Brain Perfusion SPECT: HMPAO and ECD Imaging

Brain perfusion SPECT maps regional cerebral blood flow by injecting a lipophilic Tc-99m tracer — HMPAO (exametazime) or ECD (bicisate) — that is trapped in brain tissue in proportion to perfusion at the moment of injection, then imaged tomographically. The physics that separates a diagnostic study from a misleading one lives in the details: radiopharmaceutical stability, correct energy windowing and collimation, disciplined center-of-rotation and uniformity quality control, and attenuation and scatter correction — increasingly paired with semiquantitative comparison to a normal database.¹²

This guide walks through what the tracers do, how the study is acquired and reconstructed, the corrections and quality control that protect quantitative accuracy, the clinical settings where the technique earns its place, and the regulatory framework that governs its use.

Introduction

Regional cerebral blood flow (rCBF) is tightly coupled to neuronal activity, so a snapshot of perfusion is a useful surrogate for regional brain function. Brain perfusion SPECT exploits this by using Tc-99m-labeled lipophilic tracers that cross the intact blood–brain barrier and are then converted to hydrophilic species that cannot easily diffuse back out — effectively freezing a picture of perfusion at the instant of injection.¹

That "freeze-frame" property is the technique's defining feature. Because the distribution is fixed within the first minute or two after injection but imaged minutes to hours later, the study reflects the physiologic state at injection, not during acquisition. This is what makes ictal seizure imaging possible: inject during a seizure, image afterward, and the tracer preserves the hyperperfusion pattern of the seizure onset zone.³⁴

Brain perfusion SPECT sits alongside PET and MRI in neuroimaging. It is comparatively inexpensive, widely available, and — with careful physics — quantitatively informative. DRPS supports nuclear medicine programs across Florida, Maryland, Virginia, Washington DC, California, and Nevada with PET/CT and nuclear medicine physics support that covers exactly the acquisition, QC, and correction steps described here.

Topic Explanation

The tracers: HMPAO and ECD

Two Tc-99m tracers dominate clinical brain perfusion SPECT:

• Tc-99m HMPAO (exametazime) — a lipophilic complex that crosses the blood–brain barrier and is converted intracellularly to a less lipophilic form, trapping it in proportion to flow. Its principal handling caveat is limited in-vitro stability after reconstitution unless a stabilized formulation is used, so prompt injection is important.¹
• Tc-99m ECD (bicisate) — also lipophilic and flow-dependent, but with greater in-vitro stability and faster blood and soft-tissue clearance. In a direct intrasubject comparison in healthy volunteers, ECD produced markedly higher brain-to-background contrast on delayed images (a brain-to-neck ratio of roughly 17:1 at five hours versus about 2:1 for HMPAO) and cleaner, easier-to-interpret images with less extracerebral activity.²

Both are read as maps of rCBF, but their differing kinetics affect background, timing flexibility, and the appearance of certain pathologies. The choice between them is partly clinical and partly logistical.

The acquisition chain

A brain perfusion SPECT study is a sequence of physics-sensitive steps:

1. Preparation and injection. The tracer is reconstituted and quality-checked (radiochemical purity), then injected under controlled, low-stimulation conditions when a resting baseline is desired.
2. Uptake and trapping. The tracer distributes and is trapped within about one to two minutes; the distribution then remains stable long enough to image.
3. Tomographic acquisition. A gamma camera with appropriate collimation rotates around the head, collecting projections at the 140 keV photopeak of Tc-99m.
4. Reconstruction. Projections are reconstructed — historically by filtered back-projection, now usually by iterative methods (OSEM).
5. Correction. Attenuation and scatter corrections are applied, ideally with CT on a SPECT/CT system.
6. Analysis. Images are reoriented, optionally spatially normalized, and compared qualitatively and semiquantitatively to expected patterns or a normal database.

For deeper treatment of the reconstruction and correction steps, see our companion guides to SPECT reconstruction with FBP and OSEM and SPECT scatter correction.

Key Technical Principles

Collimation, energy window, and sampling

Brain SPECT rewards high spatial resolution because the structures of interest — cortical ribbon, basal ganglia, thalami — are small and symmetric. Low-energy high-resolution (LEHR) parallel-hole collimators are the common choice, and dedicated fan-beam or multi-detector brain systems improve the resolution–sensitivity trade-off. Collimator choice is itself a physics decision; see our guide to gamma-camera collimator selection.

The energy window is centered on the 140 keV Tc-99m photopeak, typically ±10% to ±20%, chosen to balance count sensitivity against scatter acceptance. Angular sampling should be fine (small angular steps over 360°) and the radius of rotation as small as the head allows, because spatial resolution degrades with distance from the collimator.¹

Reconstruction: FBP and OSEM

Historically, brain SPECT projections were reconstructed by filtered back-projection (FBP), which is fast and linear but couples noise to the choice of reconstruction filter and produces streak artifacts around high-count regions. Modern practice uses iterative reconstruction, most commonly ordered-subsets expectation maximization (OSEM), which models the acquisition statistics and can incorporate resolution recovery, attenuation, and scatter within the reconstruction itself.⁶

OSEM speed and image properties depend on the number of iterations and the number of subsets, whose product sets the effective number of expectation-maximization updates:

$$n_{\text{eff}}=n_{\text{iterations}}\times n_{\text{subsets}}$$

More effective iterations recover resolution and contrast but also amplify noise, so brain protocols pair a modest number of updates with a smoothing post-filter (for example a Butterworth or Gaussian filter) tuned to the count statistics. The key physics discipline is to lock these settings — iterations, subsets, filter, and matrix/voxel size — and keep them consistent, because changing them alters the apparent perfusion pattern and breaks comparability across serial studies and against a normal database. For a fuller treatment, see our guide to SPECT reconstruction with FBP and OSEM.

Attenuation

Photons from deep structures traverse more tissue and are attenuated more than those from the cortex, artificially depressing central counts. Attenuation follows the exponential law for a monoenergetic beam:

$$I=I_0{e}^{-\mu x}$$

where $I_0$ is the unattenuated intensity, $x$ is the path length, and $\mu$ is the linear attenuation coefficient. For uniform (Chang) first-order correction of the head at 140 keV, an effective broad-beam coefficient of about $\mu \approx 0.11\text{–}0.12{\text{cm}}^{-1}$ is commonly assumed — lower than the roughly $0.15{\text{cm}}^{-1}$ narrow-beam value for water because scattered photons are included in the measured signal.⁶ On modern SPECT/CT systems, measured CT-based attenuation maps replace the uniform assumption and improve accuracy, and deep-learning methods to synthesize attenuation maps from the emission data are an active research area.⁶

Scatter

Scattered photons that fall within the energy window carry misplaced spatial information, reducing contrast and biasing quantification. Scatter correction — energy-window-based methods or model-based approaches integrated into iterative reconstruction — restores contrast and improves the accuracy of any subsequent quantification.¹⁶

Semiquantitative analysis and z-scores

Beyond visual reading, brain perfusion SPECT is increasingly analyzed against a healthy normal database. After spatial normalization to a template, regional counts are compared voxel-by-voxel and deviations expressed as a z-score:

$$z=\frac{C_{\text{region}}-{\mu }_{\text{normal}}}{{\sigma }_{\text{normal}}}$$

where $C_{\text{region}}$ is the (normalized) regional count value, and ${\mu }_{\text{normal}}$ and ${\sigma }_{\text{normal}}$ are the mean and standard deviation of that region in the normal database. Voxel-based tools such as statistical parametric mapping (SPM) apply this framework across the whole brain, improving detection of subtle, symmetric, or diffuse abnormalities.⁷ Semiquantitative analysis is only as valid as the corrections and QC that precede it — a bias in attenuation, scatter, or uniformity propagates directly into the z-map.

HMPAO versus ECD: a physics-and-handling comparison

Contrast ratios and stability notes are drawn from the cited comparison and guideline; individual product labeling and local formulation govern actual handling.¹²

Clinical Impact

Brain perfusion SPECT is most valuable when the clinical question is about regional function rather than structure, or when a physiologic snapshot must be captured at a specific moment. Its established roles include:

• Epilepsy presurgical evaluation. Ictal SPECT — injecting at seizure onset and imaging afterward — localizes the seizure onset zone as a region of hyperperfusion, and subtraction of ictal from interictal studies coregistered with MRI (SISCOM) sharpens localization, particularly in MRI-negative focal epilepsy.³⁴
• Dementia and neurodegeneration. Characteristic hypoperfusion patterns help differentiate dementia subtypes, complementing structural MRI and, where available, FDG PET.¹⁷
• Cerebrovascular disease. Perfusion mapping, sometimes with vasodilator challenge, assesses hemodynamic reserve.⁷
• Brain death confirmation. Cerebral scintigraphy is a recognized ancillary confirmatory test of brain death; absence of intracranial perfusion is highly specific, and lipophilic Tc-99m agents allow tomographic confirmation with high sensitivity and specificity.⁵

Across these uses, the diagnostic payoff depends on physics. A center-of-rotation error or a nonuniform detector can create a perfusion asymmetry that looks like pathology; an uncorrected attenuation gradient can mimic diffuse cortical hypoperfusion. The clinical value of the study is inseparable from the quality of the acquisition and correction chain.

Practical Optimization Tips

1. Protect radiopharmaceutical quality

Verify radiochemical purity and respect the stability window of the agent — especially HMPAO, which should be injected promptly unless a stabilized formulation is used. Poor tracer quality degrades every downstream step.¹

2. Control the injection environment

For resting baseline studies, inject in a quiet, dimly lit room with eyes and ears in a controlled state, because sensory stimulation redistributes flow. For ictal studies, inject as early in the seizure as possible; late injection captures postictal hypoperfusion instead of ictal hyperperfusion.³

3. Minimize radius of rotation

Keep the collimator as close to the head as safely possible and use a small, fixed radius of rotation. Spatial resolution falls off with distance, and the head allows a tighter orbit than most body imaging.¹

4. Commit to center-of-rotation and uniformity QC

Brain SPECT is unforgiving of instrument error. Maintain daily energy-peak and uniformity checks, periodic high-count extrinsic uniformity for the collimators in use, and center-of-rotation verification, because these errors produce ring and streak artifacts and left–right asymmetries that mimic disease. See our guides to gamma-camera uniformity QC and SPECT/CT quality control.

5. Apply attenuation and scatter correction consistently

Use CT-based attenuation correction on SPECT/CT where available, or a validated uniform method otherwise, and apply scatter correction consistently. Consistency matters as much as method when comparing serial studies or using a normal database.⁶

6. Standardize for semiquantitative analysis

If you use z-score or SPM analysis, lock down acquisition, reconstruction, and correction parameters to match the normal database's acquisition conditions. A mismatch between patient and database processing invalidates the comparison.⁷

Common pitfalls to avoid

• Injecting HMPAO outside its stability window. Degraded tracer yields poor, non-diagnostic images.
• Late ictal injection. Capturing postictal hypoperfusion defeats the purpose of ictal SPECT.
• Skipping center-of-rotation QC. Small COR errors create artifacts that read as pathology.
• Inconsistent correction. Changing attenuation or scatter methods between studies breaks comparability.
• Over-reading asymmetry. Not every left–right difference is disease; confirm it survives QC and correction.

Regulatory Considerations

Brain perfusion SPECT uses byproduct material, so its regulatory framework is NRC or Agreement State medical-use licensing, layered with equipment performance standards and professional practice guidelines. The relevant frameworks include:

• 10 CFR Part 35 — Medical Use of Byproduct Material, which governs authorized use of Tc-99m radiopharmaceuticals, the radiation safety officer's responsibilities, dosage determination, and instrument requirements.
• 10 CFR Part 20 — Standards for Protection Against Radiation, which sets occupational and public dose limits framing the safety program.
• ICRP Publication 128, which tabulates radiation dose to patients from radiopharmaceuticals, including Tc-99m HMPAO and ECD, and supports informed administered-activity choices.⁸
• IAEA and NEMA performance standards — IAEA Human Health Series guidance on SPECT quality assurance and NEMA NU 1 performance measurements of gamma cameras define the QC methods that keep the system quantitatively trustworthy.⁹¹⁰

Agreement States administer equivalent medical-use programs. Of the states DRPS serves, Florida, Maryland, Virginia, California, and Nevada are NRC Agreement States that license medical use under their own radiation-control rules, while Washington, DC is regulated directly by the NRC. Facilities should document tracer QC, camera QC, correction methods, and administered-activity rationale so the program is defensible during inspection. For related program elements, see our guide to quantitative SPECT/CT calibration.

Frequently Asked Questions (FAQs)

What does brain perfusion SPECT actually measure?

It measures the regional distribution of cerebral blood flow at the moment of injection. Lipophilic Tc-99m tracers cross the blood–brain barrier and are trapped in proportion to perfusion, then held long enough to image, so the resulting three-dimensional map reflects the flow pattern fixed at injection.¹

What is the difference between Tc-99m HMPAO and Tc-99m ECD?

Both are lipophilic Tc-99m brain perfusion agents. HMPAO has limited in-vitro stability after reconstitution unless stabilized and should be injected promptly; ECD is more stable, clears faster, and tends to produce lower extracerebral background and higher brain-to-background contrast on delayed images.¹²

Why are attenuation and scatter correction important in brain SPECT?

Deep photons are attenuated more than cortical photons, depressing central counts, and scattered photons blur contrast and bias quantification. Attenuation correction (Chang's uniform method or CT-based on SPECT/CT) and scatter correction restore a more accurate distribution, which is essential for normal-database comparison.⁶

How much radiation dose does a brain perfusion SPECT deliver?

Adult administered activity is commonly about 555 to 1110 MBq (roughly 15 to 30 mCi) per EANM guidance, giving an effective dose on the order of several millisieverts. Agent- and activity-specific coefficients are tabulated in ICRP Publication 128.¹⁸

What quality control does a brain SPECT system need?

Daily energy-peak and uniformity checks, periodic high-count extrinsic uniformity, center-of-rotation verification, spatial-resolution and SPECT phantom testing, and pixel-size calibration. Center-of-rotation and uniformity errors are especially damaging because they create artifacts that mimic or mask perfusion deficits.⁹¹⁰

Can brain perfusion SPECT be quantified?

It is most often interpreted semiquantitatively: after correction and spatial normalization, regional counts are compared to a normal database and deviations expressed as z-scores using tools such as statistical parametric mapping.⁷

Key Takeaways

• The technique freezes perfusion at injection. Tc-99m HMPAO and ECD trap in proportion to flow within a minute or two, enabling delayed imaging and ictal studies.¹³
• HMPAO and ECD differ in handling and contrast. HMPAO needs prompt injection unless stabilized; ECD is more stable with lower background and higher delayed contrast.¹²
• Corrections protect accuracy. Attenuation and scatter correction, ideally CT-based, are prerequisites for reliable reading and quantification.⁶
• QC is not optional. Center-of-rotation and uniformity errors create artifacts that read as disease.⁹¹⁰
• Semiquantitative analysis adds value. Z-score and SPM comparison to a normal database standardizes interpretation of subtle findings.⁷
• Clinical roles are functional. Epilepsy localization, dementia differentiation, cerebrovascular reserve, and brain-death confirmation are the established uses.³⁴⁵

Conclusion

Brain perfusion SPECT remains a clinically valuable window on regional cerebral blood flow, but its value is earned through physics. The tracer must be prepared and injected correctly; the camera must be collimated, peaked, and QC-verified; the data must be reconstructed and corrected for attenuation and scatter; and, where quantification is used, the whole chain must match the normal database. When these steps are disciplined, brain perfusion SPECT delivers robust, reproducible functional images that complement structural imaging and inform real clinical decisions. When they are not, the same study can generate artifacts that masquerade as disease. The difference is a well-run physics and quality-control program.

How DRPS Can Help

Diagnostic Radiation Physics Services helps nuclear medicine facilities keep SPECT and SPECT/CT systems quantitatively trustworthy. For brain perfusion imaging, this includes gamma-camera and SPECT/CT acceptance and annual testing, center-of-rotation and uniformity QC review, attenuation- and scatter-correction validation, protocol and administered-activity review, and documentation for accreditation and inspection, delivered by board-certified medical physicists through our PET/CT and nuclear medicine physics and medical physics consulting services.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware.

A brain SPECT program is only as good as the physics behind the picture — that is where we help.

Related Resources

• SPECT reconstruction with FBP and OSEM
• SPECT scatter correction
• Gamma-camera uniformity QC
• SPECT/CT quality control
• Quantitative SPECT/CT calibration
• Gamma-camera collimator selection
• PET/CT and nuclear medicine physics
• Medical physicist consulting

References

1. Kapucu ÖL, Nobili F, Varrone A, et al. EANM procedure guideline for brain perfusion SPECT using 99mTc-labelled radiopharmaceuticals, version 2. Eur J Nucl Med Mol Imaging. 2009;36(12):2093-2102. doi:10.1007/s00259-009-1266-y. doi.org
2. Léveillé J, Demonceau G, Walovitch RC. Intrasubject comparison between technetium-99m-ECD and technetium-99m-HMPAO in healthy human subjects. J Nucl Med. 1992;33(4):480-484. PubMed
3. Van Paesschen W. Ictal SPECT. Epilepsia. 2004;45(Suppl 4):35-40. doi:10.1111/j.0013-9580.2004.04008.x. doi.org
4. Cendes F, Theodore WH, Brinkmann BH, Sulc V, Cascino GD. Neuroimaging of epilepsy. Handb Clin Neurol. 2016;136:985-1014. doi:10.1016/B978-0-444-53486-6.00051-X. doi.org
5. Conrad GR, Sinha P. Scintigraphy as a confirmatory test of brain death. Semin Nucl Med. 2003;33(4):312-323. doi:10.1016/s0001-2998(03)00034-5. doi.org
6. Du Y, Jiang H, Lin CN, et al. Generative adversarial network-based attenuation correction for 99mTc-TRODAT-1 brain SPECT. Front Med (Lausanne). 2023;10:1171118. doi:10.3389/fmed.2023.1171118. doi.org
7. Taghizadeh Asl M, Nemati R, Chabi N, et al. Brain perfusion imaging with voxel-based analysis in secondary progressive multiple sclerosis patients with a moderate to severe stage of disease: a boon for the workforce. BMC Neurol. 2016;16:79. doi:10.1186/s12883-016-0605-4. doi.org
8. International Commission on Radiological Protection. ICRP Publication 128: Radiation Dose to Patients from Radiopharmaceuticals — A Compendium of Current Information Related to Frequently Used Substances. Ann ICRP. 2015;44(2S). icrp.org
9. International Atomic Energy Agency. Quality Assurance for SPECT Systems. IAEA Human Health Series No. 6. Vienna: IAEA; 2009. iaea.org
10. National Electrical Manufacturers Association. NEMA NU 1: Performance Measurements of Gamma Cameras. Rosslyn, VA: NEMA. nema.org
11. U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. nrc.gov
12. U.S. Nuclear Regulatory Commission. 10 CFR Part 20: Standards for Protection Against Radiation. nrc.gov