SPECT/CT Attenuation Correction Explained

SPECT/CT attenuation correction uses the co-acquired CT to build a patient-specific map of how gamma rays are absorbed inside the body, scales that map to the emission photon energy, and recovers the counts lost to attenuation during reconstruction. It is what makes SPECT quantitative and what removes the depth-dependent shading that once produced false cardiac defects — but the correction is only as trustworthy as the registration, the map, and the physicist's artifact review behind it.¹²

Attenuation is the largest physical degradation in single-photon imaging. A gamma ray emitted deep in the thorax can be attenuated by an order of magnitude before it reaches the detector, so uncorrected SPECT systematically underrepresents deep activity and makes image intensity a function of location as much as uptake. Hybrid SPECT/CT solved the practical problem of measuring attenuation directly, but it also introduced new, correction-specific artifacts that every nuclear medicine program must understand and control.²³

This guide explains the physics of photon attenuation, how a CT scan becomes an energy-appropriate attenuation map, why correction is essential for cardiac, bone, and quantitative imaging, the artifacts it can create, and how a medical physicist verifies it. DRPS supports SPECT/CT programs through PET/CT and nuclear medicine physics services across Florida, Maryland, Virginia, Washington DC, California, Nevada, and our other service areas.

Introduction

Single-photon emission computed tomography reconstructs the three-dimensional distribution of a gamma-emitting radiotracer. The measurement it makes, however, is not the emitted activity but the detected activity — and detection is reduced by every centimeter of tissue a photon must traverse. The reduction is exponential, so the effect is severe and strongly depth-dependent.²

Before hybrid systems, nuclear medicine managed attenuation with transmission line sources or with uniform-attenuation assumptions. Both were limited: transmission sources were noisy and slow, and uniform maps fail wherever the body is heterogeneous, as in the chest. The integration of a CT scanner onto the gamma-camera gantry — SPECT/CT — provided a fast, low-noise, high-resolution, patient-specific measurement of attenuation, and it enabled truly quantitative SPECT.¹²

But CT measures attenuation at diagnostic X-ray energies, not at the emission energy of the radiotracer, and the CT is acquired at a slightly different time and breathing state than the emission data. Bridging those gaps — energy scaling and spatial registration — is where the physics and the quality control live. This guide follows that chain end to end.

Topic Explanation

What attenuation correction actually corrects

Attenuation correction compensates the SPECT reconstruction for the fraction of emitted photons absorbed or scattered out of the acquired projection line before reaching the detector. For a photon emitted at depth $d$ traveling toward the detector through tissue with linear attenuation coefficient $μ$ , the probability of reaching the detector without interaction falls exponentially with path length. Correcting for this restores the relationship between reconstructed counts and true activity concentration.²

Attenuation correction is distinct from — but works alongside — scatter correction and resolution recovery. Scatter correction removes counts that were deflected into a projection line; attenuation correction restores counts that were removed from it. The two are complementary and are usually applied together in modern iterative reconstruction, as covered in our companion guide to SPECT scatter correction.¹³

From Hounsfield units to a mu-map

The CT does not directly measure attenuation at the tracer's emission energy. CT reconstructs linear attenuation at the CT beam's effective energy and reports it in Hounsfield units (HU), where water is 0 HU and air is −1000 HU. To turn this into a usable attenuation map, the system applies a bilinear (two-segment) scaling model: one linear relationship converts HU to the emission-energy linear attenuation coefficient in the air-to-soft-tissue range, and a second, steeper relationship applies in the soft-tissue-to-bone range, because bone's attenuation scales differently with energy than soft tissue does.¹²

The output is a voxel-by-voxel mu-map at the correct emission energy — for example, at 140 keV for technetium-99m or 208 keV for lutetium-177. That map is then incorporated into the iterative reconstruction (typically OSEM), where it defines the attenuation factors applied to each projection ray. For the reconstruction framework these corrections plug into, see SPECT reconstruction: FBP and OSEM.

Key Technical Principles

The attenuation integral and the correction factor

For a projection line $L$ through the patient, the total attenuation experienced by photons is governed by the line integral of the attenuation coefficient. The attenuated projection $p$ relates to the unattenuated projection $p_{0}$ by:

p = p_{0} exp (- \int_{L} μ (x) d x)

The attenuation correction factor (ACF) is the reciprocal of that transmission term:

ACF = exp (+ \int_{L} μ (x) d x)

In a heterogeneous body the integral must be evaluated voxel by voxel along each ray — which is exactly the information the CT-derived mu-map supplies. In a uniform-attenuation approximation the integral collapses to $μ \cdot L$ , which is why uniform maps are only acceptable where tissue is genuinely homogeneous.²

Worked example: how much signal is lost at depth

Consider technetium-99m (140 keV), whose photons travel through soft tissue with an approximate narrow-beam linear attenuation coefficient of $μ \approx 0.15 cm^{- 1}$ . For a source 10 cm deep, the fraction of photons reaching the detector without attenuation is:

\frac{p}{p _{0}} = e^{- μ d} = e^{- (0.15) (10)} = e^{- 1.5} \approx 0.22

Roughly 78% of the signal is lost, and the required correction factor is:

ACF = e^{+ 1.5} \approx 4.5

A source 15 cm deep would need $ACF = e^{(0.15) (15)} = e^{2.25} \approx 9.5$ . Because the correction factor grows exponentially with depth, a deep myocardial or pelvic source can require nearly an order-of-magnitude correction, while a superficial source needs almost none. This is precisely the depth dependence that makes uncorrected SPECT unreliable and quantification impossible without a map.²

Emission energy matters: the mu-map is energy-specific

Soft-tissue attenuation depends on photon energy, so the mu-map must match the radionuclide. Representative approximate narrow-beam soft-tissue (water) linear attenuation coefficients are:

Radionuclide	Principal imaging photon	Approx. soft-tissue μ	Relative correction at 10 cm depth (ACF)
Tc-99m	140 keV	≈ 0.150 cm⁻¹	≈ 4.5
I-123	159 keV	≈ 0.145 cm⁻¹	≈ 4.3
In-111	171 / 245 keV	≈ 0.132 cm⁻¹	≈ 3.7
Lu-177	208 keV	≈ 0.128 cm⁻¹	≈ 3.6

(Values are approximate, energy-dependent, and dataset-dependent; a facility mu-map is derived from the CT and scaled to the specific emission energy.) The table shows why the same CT scan must be scaled differently for a Tc-99m bone study, an I-123 DaTscan, and a Lu-177 post-therapy scan — and why applying the wrong energy scaling systematically biases quantification.¹²

Why quantification needs it

Quantitative SPECT/CT reports activity concentration (Bq/mL) or standardized uptake values, which are only meaningful once attenuation, scatter, and resolution effects are corrected and the system is calibrated against a known activity. The EANM practice guideline for quantitative SPECT-CT identifies CT-based attenuation correction as a core element of the quantification chain, alongside scatter correction and system calibration. Quantitative bone SPECT/CT, for example, uses corrected SUVs to separate active metastatic uptake from benign degenerative uptake — a discrimination that collapses without accurate attenuation correction.¹⁵ For the calibration side of this chain, see quantitative SPECT/CT calibration.

Clinical Impact

Attenuation correction changes interpretation, not just image cosmetics. In cardiac perfusion SPECT, soft-tissue attenuation from the diaphragm or breast produces classic false defects — inferior-wall defects in men, anterior defects in women — that mimic ischemia or infarction. CT-based attenuation correction substantially reduces these false positives and improves specificity, which is one of the strongest clinical arguments for hybrid cardiac SPECT/CT.³⁴

For oncology, hybrid SPECT/CT with attenuation correction improves lesion localization and characterization and enables the volumetric dosimetry that underpins theranostics; reviews of neuroendocrine, adrenal, and skeletal imaging document these gains in specificity and management impact.⁵⁶ In bone imaging, quantitative uptake values derived from attenuation-corrected SPECT/CT help discriminate malignant from degenerative foci.⁵

The clinical caveat is equally important: because attenuation correction can create artifacts, standard practice is to review corrected and uncorrected data together. A defect that appears only after correction, or disappears only after correction, is a signal to check registration, truncation, and metal before making a clinical call.³ This dual-review discipline connects directly to cardiac SPECT MPI quality control.

Practical Optimization Tips

Guard registration first

Check SPECT/CT alignment on every cardiac study. Respiratory and bulk motion between the emission and CT acquisitions is the most common and most damaging error; even a one-to-two-pixel shift can shift the myocardial mu-map onto lung and create a false defect. Use the vendor's fusion display and manual re-registration tools before accepting corrected images.³
Coach breathing. Acquire the CT in a breathing state that matches the average of the emission acquisition, per vendor guidance, rather than a deep inspiration that mismatches the SPECT.

Manage truncation, metal, and contrast

Watch for truncation. Large patients can extend beyond the CT field of view; the truncated mu-map underestimates attenuation at the body edge and can bias correction. Position to keep the region of interest within the CT field of view.
Recognize metal and contrast. Hip prostheses, pacemakers, port catheters, and iodinated contrast inflate HU and, through the bilinear model, overcorrect the mu-map, producing false hot spots adjacent to the metal. Review the CT and the fused images for these features.

Use the right map for the region

Non-uniform CT maps for the torso; uniform maps only where justified. In the brain, a uniform or measured map may both perform acceptably; in the thorax and abdomen, only a measured, non-uniform CT map captures the lung-soft-tissue-bone heterogeneity.²
Keep the CT correction-appropriate. A low-dose CT is sufficient for attenuation and localization; do not over-scan. But confirm the CT technique yields a map free of excessive noise or beam-hardening that would corrupt the coefficients.

Common pitfalls to avoid

Trusting corrected images without the uncorrected set. Always review both.
Ignoring a shifted fusion overlay on cardiac studies.
Applying the wrong emission-energy scaling for the radionuclide.
Overlooking truncation in large patients.
Reading a peri-prosthetic hot spot as pathology without checking the CT.

Regulatory Considerations

A SPECT/CT system sits under two regulatory frameworks at once: the radioactive material rules governing the radiopharmaceutical, and the machine rules governing the CT subsystem. Both apply, and the attenuation-correction CT — however low its dose — is a regulated X-ray exposure.

Byproduct material. Medical use of the radiotracer is governed by 10 CFR Part 35 (or the equivalent Agreement State program), with dose limits under 10 CFR Part 20.⁷
CT subsystem. The CT is a radiation-producing machine subject to the FDA performance standard at 21 CFR 1020.33 and to state radiation-control programs; CTDIvol should be recorded and justified even for a localization/attenuation CT.⁸
Performance standards. NEMA NU 1-2023, the current standard for gamma-camera performance, includes SPECT/CT co-registration accuracy and absolute quantification accuracy among its parameters, formalizing what a quantitative SPECT/CT program must verify.⁹
Guidance. The IAEA Human Health Reports No. 9 and the EANM practice guideline for quantitative SPECT-CT provide the methodological framework for attenuation correction, calibration, and harmonization.¹¹⁰
State and Agreement-State jurisdiction. 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 (in Florida, Chapter 64E-5, F.A.C.), while Washington DC and Delaware are regulated directly by the NRC. Confirm the licensing authority and the CT-machine registration requirements that apply.

Documented SPECT/CT co-registration checks, mu-map and calibration verification, and CT technique review are what make a quantitative program defensible. See our overviews of SPECT/CT quality control and PET/CT attenuation correction for the parallel workflows.

Frequently Asked Questions (FAQs)

What is SPECT/CT attenuation correction?

SPECT/CT attenuation correction is the process of using the co-registered CT scan to estimate how much of the gamma-ray signal from a radiotracer is absorbed as it travels out of the patient, then compensating the SPECT reconstruction for that loss. The CT provides a patient-specific map of attenuation (a mu-map) that is scaled from CT X-ray energies to the emission photon energy, so that reconstructed counts more faithfully represent the true activity distribution.

Why is attenuation correction important in SPECT?

Photons emitted deep in the body are attenuated far more than photons from superficial structures, so without correction, deep activity appears artificially low and image intensity depends on depth rather than true uptake. Attenuation correction removes most of this depth dependence, which reduces false defects (for example, in cardiac perfusion imaging), improves uniformity, and is a prerequisite for quantitative SPECT that reports activity concentration or standardized uptake values.

How does the CT scan become an attenuation map?

CT reconstructs linear attenuation at the CT beam's effective energy, expressed in Hounsfield units. A bilinear (two-segment) scaling model converts Hounsfield units to the linear attenuation coefficient at the radionuclide's emission energy — one slope for air-to-water tissue and another for water-to-bone. The result is an energy-appropriate mu-map used to compute correction factors during iterative reconstruction.

What artifacts can SPECT/CT attenuation correction introduce?

The main failure modes are misregistration between the SPECT and CT (from patient motion or respiration, which is especially damaging in cardiac imaging), CT truncation when the patient extends beyond the CT field of view, and metal or contrast artifacts that inflate Hounsfield units and overcorrect the mu-map. Each can create false defects or false hot spots, so both corrected and uncorrected images are typically reviewed together.

Do I need a CT for attenuation correction, or can I use a uniform map?

Historically, transmission sources and uniform (Chang) attenuation maps were used, and uniform maps can be adequate in regions of relatively homogeneous soft tissue such as the brain. For the thorax, abdomen, and any quantitative work, a measured, non-uniform CT-based map is strongly preferred because attenuation there varies sharply between lung, soft tissue, and bone. Newer deep-learning methods can estimate maps without CT on some systems but remain under validation.

How is SPECT/CT attenuation correction quality-controlled?

A medical physicist verifies SPECT/CT co-registration with a phantom, checks the accuracy of the mu-map and calibration against known activity, confirms that CTDIvol and CT technique are appropriate for a correction-quality scan, and reviews for misregistration, truncation, and metal artifacts. Absolute quantification accuracy and SPECT/CT co-registration are explicit performance parameters in modern gamma-camera standards.

Key Takeaways

Attenuation is exponential with depth. A 10 cm-deep Tc-99m source loses roughly 78% of its signal and needs a correction factor near 4.5; deeper sources need nearly an order of magnitude.²
The CT becomes an energy-scaled mu-map. A bilinear model converts Hounsfield units to the linear attenuation coefficient at the tracer's emission energy.¹²
Correction is essential for quantification and cardiac accuracy. It underpins Bq/mL and SUV reporting and reduces false cardiac defects.¹³⁴
Correction can create artifacts. Misregistration, truncation, and metal/contrast are the key failure modes; review corrected and uncorrected data together.³
Match the map to the radionuclide and region. Use non-uniform CT maps in the torso and the correct emission-energy scaling.²
Verify it. Co-registration, mu-map/calibration accuracy, and CT technique are the QC backbone, and are named in NEMA NU 1-2023.⁹

Conclusion

Attenuation correction is the single most important physical correction in SPECT, and hybrid SPECT/CT made it fast, patient-specific, and quantitative. The physics is straightforward — recover the exponential signal loss along each ray — but the execution depends on two bridges that can fail: scaling the CT to the emission energy, and registering the CT to the emission data. When those bridges hold, attenuation correction removes depth-dependent shading, cuts false cardiac defects, and enables absolute quantification and theranostic dosimetry. When they fail, the same correction manufactures defects and hot spots. That is why a disciplined program pairs every corrected dataset with its uncorrected counterpart and backs both with documented physics QC.

How DRPS Can Help

Diagnostic Radiation Physics Services helps nuclear medicine programs deploy and defend quantitative SPECT/CT. Our PET/CT and nuclear medicine physics support includes SPECT/CT co-registration verification, mu-map and calibration accuracy checks, attenuation- and scatter-correction validation, CT-technique and CTDIvol review, artifact troubleshooting, and acceptance and annual survey documentation prepared by board-certified medical physicists. We also provide medical physicist consulting and accreditation support across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware.

A strong SPECT/CT program is not just about turning correction on. It is about proving that the correction is registered, energy-scaled, calibrated, and artifact-checked on every study.

Related Resources

References

Dickson JC, Armstrong IS, Gabiña PM, et al. EANM practice guideline for quantitative SPECT-CT. European Journal of Nuclear Medicine and Molecular Imaging. 2023;50(4):980-995. doi:10.1007/s00259-022-06028-9. doi.org
Lee TC, Alessio AM, Miyaoka RM, Kinahan PE. Morphology supporting function: attenuation correction for SPECT/CT, PET/CT, and PET/MR imaging. The Quarterly Journal of Nuclear Medicine and Molecular Imaging. 2016;60(1):25-39. PubMed
Celler A, Shcherbinin S, Hughes T. An investigation of potential sources of artifacts in SPECT-CT myocardial perfusion studies. Journal of Nuclear Cardiology. 2010;17(2):232-246. doi:10.1007/s12350-009-9171-0. doi.org
Yang J, Shi L, Wang R, et al. Direct attenuation correction using deep learning for cardiac SPECT: a feasibility study. Journal of Nuclear Medicine. 2021;62(11):1645-1652. doi:10.2967/jnumed.120.256396. doi.org
Kuji I, Yamane T, Seto A, et al. Skeletal standardized uptake values obtained by quantitative SPECT/CT as an osteoblastic biomarker for the discrimination of active bone metastasis in prostate cancer. European Journal of Hybrid Imaging. 2017;1(1):2. doi:10.1186/s41824-017-0006-y. doi.org
Wong KK, Chondrogiannis S, Fuster D, et al. Additional value of hybrid SPECT/CT systems in neuroendocrine tumors, adrenal tumors, pheochromocytomas and paragangliomas. Revista Española de Medicina Nuclear e Imagen Molecular. 2017;36(2):103-109. doi:10.1016/j.remn.2016.09.003. doi.org
U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. ecfr.gov
U.S. Food and Drug Administration. 21 CFR 1020.33 — Computed tomography (CT) equipment. ecfr.gov
National Electrical Manufacturers Association. NEMA NU 1-2023: Performance Measurements of Gamma Cameras. Rosslyn, VA: NEMA; 2023. nema.org
International Atomic Energy Agency. Quantitative Nuclear Medicine Imaging: Concepts, Requirements and Methods. Human Health Reports No. 9. Vienna: IAEA; 2014. iaea.org
National Institute of Standards and Technology. XCOM: Photon Cross Sections Database (mass attenuation coefficients for water and tissue). nist.gov