Quantitative Myocardial Blood Flow with Cardiac PET

Quantitative myocardial blood flow (MBF) turns cardiac PET from a relative picture into an absolute measurement — perfusion in milliliters per minute per gram of tissue at rest and at stress, and the ratio between them, myocardial flow reserve. That absolute number is what lets PET uncover balanced multivessel disease and coronary microvascular dysfunction that relative perfusion images can miss, but it is only as trustworthy as the tracer physics, the dynamic acquisition, and the kinetic model behind it. ^{1, 2}

Introduction

For decades, myocardial perfusion imaging answered a relative question: does one part of the heart take up less tracer than the apparently normal part? That comparison detects flow-limiting coronary stenosis well, but it has a blind spot. When disease is diffuse — three vessels narrowed to a similar degree, or dysfunction of the small vessels the coronary angiogram cannot even see — every region can be reduced together, and the relative image looks deceptively uniform. ^{1, 3}

Cardiac PET closes that blind spot by measuring absolute flow. Because PET quantifies activity concentration accurately and acquires data fast enough to follow a tracer bolus through the blood and into the myocardium, it can fit a physiologic model and report rest and stress flow in absolute units. Dividing stress by rest yields myocardial flow reserve (MFR), also called coronary flow reserve (CFR) — a single number that integrates the effects of epicardial stenosis, diffuse atherosclerosis, and microvascular dysfunction. Large outcome studies show that a reduced flow reserve is a powerful, independent predictor of cardiac death, over and above the relative perfusion result. ^{3, 5}

This article explains how the measurement actually works: the tracer-kinetic principles, the dynamic acquisition, the differences among rubidium-82, N-13 ammonia, and O-15 water (and the newer F-18 flurpiridaz), the worked mathematics of extraction correction and flow reserve, the clinical thresholds, and the quality-control discipline the number demands. DRPS supports PET/CT and nuclear cardiology programs across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware with PET/CT and nuclear medicine physics support.

Topic Explanation

What quantitative MBF measures

Myocardial blood flow is the rate at which arterial blood is delivered to a gram of myocardium, in mL/min/g. A dynamic PET acquisition records how the tracer's concentration in the blood pool (the arterial input) and in the myocardial tissue evolve second by second after injection. A tracer-kinetic model relates those two time–activity curves to flow, and the fit returns MBF for the whole left ventricle and for regional territories. ^{1, 2}

Two numbers come out of a rest–stress study:

Rest MBF, typically on the order of ~0.6–1.1 mL/min/g in normal myocardium, though it must be interpreted against the patient's resting workload.
Stress MBF, the flow achieved under a vasodilator (regadenoson, adenosine, or dipyridamole) or exercise, normally rising several-fold.

Their ratio is myocardial flow reserve:

MFR = \frac{MBF _{stress}}{MBF _{rest}}

Because MFR is a ratio measured within the same patient, it is robust to many global scaling errors — but it inherits any error in either flow value, which is why both the rest and stress measurements must be controlled. ^{1, 2}

Why rest flow must be normalized to workload

Resting myocardial flow tracks the heart's workload, so a patient who is anxious, tachycardic, or hypertensive at rest has a genuinely elevated rest MBF — which would artificially depress the flow reserve. Rest flow is therefore commonly corrected to a reference workload using the rate–pressure product (RPP), the product of heart rate and systolic blood pressure:

RPP = HR \times SBP

MBF_{rest,corr} = MBF_{rest} \times \frac{RPP _{ref}}{RPP _{measured}}

with a reference RPP near 8,500–10,000 mmHg·min⁻¹. This correction is one of several documented in the joint position paper on clinical MBF quantification and should be applied consistently within a program. ¹

Key Technical Principles

The tracer-kinetic model

The workhorse is the one-tissue-compartment model, which treats the myocardium as a single compartment exchanging tracer with arterial blood. If $C_{a} (t)$ is the arterial (blood-pool) concentration and $C_{t} (t)$ is the tissue concentration, then

\frac{d C _{t} ( t )}{d t} = K_{1} C_{a} (t) - k_{2} C_{t} (t)

where $K_{1}$ (mL/min/g) is the uptake rate constant and $k_{2}$ (min⁻¹) is the washout rate constant. Fitting this equation to the measured tissue curve, using the measured arterial input, returns $K_{1}$ . Some tracers and workflows use a two-tissue-compartment formulation; the outcome study that established the prognostic value of flow reserve used a two-compartment kinetic model with factor analysis to extract the blood and tissue curves. ^{1, 3}

Extraction and the flow–uptake relationship

Here is the crucial physics: $K_{1}$ is not flow. It is the product of flow $F$ and the first-pass extraction fraction $E$ :

K_{1} = E \cdot F

For a diffusible tracer that must cross the capillary and cell membrane, the extraction fraction follows the Renkin–Crone relationship,

E = 1 - e^{- P S / F}

where $P S$ is the permeability–surface-area product of the tracer. The consequence is decisive: as flow $F$ rises, $E$ falls, so uptake $K_{1}$ increasingly underestimates flow. Consider an illustrative tracer with $P S = 1.5 mL/min/g$ :

At rest (F = 1.0) : E = 1 - e^{- 1.5/1.0} = 1 - e^{- 1.5} \approx 0.78

At stress (F = 3.5) : E = 1 - e^{- 1.5/3.5} = 1 - e^{- 0.43} \approx 0.35

Extraction has more than halved. If the software reported uptake without correcting for this roll-off, it would badly compress the true stress flow and understate the flow reserve. Every extracted flow tracer therefore carries a validated, tracer-specific extraction correction that inverts this relationship to recover $F$ from $K_{1}$ . This is why the flow number is only as good as the tracer model applied to it. ^{1, 4}

The three established flow tracers

The tracers differ in half-life, production, positron range (which sets image sharpness), and extraction behavior.

Property	Rubidium-82	N-13 ammonia	O-15 water
Physical half-life	~76 s	~9.97 min	~2 min (122 s)
Production	Sr-82/Rb-82 generator (no cyclotron)	On-site cyclotron	On-site cyclotron
Decay	~95% β⁺, ~5% electron capture	~100% β⁺	~100% β⁺
Positron range (RMS in water)	~2.6 mm (higher; softer images)	~0.57 mm (sharper)	Intermediate
Myocardial extraction	Flow-dependent; rolls off at high flow	High; modest roll-off	Freely diffusible; ~complete, flow-independent
Typical role	High-throughput clinical rest–stress MPI + flow	Flow + high-quality perfusion images	Physiologic reference standard for flow

Rubidium-82 dominates clinical practice because its generator removes the need for an on-site cyclotron and its 76-second half-life allows fast, repeated rest–stress imaging — at the cost of a high positron energy and a ~2.6 mm positron range that soften spatial resolution and produce noisier time–activity curves. N-13 ammonia gives sharper images and higher extraction but needs a cyclotron and a ~10-minute half-life that slows throughput. O-15 water is the physiologic gold standard for flow because it is freely diffusible and its extraction stays essentially complete across the flow range, but it produces no retained myocardial image, so it is used mostly in research and expert centers. ^{1, 2, 4}

A newer option, F-18 flurpiridaz, received U.S. approval in 2024. Its ~110-minute F-18 half-life allows unit-dose distribution without an on-site generator or cyclotron, a small positron range for sharp images, high extraction, and compatibility with exercise stress; a 2025 multi-society procedure standard now covers its use for perfusion imaging and blood-flow quantitation. ^{6, 8}

A worked flow-reserve example

Suppose a rest–stress rubidium-82 study, after extraction correction and rest-workload normalization, yields corrected rest MBF of 0.90 mL/min/g and stress MBF of 2.70 mL/min/g. Then

MFR = \frac{2.70}{0.90} = 3.0

A global MFR of 3.0 is comfortably normal. Now suppose a different patient shows visually uniform relative images but corrected rest MBF of 1.05 mL/min/g and stress MBF of 1.45 mL/min/g:

MFR = \frac{1.45}{1.05} \approx 1.4

Despite the "normal-looking" relative scan, an MFR near 1.4 signals a severely blunted vasodilator response — the physiologic fingerprint of balanced multivessel disease or diffuse microvascular dysfunction, and a high-risk finding that the relative image alone would have hidden. ^{1, 3}

Clinical Impact

Absolute flow changes what cardiac PET can detect and how it stratifies risk. In a landmark study of 2,783 patients, the lowest tertile of coronary flow reserve (below ~1.5) carried roughly a 5.6-fold increase in the risk of cardiac death compared with the highest tertile, and adding flow reserve improved risk reclassification beyond clinical variables and relative perfusion imaging. ³ The prognostic value extends to populations where relative imaging is especially limited: in dialysis-dependent end-stage renal disease, where diffuse microvascular disease is the rule, global CFR independently predicted all-cause and cardiovascular mortality and reclassified risk in a substantial fraction of patients. ⁵

Clinically, quantitative MBF and MFR help in several specific situations the joint guidance highlights: unmasking balanced ischemia in multivessel disease, diagnosing coronary microvascular dysfunction in patients with angina and non-obstructive coronaries (INOCA), gauging the physiologic significance of a known stenosis, and refining risk in diabetes and chronic kidney disease. The number does not replace the relative images or the clinical picture; it adds a physiologic axis that sharpens both diagnosis and prognosis. ^{1, 3, 5} For the relative-perfusion and quality-control foundation this builds on, see our guides to Rb-82 cardiac PET myocardial perfusion imaging and cardiac SPECT MPI quality control.

Practical Optimization Tips

Protect the dynamic acquisition

Start the scan before the injection so the arterial input peak is captured; a late or truncated dynamic series is the single most common way to corrupt the input function.
Use list-mode or a validated fast-frame dynamic protocol with short early frames (a few seconds) to resolve the sharp blood-pool bolus, lengthening later frames as tissue uptake dominates.
Standardize injection — activity, bolus profile, and flush — because the shape of the input bolus feeds directly into the kinetic fit.

Control the patient and the registration

Minimize and correct motion. The heart can shift between the blood-pool and uptake phases; even small motion misassigns blood and tissue activity. Review and, where available, apply motion correction.
Verify emission–CT alignment. Attenuation-map misregistration distorts the tissue curve and the flow value; check and correct the co-registration for both rest and stress. Our discussion of PET/CT attenuation correction explains why this matters quantitatively.
Account for residual activity. With longer-lived tracers, residual counts from the rest scan can contaminate the stress measurement; allow adequate decay or apply a correction.

Standardize the quantitative chain

Fix the tracer model. Use the extraction correction validated for your specific tracer and software, and do not compare flow numbers across tracers or software packages as if they were interchangeable.
Normalize rest flow to the rate–pressure product consistently, and record HR and SBP at rest.
Build quality control for the numbers, not just the pictures. A simplified stress/rest activity-ratio check has been validated as a routine quality-assurance cross-check on PET flow-reserve measurements and can flag a bad fit that still produces a plausible-looking image. ⁴

Common pitfalls to avoid

Reading MFR without the relative images — or vice versa. They answer different questions and are strongest together.
Treating uptake as flow. Uncorrected $K_{1}$ underestimates high flows; always apply the extraction model.
Ignoring hemodynamics. An elevated resting rate–pressure product depresses MFR unless rest flow is normalized.
Cross-comparing tracers and software. Thresholds are tracer- and platform-specific; validate your own.

Regulatory Considerations

Quantitative cardiac PET sits inside the same regulatory framework as any PET radiopharmaceutical program, with added quality-control expectations for the quantitative result. Key frameworks:

10 CFR Part 35 — Medical Use of Byproduct Material governs the medical use of PET radiopharmaceuticals, including authorized users, dosage determination and records, and the radiation safety officer's responsibilities. Positron emitters are byproduct material regulated by the NRC or an Agreement State. ⁷
Agreement State vs. NRC jurisdiction. Of the states DRPS serves, Florida, Maryland, Virginia, California, Nevada, Pennsylvania, New York, and New Jersey are NRC Agreement States that license medical use under their own radiation-control rules, while Washington DC and Delaware are regulated directly by the NRC. A generator-based Rb-82 program and a cyclotron-based N-13 or F-18 program carry different licensing and facility considerations, so confirm requirements with the authority having jurisdiction.
Professional procedure standards. The SNMMI/ASNC PET nuclear cardiology procedure standard and the joint SNMMI Cardiovascular Council–ASNC position paper on clinical MBF quantification define acquisition, processing, and reporting expectations, and the 2025 multi-society flurpiridaz guideline extends them to the new F-18 tracer. ^{1, 2, 6}
Accreditation and the physicist's role. Imaging programs are typically maintained under ACR or IAC accreditation, with an annual evaluation by a qualified medical physicist and equipment performance monitoring consistent with a recognized accrediting organization. Adding absolute flow adds calibration, dynamic-acquisition, and software-validation obligations on top of standard perfusion QC. ^{2, 6}

Because the quantitative result feeds clinical decisions, the program should document its tracer model, correction methods, software version, and normal ranges so the flow numbers are defensible. For the accreditation context, see ACR accreditation physics requirements and our PET/CT NEMA NU-2 performance testing guide.

Frequently Asked Questions (FAQs)

What is quantitative myocardial blood flow on cardiac PET?

Quantitative myocardial blood flow (MBF) is the absolute rate of blood delivery to the myocardium, expressed in milliliters per minute per gram of tissue (mL/min/g). Cardiac PET measures it by acquiring a dynamic scan that captures the tracer as it first passes through the blood pool and is taken up by the myocardium, then fitting a tracer-kinetic model to recover rest and stress flow. Dividing stress by rest gives myocardial flow reserve.

How is myocardial flow reserve (MFR) different from a relative perfusion scan?

A standard perfusion scan shows relative uptake — how one region compares with the apparently normal region in the same patient. Myocardial flow reserve (also called coronary flow reserve) is an absolute number: stress MBF divided by rest MBF. Because it does not rely on a normal reference region, MFR can reveal balanced, multivessel disease and microvascular dysfunction that look deceptively uniform on relative images.

Which PET tracers are used to measure myocardial blood flow?

The three established flow tracers are rubidium-82 (generator-produced, no cyclotron needed), N-13 ammonia (cyclotron-produced, higher myocardial extraction and sharper images), and O-15 water (cyclotron-produced, freely diffusible and flow-independent, the physiologic reference standard). F-18 flurpiridaz, approved in the United States in 2024, adds a longer-lived, cyclotron-distributed F-18 flow tracer with favorable imaging properties.

What is a normal myocardial flow reserve value?

In broad terms, a global myocardial flow reserve above about 2.0 is generally considered normal, and a value below roughly 1.5 signals a high-risk physiology. These thresholds are guides, not hard cutoffs; they depend on tracer, software, patient hemodynamics, and the population studied, and they should be interpreted together with the relative perfusion images and clinical context rather than in isolation.

Why does the choice of tracer affect the flow number?

Flow tracers that are extracted by the myocardium (rubidium-82 and N-13 ammonia) under-extract as flow rises — their first-pass extraction rolls off at high flow — so the kinetic model must apply a tracer-specific extraction correction to recover true flow. O-15 water is freely diffusible and its extraction stays near complete across the flow range, which is why it is the physiologic reference. Using the wrong tracer model, or an uncorrected uptake value, biases the reported flow.

What can go wrong in an MBF measurement?

The common failure modes are a mistimed or truncated dynamic acquisition that misses the arterial input peak, patient motion between the blood-pool and uptake phases, an incorrectly sampled arterial input function, residual tracer from the rest scan contaminating the stress scan, and misregistration between the emission data and the CT attenuation map. Each can shift the flow number without obviously degrading the static image, which is why quantitative PET needs dedicated quality control.

Does measuring absolute flow require special equipment or accreditation?

It requires a PET/CT capable of list-mode or fast dynamic acquisition, validated flow-quantification software, and a disciplined protocol. The medical use of PET radiopharmaceuticals falls under NRC 10 CFR Part 35 or an Agreement State program, and the imaging program is typically maintained under ACR or IAC accreditation with an annual qualified-medical-physicist evaluation. Absolute flow adds quality-control obligations on top of standard perfusion imaging rather than replacing them.

Key Takeaways

Absolute beats relative for diffuse disease. MBF and myocardial flow reserve reveal balanced multivessel disease and microvascular dysfunction that uniform-looking relative images can hide.
Uptake is not flow. Measured uptake $K_{1} = E \cdot F$ , and extraction $E$ rolls off at high flow, so every extracted tracer needs a validated extraction correction to recover true flow.
Tracers are not interchangeable. Rubidium-82, N-13 ammonia, O-15 water, and F-18 flurpiridaz differ in half-life, positron range, extraction, and logistics; flow thresholds are tracer- and software-specific.
Hemodynamics matter. Rest flow tracks workload, so normalize it to the rate–pressure product before computing flow reserve.
Reserve is prognostic. A low flow reserve (roughly below 1.5) independently predicts cardiac death, adding information beyond the relative scan.
Quantitative PET needs quantitative QC. Dynamic timing, motion, input sampling, and emission–CT registration can shift the number while the static image still looks fine.

Conclusion

Quantitative myocardial blood flow is one of cardiac PET's most valuable capabilities — and one of its most demanding. When the dynamic acquisition captures the input, the tracer model corrects extraction correctly, rest flow is normalized to workload, and the emission and attenuation data are properly registered, the resulting flow and flow-reserve numbers add a physiologic dimension that changes diagnosis and risk stratification. When any of those links is weak, the software still returns a plausible-looking number that can mislead. Treating absolute flow as a controlled, documented measurement — with tracer-specific models, standardized corrections, and dedicated quality control — is what makes it clinically trustworthy.

How DRPS Can Help

Diagnostic Radiation Physics Services supports nuclear cardiology and PET/CT programs that want to add or strengthen absolute-flow imaging. Our board-certified medical physicists provide PET/CT and nuclear medicine physics support, including scanner performance and calibration testing, dynamic-acquisition and flow-software commissioning, quantitative quality-control design, protocol standardization, and accreditation support aligned with ACR/IAC and state requirements.

DRPS serves facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware. To build defensible quantitative cardiac PET into your program, contact our team.

Related Resources

References

Murthy VL, Bateman TM, Beanlands RS, et al. Clinical quantification of myocardial blood flow using PET: joint position paper of the SNMMI Cardiovascular Council and the ASNC. Journal of Nuclear Medicine. 2018;59(2):273-293. doi:10.2967/jnumed.117.201368. PubMed
Dilsizian V, Bacharach SL, Beanlands RS, et al. ASNC imaging guidelines/SNMMI procedure standard for positron emission tomography (PET) nuclear cardiology procedures. Journal of Nuclear Cardiology. 2016;23(5):1187-1226. doi:10.1007/s12350-016-0522-3. PubMed
Murthy VL, Naya M, Foster CR, et al. Improved cardiac risk assessment with noninvasive measures of coronary flow reserve. Circulation. 2011;124(20):2215-2224. doi:10.1161/CIRCULATIONAHA.111.050427. PubMed
Juneau D, Wu KY, Kaps N, et al. Internal validation of myocardial flow reserve PET imaging using stress/rest myocardial activity ratios with Rb-82 and N-13-ammonia. Journal of Nuclear Cardiology. 2021;28(3):835-850. doi:10.1007/s12350-020-02464-y. PubMed
Shah NR, Charytan DM, Murthy VL, et al. Prognostic value of coronary flow reserve in patients with dialysis-dependent ESRD. Journal of the American Society of Nephrology. 2016;27(6):1823-1829. doi:10.1681/ASN.2015030301. PubMed
Packard RRS, Maddahi J, Pelletier-Galarneau M, et al. SNMMI/EANM/ASNC/ACNM procedure standard/practice guideline for 18F-flurpiridaz PET myocardial perfusion imaging and blood flow quantitation. Journal of Nuclear Cardiology. 2025;53:102473. doi:10.1016/j.nuclcard.2025.102473. PubMed
U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. ecfr.gov
U.S. Food and Drug Administration. Flyrcado (flurpiridaz F 18) injection — prescribing information. 2024. accessdata.fda.gov
American Society of Nuclear Cardiology. Cardiac PET clinical resources and guidelines. asnc.org
National Nuclear Data Center, Brookhaven National Laboratory. NuDat nuclear decay data (Rb-82, N-13, O-15). nndc.bnl.gov