What Is Myocardial Perfusion and How Is It Tested?

Myocardial perfusion is the flow of blood through the muscle tissue of the heart, delivering the oxygen and nutrients that keep cardiac cells alive and contracting. When doctors talk about “testing” myocardial perfusion, they mean using imaging or invasive tools to measure whether that blood supply is adequate, especially under physical or chemical stress. The concept sits at the center of how coronary artery disease is diagnosed and managed, but the testing landscape is broader and more varied than most people realize.

How the Heart Feeds Itself

Every organ needs blood, but the heart is uniquely demanding. It extracts a higher percentage of oxygen from its blood supply than almost any other tissue, which means it has very little reserve to draw on when demand spikes. To cope, the coronary arteries rely on a layered system of self-regulation. Multiple mechanisms work together: the arteries respond to changes in pressure, to chemical signals released by heart muscle cells when they need more fuel, to signals from the cells lining the vessel walls, and to input from the nervous system and circulating hormones. These mechanisms collectively adjust resistance in the coronary vessels so that blood flow matches what the heart muscle needs at any given moment.¹

Researchers describe this self-adjusting system as “autoregulation,” and its components do not contribute equally. Metabolic signals, which are chemical cues from the heart tissue itself saying “I need more oxygen,” appear to have the largest influence. The myogenic response, where the vessel wall reacts to pressure changes by constricting or relaxing, is the second most important. Shear-dependent mechanisms, driven by the friction of blood flowing along vessel walls, contribute the least to maintaining steady flow across different pressures.² This hierarchy matters clinically because diseases that disrupt metabolic signaling or damage vessel walls can undermine the heart’s ability to keep itself fed, even before a blockage becomes severe enough to show up on a standard angiogram.

What Happens When Perfusion Falls Short

When coronary blood flow cannot keep up with demand, the result is myocardial ischemia. Most people associate ischemia with chest pain, but the actual sequence of events starts silently. The earliest detectable change is a perfusion defect: a patch of heart muscle receiving less blood than its neighbors. Diastolic dysfunction follows, meaning the affected region has trouble relaxing between beats. Then comes systolic dysfunction, where contraction weakens. Only after that do the electrical changes appear on an ECG, and chest pain typically arrives last.³ This sequence, sometimes called the ischemic cascade, explains why perfusion imaging can catch problems that neither symptoms nor a resting ECG would reveal. The perfusion defect is the first domino, and imaging aims to spot it before the later, more dangerous dominoes fall.

The cause is not always a classic cholesterol plaque narrowing a large coronary artery. Microvascular dysfunction, where tiny vessels deep in the heart muscle malfunction, can produce the same mismatch between supply and demand. This distinction has become increasingly important, because patients with microvascular problems can have completely normal-looking large arteries on a catheterization yet still face real risks.

SPECT Imaging

Single-photon emission computed tomography, or SPECT, is the workhorse of myocardial perfusion testing and has been in clinical use for decades.⁴ In a SPECT scan, a small amount of radioactive tracer, usually a technetium-99m-labeled compound, is injected into a vein. The tracer travels through the bloodstream and is taken up by heart muscle cells in proportion to blood flow. A gamma camera then rotates around the chest, detecting the radiation emitted by the tracer and constructing a three-dimensional map of how the tracer distributed itself through the heart wall.

Two sets of images are typically acquired: one at rest and one during stress. Where blood flow is normal, the tracer lights up evenly. Where an artery is narrowed, the downstream segment receives less tracer during stress, showing up as a darker region on the image. If that same region fills in normally at rest, the defect is “reversible,” suggesting ischemia. If the dark patch persists at rest, it points to scarring from a prior heart attack.

The prognostic power of a SPECT scan is well established. In a study of over 1,700 patients, those with normal or equivocal scan results had an extremely low rate of hard cardiac events like death or heart attack, while those with abnormal results had a significantly higher event rate. SPECT added prognostic information beyond what clinical data and exercise testing alone could provide.⁵ Separately, in patients with high exercise tolerance who might seem low-risk on a treadmill test, perfusion imaging still identified those at substantially higher risk. Patients with abnormal perfusion scans had roughly eight times the risk of adverse cardiac events compared to those with normal scans, whereas the exercise ECG itself had no meaningful predictive power in this group.⁶

PET for Blood Flow Measurement

Positron emission tomography, or PET, works on a similar principle to SPECT but uses different tracers and detectors that give it several advantages. PET offers better diagnostic accuracy, lower radiation doses because the tracers decay faster, and a unique capability: it can measure myocardial blood flow in absolute terms, expressed as milliliters of blood per minute per gram of tissue.⁷ That absolute measurement opens up clinical territory that SPECT cannot easily reach. With SPECT, the scan compares regions of the heart to each other: one area looks darker than the rest. But if all three major coronary arteries are equally diseased, every region may look similarly dim and the scan can appear misleadingly normal. PET’s ability to quantify actual flow sidesteps this problem, making it better suited to catching multivessel disease.⁸

From a PET scan, clinicians can derive myocardial flow reserve, or MFR, which is the ratio of blood flow during stress to blood flow at rest. A normal MFR above 2.0 has a very high negative predictive value for ruling out high-risk obstructive coronary artery disease. A severely reduced MFR below 1.5 identifies patients at high risk for adverse events, even when no regional perfusion defect shows up on the images.⁹ This makes PET particularly useful for monitoring patients after heart transplants, where subtle diffuse disease in the graft’s vessels can be hard to detect by other means.

Cardiac MRI and CT Perfusion

Not every perfusion test involves a radioactive tracer. Stress cardiac magnetic resonance imaging, often called stress CMR, uses a gadolinium-based contrast agent injected during pharmacological stress. The MRI captures how the contrast washes through the heart wall in real time, with excellent spatial resolution. Because ischemia preferentially affects the inner layers of the heart muscle, CMR can identify perfusion defects within a single segment of the heart wall without needing to compare it to other regions.¹⁰ This fine detail makes it a powerful tool, especially when the question involves subtleties like whether ischemia is confined to the subendocardium.

CT perfusion is a newer approach that piggybacks on the technology of coronary CT angiography. Instead of just looking at anatomy, the scan acquires images during contrast passage to assess whether blood is actually reaching the heart muscle. In one study comparing CT perfusion and a computational method called CT-derived fractional flow reserve, CT perfusion showed sensitivity around 73% and specificity around 68% for detecting significant disease.¹¹ These numbers are respectable but not yet at the level of PET or stress CMR for standalone perfusion assessment. The real appeal of CT perfusion is the convenience of getting both anatomical and functional information in a single scan session.

Exercise Versus Drug-Induced Stress

Perfusion imaging almost always involves stressing the heart to widen the gap between healthy and diseased vessels. The preferred method, when the patient can manage it, is exercise on a treadmill or stationary bike. Exercise is physiological, familiar, and provides additional data like exercise capacity, heart rate response, and symptom reproduction. But many patients cannot exercise adequately because of arthritis, lung disease, peripheral vascular problems, or simple deconditioning.

For those patients, pharmacological stress agents fill in. Vasodilators like adenosine, dipyridamole, and regadenoson work by dilating coronary arteries. Healthy arteries open wide, increasing flow, while diseased arteries that are already maximally dilated at rest cannot respond much further. The result is a flow differential that lights up on the scan. Regadenoson has become the most commonly used vasodilator in the United States because it is given as a single quick injection rather than a continuous infusion.

However, vasodilators are not a perfect substitute for exercise. Preliminary findings in patients with hypertrophic cardiomyopathy suggest that regadenoson may overestimate perfusion defects compared to exercise, frequently producing impaired flow reserve values and side effects that do not reflect the patient’s daily symptoms.¹² In clinical practice, hybrid protocols are sometimes used: the patient exercises as much as they can, and if they fail to reach their target heart rate, a vasodilator is injected at peak exercise to supplement the stress.¹³

Invasive Perfusion Assessment in the Catheterization Lab

Sometimes the question about perfusion needs to be answered during a cardiac catheterization, with a wire already threaded into the coronary artery. Fractional flow reserve, or FFR, is the established invasive measurement. A pressure-sensing wire is advanced past a narrowing while a vasodilator is given to maximize blood flow. The ratio of pressure downstream of the blockage to pressure upstream tells clinicians whether that specific narrowing is limiting blood flow enough to warrant a stent or bypass. An alternative index called the instantaneous wave-free ratio, or iFR, measures the same concept during a specific window of the cardiac cycle when resistance is naturally stable, without needing a drug to be infused. Head-to-head comparisons have found that the two methods have similar diagnostic accuracy for guiding stent placement decisions.¹⁴

Beyond the large arteries, the index of microcirculatory resistance, or IMR, has emerged as a way to evaluate the tiny vessels that are invisible on an angiogram. IMR provides a quantitative assessment of microvascular function and has become important for understanding cases where patients have chest pain and abnormal perfusion imaging but no visible blockages in their main coronary arteries.

Artifacts and Diagnostic Pitfalls

Perfusion imaging is powerful, but it is not infallible. In SPECT scanning, soft tissue attenuation is the most common source of artifacts. Tissues of varying density surrounding the heart, including bone, the diaphragm, lungs, and breast tissue, absorb some of the radiation before it reaches the camera, creating dark patches that can mimic real perfusion defects.¹⁵ These artifacts depend on body size, sex, heart position, and whether the patient is lying down or upright. Misinterpreting an attenuation artifact as a true perfusion defect can send a patient down a path of unnecessary additional testing, including invasive catheterization. Modern SPECT systems use CT-based attenuation correction to reduce this problem, but recognizing artifact patterns remains a core skill for anyone reading these scans.

Patient motion is another common issue. Even small shifts during a scan can blur the images and create apparent defects. PET and CMR are less susceptible to some of these artifacts because of their higher resolution and shorter acquisition times, but each modality has its own set of pitfalls. CMR can be affected by irregular heart rhythms that disrupt the timing of image acquisition, and CT perfusion can be limited by beam-hardening artifacts from calcified plaques.

The Caffeine Question

If you have ever been told to skip your morning coffee before a stress test, this is why: caffeine blocks the receptors that adenosine and similar vasodilators act on. In theory, caffeine intake before a vasodilator stress test could blunt the drug’s ability to dilate the coronary arteries, masking a real perfusion defect. Standard practice is to abstain from caffeine for 12 to 24 hours before an adenosine-based test.¹⁶

The real-world picture is more nuanced than that strict rule suggests. In studies using cardiac MRI, caffeine intake reduced the apparent amount of ischemia slightly, but the differences were small, on the order of one segment in a 16-segment model, and no patient who had clinically significant ischemia was reclassified as normal because of caffeine.¹⁷ Still, the safest approach is to follow your testing center’s instructions. Canceling and rescheduling a test because of a forgotten latte is disruptive, but it is less harmful than getting a potentially unreliable result that delays diagnosis. Regadenoson, the most commonly used vasodilator in the U.S., was specifically developed to be less sensitive to caffeine interference than adenosine, though most labs still advise avoidance to be safe.

When Perfusion Results Predict the Future

A perfusion scan does more than answer the binary question of “is there a blockage?” It stratifies risk. A large study examining outcomes across different risk categories found that ischemia on a stress perfusion scan was significantly associated with death or heart attack in intermediate- and high-risk patients over a median follow-up period of about four years. In the high-risk group, patients with ischemia had roughly 50% higher odds of a major event. In low-risk patients, however, the association between scan-detected ischemia and outcomes did not reach statistical significance.¹⁸ This finding matters for who should get tested in the first place: perfusion imaging adds the most value for people at intermediate or higher baseline risk, not for low-risk individuals where the test is less likely to change management.

PET-derived flow measurements add another layer. A globally normal myocardial flow reserve identifies patients at lower clinical risk regardless of what the visual images show, while a severely reduced flow reserve flags high risk even when no regional defect is visible.⁹ These quantitative flow measurements have increasingly been used to monitor responses to lifestyle changes and medical therapy over time, giving clinicians a way to see whether interventions are actually improving blood delivery to the heart.

AI-Assisted Interpretation

Reading perfusion images is inherently subjective, and experienced readers sometimes disagree on borderline cases. Artificial intelligence is starting to change that. Deep learning models trained on large databases of perfusion scans can now generate automated scores that modulate and improve the traditional quantitative analysis. In one study, combining AI-generated predictions with conventional automated scoring lifted the diagnostic accuracy for obstructive coronary artery disease substantially: the area under the curve for the AI-enhanced score was 0.84, compared to 0.74 for the traditional automated score. Sensitivity improved from about 72% to 79%, and specificity from about 62% to 70%.¹⁹

The applications extend beyond SPECT. In dynamic CT perfusion, deep learning models have been developed to automatically segment the heart and identify which coronary artery territory has a perfusion defect, achieving accuracy above 0.94 area under the curve for each major coronary territory.²⁰ These tools are not replacing human readers yet, but they are increasingly used as a second set of eyes, flagging cases that a reader might underestimate and smoothing out variability between institutions.

Perfusion Testing Beyond Blocked Arteries

Although coronary artery disease drives most referrals for perfusion testing, the technique has found a growing role in conditions that have nothing to do with cholesterol plaques. Microvascular dysfunction is present in a strikingly large proportion of patients with various heart muscle diseases: roughly half to two-thirds of those with dilated cardiomyopathy, 40 to 80% of those with hypertrophic cardiomyopathy, and over 95% of those with cardiac amyloidosis.²¹ In each of these conditions, impaired perfusion contributes to disease progression and worse outcomes, and detecting it can inform treatment decisions.

Hypertrophic cardiomyopathy is a particularly instructive example. Even carriers of a genetic mutation linked to the disease who have not yet developed significant thickening of their heart walls can already show perfusion abnormalities. One study using CMR perfusion mapping found visible defects in about 20% of mutation carriers versus none of the healthy controls, and globally reduced perfusion reserve across the group.²² Conditions like sarcoidosis, takotsubo cardiomyopathy, and amyloidosis each produce recognizable perfusion patterns on PET that can help clinicians reach the right diagnosis faster when typical imaging clues are ambiguous.²³

How the Technology Got Here

The speed at which perfusion imaging has advanced is remarkable. The field essentially started with crude planar images using potassium-43 and thallium-201 in the 1970s and 1980s, graduated to SPECT, and then moved to hybrid systems that combine SPECT or PET with CT for anatomical context. Meanwhile, the roster of available tracers expanded from thallium and early technetium agents to rubidium-82 and nitrogen-13 ammonia for PET, with newer fluorine-18 labeled perfusion tracers under active development.²⁴ Fluorine-18 tracers are appealing because they can be produced at a central facility and shipped to hospitals that do not have their own on-site cyclotron, which could make PET perfusion imaging more accessible outside major academic centers. Alongside nuclear techniques, CMR and CT perfusion have matured as radiation-free or low-radiation alternatives, and the integration of AI analysis promises to make all of these modalities more consistent and more accurate across different hospitals and readers.