and again for all manner of cardiovascular diseases. For valvular heart disease, we need anatomy, quantitation of regurgitation and stenosis, and impact on the ventricle. Echocardiography is the mainstay for this, supplemented in some cases by CMR and CCT. Similarly, pericardial disease is easily screened for by echo, which can reliably diagnose tamponade and guide pericardiocentesis. CCT and CMR provide detailed anatomic and hemodynamic evidence of constriction. Aortic aneurysm and dissection; pulmonary embolus and hypertension; congenital heart disease; cardiac tumors, masses, thrombi, and other sources of emboli; thrombotic and atherosclerotic vascular disease—the list goes on and on. There simply are few disorders for which standard imaging cannot suffice.
Guidance of interventions
Finally, conventional cardiovascular imaging has proven invaluable for guidance of percutaneous and operative interventions. Intraoperative echo has long been used to guide valve repair and other procedures. Recently, echocardiographic guidance has been extended to catheterization and electrophysiology lab procedures, including percutaneous aortic valve replacement and mitral valve repair, closure of atrial and ventricular septal defects and paravalvular leaks (Fig. 3), and pulmonary vein isolation, among many others. Multimodality imaging is commonly used, such as when a preoperative CT scan is fused with live 2-dimensional and 3-dimensional transesophageal echocardiography to improve the safety margin for percutaneous aortic valve replacement.
Barriers to molecular imaging
Now, let us consider the challenges that developmental molecular imaging faces before it will move into a significant role in cardiovascular diagnosis. The vast majority of agents touted for molecular diagnosis are investigational and not currently and generally available for clinical use. Most of these new imaging techniques are only in the early investigational-development phase, and many are not even undergoing formal controlled clinical trials. And why wouldn't commercial imaging and pharmaceutical companies be actively pursuing all such promising agents? Consider the development costs and the prolonged review process any imaging agent must face to achieve regulatory approval. Based on publically available data, Adrian Nunn (12) has estimated that Schering and Amersham each spent about $150 million per year on research and development between 1999 and 2004 and yet did not get a single new agent or important new indication. Even a modest development program is likely to cost in excess of $200 million, with a very uncertain regulatory and reimbursement outcome.
“We [FDA] recommend that a medical imaging agent intended for the indication diagnostic or therapeutic patient management be able to improve patient management decisions… or improve patient outcomes…” (13). The requirement for patient management and outcomes improvement means the regulatory requirements for approval has been raised much higher than when thallium-201 and Tc99m sestamibi were approved. The FDA has made it perfectly clear that simply producing a prettier picture of some structure is not enough to gain approval. Indeed, the FDA has approved only a few new imaging agents in the past 10 years. And for those agents that do make it through the regulatory gauntlet, there is a very uncertain reimbursement reward. A general rule of thumb is that the peak yearly earnings of a diagnostic or therapeutic agent must roughly equal its overall development cost, thus covering the expense of the many agents that never receive approval and to allow investment in future products. With the $200+ million benchmark in mind, there is a clear tradeoff between market size and anticipated reimbursement. One of the characteristics of many molecular imaging agents is focused targeting on very specific clinical situations, which by necessity are relatively small markets. The smaller the market, the higher the required reimbursement (for 100,000 cases/year, reimbursement of $2,000 is required), which flies in the face of the prevailing climate of ever smaller reimbursement for imaging tests. Just last year, outpatient nuclear cardiology exams saw a threat of 36% to 42% cuts in Medicare reimbursement. Although these cuts will now be phased in over time, the long-term trend is clear. For a new molecular imaging agent to justify substantially higher costs than existing agents will require convincing evidence for improvement in patient care outcomes in controlled clinical trials, evidence that is both elusive and extraordinarily expensive to demonstrate.
The past 40 years have seen stunning improvements in the ability of noninvasive imaging to characterize cardiovascular structure, function, and perfusion. These strides have come from the progressive evolution of conventional imaging techniques, with relatively little impact from imaging targeted to specific molecular moieties. Although the basic science of molecular imaging continues to make impressive strides, the regulatory and commercial landscape is limiting to these investigational imaging agents. To continue our impressive progress in cardiac imaging, we will need to rely on continuous improvements in our existing techniques, while we await the magic of a molecular Tricorder.
Address correspondence: Dr. Christopher M. Kramer, University of Virginia Health System, Departments of Medicine and Radiology, PO Box 800170, Lee Street, Charlottesville, Virginia 22908-0001. E-mail:
- American College of Cardiology Foundation
- Kontos M.C.,
- Dilsizian V.,
- Weiland F.,
- et al.
- Jacobson A.F.,
- Senior R.,
- Cerqueira M.D.,
- et al.
- Rudd J.H.F.,
- Narula J.,
- Strauss H.W.,
- et al.
- Sahul Z.H.,
- Mukherjee R.,
- Song J.,
- et al.
- Sladek J.
- Kaufmann B.A.,
- Lewis C.,
- Xie A.,
- et al.
- Notomi Y.,
- Martin-Miklovic M.G.,
- Oryszak S.J.,
- et al.
- ↵Guidance for Industry: Developing Medical Imaging Drug and Biological Products Part 2: Clinical Indications. http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm071603.pdf. Accessed June 27, 2011.