Dual-Isotope SPECT Imaging: Crosstalk Correction

Dual-isotope SPECT images two radionuclides in a single study, but the two energy windows contaminate each other through downscatter, photopeak spillover, and septal penetration — so a defensible protocol is defined by its crosstalk correction, not just its tracer pair. The medical physicist's job is to pick isotopes whose photopeaks can be separated, set energy windows that the camera can actually resolve, and apply and validate a correction that removes the contamination without destroying the counts.

Imaging two tracers at once is attractive: one patient visit, one set of positioning, perfectly registered images, and in some protocols a shorter study. But single-photon detectors sort photons by energy, and two radionuclides emitting photons of similar energy will always leak into each other's energy windows. That leakage — crosstalk — is the central physics problem of dual-isotope SPECT, and whether it is handled well or badly determines whether the images are quantitative or merely pretty.¹²

This guide explains where crosstalk comes from, how energy-window and model-based corrections remove it, what quantitative accuracy is realistic, and how DRPS helps nuclear medicine programs set up and validate dual-isotope protocols as part of its PET/CT and nuclear medicine physics services across Florida, Maryland, Virginia, Washington DC, California, Nevada, and the Northeast.

Introduction

A dual-isotope study depends on the camera being able to tell two photon populations apart by energy, and no detector does that perfectly. The gamma camera assigns each detected event to an energy window based on the pulse height it produces, and because of Compton scatter in the patient and the finite energy resolution of the detector, a fraction of one isotope's events always land in the other isotope's window.¹³

There are two ways to run a dual-isotope study. In separate (sequential) acquisition, the two tracers are imaged one after another — the contamination is smaller but the images can misregister if the patient moves between scans, and the study is longer. In simultaneous acquisition, both tracers are imaged at the same time with two energy windows open at once — the images are intrinsically co-registered and the study is faster, but crosstalk is unavoidable and must be corrected.²⁴

The remainder of this guide treats simultaneous acquisition as the demanding case, because that is where the physics of crosstalk correction earns its keep.

Topic Explanation

What "dual-isotope" actually means at the detector

A single-photon emission computed tomography (SPECT) camera does not measure photon energy exactly. Each photon deposits energy in the scintillator (or semiconductor), producing a pulse whose height is proportional to the deposited energy but blurred by the statistics of light (or charge) collection. The spread of that blur is the detector's energy resolution, usually quoted as the full width at half maximum (FWHM) of the photopeak expressed as a percentage of the photopeak energy.³

A conventional sodium iodide, NaI(Tl), camera has an energy resolution of roughly 9 to 10 percent at the 140 keV photopeak of Tc-99m. Solid-state cameras built from cadmium zinc telluride (CZT) or cadmium telluride (CdTe) reach about 5 to 6 percent, producing much narrower photopeaks.⁴ That difference is decisive for dual-isotope work: the narrower the photopeak, the less of one isotope spills into a neighboring window.

When two radionuclides are present, the camera sets a separate energy window on each photopeak. The counts recorded in each window are then treated as an image of that isotope — but only after the contamination from the other isotope is removed.

Where the energy windows fall

The feasibility of any dual-isotope pair comes down to how far apart the photopeaks are and how penetrating the higher-energy isotope is. A few representative pairs:

The Tl-201/Tc-99m pair is comparatively forgiving because 70 keV and 140 keV are far apart; the problem is almost entirely Tc-99m scattering down into the Tl-201 window. The Tc-99m/I-123 pair is far harder because 140 keV and 159 keV are separated by only about 19 keV — closer than the photopeak width of a NaI(Tl) camera — so the two photopeaks physically overlap and spillover dominates.⁴

For the emission characteristics that drive these choices, see our companion guide to common PET and radiopharmaceutical-therapy isotopes, and for the windowing fundamentals, SPECT scatter correction.

Key Technical Principles

The three sources of crosstalk

Crosstalk — counts from one radionuclide recorded in the other's window — has three physical origins:

1. Downscatter. A high-energy photon Compton-scatters in the patient or collimator, loses energy, and arrives within the lower-energy window. Tc-99m photons at 140 keV readily scatter down into a Tl-201 70 keV or Ho-166 81 keV window. Because scattered photons also come from the isotope's own primary emission, downscatter correction and ordinary scatter correction are the same physics applied across windows.¹⁶
2. Photopeak spillover (tailing). Even an unscattered photon is recorded with a blurred pulse height. The tail of a higher photopeak extends into an adjacent lower window and vice versa. Spillover is governed by energy resolution, which is why CZT/CdTe systems tolerate close pairs that NaI(Tl) cannot.⁴
3. Septal penetration and backscatter. Isotopes such as I-123 emit a small fraction of high-energy photons (for example near 529 keV) that penetrate collimator septa and can scatter in the crystal or backscatter from structures behind the detector, adding counts across the spectrum. Choosing the correct collimator — often a medium- or high-energy design for I-123 — limits this contribution.⁴

The triple-energy-window method

The most widely implemented energy-domain correction is the triple-energy-window (TEW) method of Ogawa and colleagues. The scatter (or downscatter) contribution inside a main photopeak window is estimated from the counts in two narrow subwindows placed immediately below and above it, assuming the scatter spectrum is approximately trapezoidal across the main window:⁹

Here (C_{\text{low}}) and (C_{\text{high}}) are the counts in the lower and upper subwindows, (w_{\text{low}}) and (w_{\text{high}}) are their widths, and (w_{\text{main}}) is the width of the main photopeak window. The estimated scatter is subtracted pixel-by-pixel from the main-window image.

Worked example. Suppose a Tc-99m main window spans 126–154 keV, so (w_{\text{main}} = 28\ \text{keV}), with 7 keV subwindows on each side. In a given pixel the subwindows record (C_{\text{low}} = 40) and (C_{\text{high}} = 10) counts. The estimated scatter in the main window is:

If the main window recorded 600 counts, the scatter-corrected (primary) estimate is about 500 counts — a 17 percent correction in this pixel. When the "high-energy" neighbor is a second isotope rather than the same isotope's own tail, the same arithmetic estimates the downscatter leaking across isotopes. TEW is attractive because it is linear, fast, and requires no model of the patient, but it is noisy in low-count subwindows, so the subwindow images are usually smoothed before subtraction.⁹

Model-based and Monte Carlo downscatter correction

When the crosstalk is large or the quantification demands are strict, a model-based approach does better. The higher-energy isotope is reconstructed first (with attenuation, collimator-response, and scatter modeling); that reconstructed distribution is then used to simulate how many of its photons scatter down into the lower isotope's window, and the simulated downscatter is incorporated into the lower isotope's iterative reconstruction.

Kangasmaa and colleagues demonstrated this for simultaneous Tl-201/Tc-99m cardiac SPECT: an ordered-subsets expectation-maximization (OS-EM) reconstruction used an efficient Monte Carlo simulator to estimate Tc-99m downscatter into the Tl-201 window, converging in under three minutes with as few as two scatter-update iterations.⁵ Van Rooij and colleagues applied the same idea to simultaneous Ho-166/Tc-99m liver SPECT for radioembolization dosimetry, modeling the crosstalk in both directions; with a Ho-166-to-Tc-99m activity ratio of about 5:1, Ho-166 activity recovery was reduced by only about 10 percent by the presence of Tc-99m.⁶ On a solid-state CdTe brain system, Takeuchi and colleagues used a five-energy-window scheme to separate simultaneous Tc-99m and I-123, recovering activity within a few percent (percentage difference between single and dual acquisition about 4 to 6 percent) despite the close photopeaks.⁴

Why the activity ratio is a design variable

A subtle but important point: the ratio of administered activities is not fixed by the clinical question alone. If the higher-energy isotope is given in far greater activity than the lower-energy one, its downscatter can swamp the weaker isotope's window and no correction will recover good statistics. Van Rooij's 5:1 Ho-166:Tc-99m ratio was chosen precisely so that Tc-99m — which contributed roughly twice as many counts as Ho-166 to the 81 keV window at equal activity — would not overwhelm the therapeutic isotope's image.⁶ The physicist should treat the activity ratio, the energy windows, and the correction as one coupled design.

Clinical Impact

Myocardial perfusion: from popular to largely retired

The best-known dual-isotope protocol is rest Tl-201 / stress Tc-99m sestamibi myocardial perfusion SPECT. Its appeal was throughput and a combined viability-plus-perfusion assessment in one sitting, and it was widely adopted.¹⁰¹¹ But it carries a higher patient radiation dose than an all-technetium study because Tl-201 has a long physical half-life and a relatively high absorbed dose per unit activity. As dose-reduction became a priority and high-sensitivity CZT cameras made low-dose technetium imaging routine, guidance moved away from dual-isotope myocardial perfusion toward single-isotope technetium protocols.¹⁰² The physics lesson endures even where the protocol has faded: the Tl-201 window was always contaminated by Tc-99m downscatter, and reported results depended on how well that was handled.¹

Parathyroid subtraction: a durable dual-tracer application

Parathyroid subtraction scintigraphy remains a mainstream dual-tracer technique. A thyroid agent (I-123 or Tc-99m pertechnetate) marks the thyroid, while Tc-99m sestamibi accumulates in both thyroid and hyperfunctioning parathyroid tissue; subtracting the thyroid image isolates the parathyroid. Pooled data show sestamibi/I-123 subtraction with higher sensitivity for abnormal glands than older Tl-201 subtraction (about 87 percent versus 71 percent), and subtraction remains useful when sestamibi dual-phase washout is ambiguous — in one series, correlative subtraction imaging was positive in roughly 88 percent of rapid-washout adenomas.⁷⁸ Here the crosstalk is spillover between the adjacent 140 and 159 keV windows, and careful window placement plus spillover correction is what makes the subtraction clean.

Quantitative imaging and theranostic dosimetry

The newest driver is quantitative dual-isotope imaging. Simultaneous Ho-166/Tc-99m liver SPECT lets a single acquisition estimate both tumor-absorbed dose (from the therapeutic Ho-166 microspheres) and healthy-liver uptake (from Tc-99m), provided the downscatter is corrected.⁶ As theranostics expands, the ability to image a therapeutic isotope and a physiologic marker together — accurately — is exactly the capability a quantitative SPECT program must validate. See quantitative SPECT/CT calibration for the calibration chain that makes any of this defensible.

Practical Optimization Tips

A dual-isotope protocol should be built, not borrowed. A practical setup sequence:

1. Choose a separable isotope pair

Favor pairs with well-separated photopeaks (Tl-201/Tc-99m) over closely spaced ones (Tc-99m/I-123) unless a solid-state camera with fine energy resolution is available. Confirm the higher-energy isotope's high-energy emissions and select the collimator that controls septal penetration.

2. Set and document the energy windows

Place a main photopeak window on each isotope and, if using TEW, add the narrow sub-windows. Record the exact window bounds — they are part of the protocol and must match the correction the reconstruction applies.

3. Pick the activity ratio deliberately

Set the ratio so the weaker-window isotope retains adequate counts after downscatter. Validate the choice with a phantom containing both isotopes at the intended ratio.

4. Select and validate a crosstalk correction

Decide between energy-window (TEW) and model-based/Monte Carlo correction based on the magnitude of the crosstalk and the quantitative goal. Then validate it: image a dual-isotope phantom, compare each isotope's recovered activity against a single-isotope acquisition of the same phantom, and document the agreement.

5. Re-validate after any change

Changing collimator, camera, activity ratio, or reconstruction can change the crosstalk. Treat a protocol change as a reason to repeat the phantom validation.

Common pitfalls to avoid

• Treating the window counts as the image. Raw window counts are contaminated; the corrected counts are the image.
• Copying a NaI(Tl) protocol onto close photopeaks. Tc-99m/I-123 on a conventional camera without adequate spillover correction yields poor separation.
• Ignoring the activity ratio. A dominant higher-energy isotope can make the weaker isotope unrecoverable.
• Skipping phantom validation. Clinical images cannot confirm quantitative accuracy; a known phantom can.
• Forgetting the collimator. High-energy emissions from I-123 or other isotopes penetrate low-energy collimator septa and add broad crosstalk.

Regulatory Considerations

Dual-isotope imaging does not change the regulatory basis of nuclear medicine, but it does raise the stakes on written directives, dosimetry, and quality control. Each radionuclide is byproduct material administered under the facility's medical-use authorization.

• 10 CFR Part 35 — Medical Use of Byproduct Material governs authorized use, dosage determination, and the records for each radiopharmaceutical administered; a dual-isotope study administers two.¹²
• SNMMI/ASNC SPECT myocardial perfusion imaging guidelines (2018) describe instrumentation, acquisition, processing, and interpretation expectations for SPECT, including camera quality control that underpins any multi-isotope protocol.²
• IAEA Human Health Series No. 9, Quality Assurance for SPECT Systems (2009) sets out the acceptance and routine QC — energy-window calibration, uniformity, and energy resolution — on which reliable dual-isotope separation depends.¹³
• NEMA NU 1, Performance Measurements of Gamma Cameras, defines how energy resolution, spatial resolution, and count-rate performance are measured and reported, the camera parameters that determine whether a given isotope pair is even feasible.¹⁴

Agreement States administer their own medical-use programs. Of the states DRPS serves, Florida, Maryland, Virginia, California, Nevada, Pennsylvania, New York, and New Jersey are NRC Agreement States, while Washington, DC is regulated directly by the NRC. A facility must verify which authority issues its license and which QC and recordkeeping requirements apply. For the program elements that surround this, see quality control for cardiac SPECT and the facility-level framework in NRC Part 35 recordkeeping requirements.

Frequently Asked Questions (FAQs)

What is dual-isotope SPECT imaging?

Dual-isotope SPECT is a single-photon imaging technique that uses two radionuclides with different photon energies in the same patient, either injected and imaged at the same time (simultaneous acquisition) or in back-to-back acquisitions (separate acquisition). Separate energy windows are set on each isotope's photopeak so the camera can form a distinct image of each tracer.

What is crosstalk in dual-isotope SPECT?

Crosstalk is unwanted counts from one radionuclide appearing in the energy window of the other. It comes mainly from downscatter (higher-energy photons that Compton-scatter down into the lower-energy window), photopeak spillover caused by finite detector energy resolution, and collimator septal penetration by high-energy emissions. Uncorrected crosstalk degrades contrast and biases quantification.

Does simultaneous dual-isotope imaging require energy-window corrections?

Yes. Because the two photopeaks and their scatter overlap, a simultaneous dual-isotope study is not quantitatively reliable without a crosstalk or scatter correction such as the triple-energy-window method, a model-based estimate, or a Monte Carlo downscatter simulation validated on that camera.

Why does detector energy resolution matter for dual-isotope SPECT?

Finer energy resolution produces narrower photopeaks, so less of one isotope's photopeak spills into the adjacent window. Conventional NaI(Tl) cameras have roughly 9 to 10 percent energy resolution at 140 keV, while solid-state CZT or CdTe detectors reach about 5 to 6 percent, which makes closely spaced pairs such as Tc-99m and I-123 more separable.

What are common clinical uses of dual-isotope SPECT?

Historical and current examples include rest thallium-201 with stress Tc-99m sestamibi myocardial perfusion imaging, I-123 or Tc-99m pertechnetate plus Tc-99m sestamibi parathyroid subtraction imaging, and simultaneous Tc-99m and I-123 brain studies. Dual-isotope myocardial perfusion has largely been replaced by lower-dose single-isotope technetium protocols.

Is dual-isotope SPECT quantitatively accurate?

With a validated crosstalk correction it can be. Phantom and patient studies report agreement within a few percent to about 10 to 15 percent between dual-isotope and single-isotope acquisitions, depending on the isotope pair, activity ratio, detector, and correction method. The physicist should verify quantitative accuracy on the specific system before clinical reliance.

When should a facility involve a medical physicist in a dual-isotope protocol?

Before clinical use and whenever the isotope pair, activity ratio, energy windows, collimator, or reconstruction changes. The physicist sets and documents the energy windows, selects and validates the crosstalk correction, confirms quantitative accuracy with phantoms, and aligns the protocol with camera performance and accreditation requirements.

Key Takeaways

• Crosstalk is the defining problem. Downscatter, photopeak spillover, and septal penetration always contaminate the two windows; the correction is what makes the study usable.
• Separation depends on photopeak spacing and detector resolution. Tl-201/Tc-99m (70 vs 140 keV) is forgiving; Tc-99m/I-123 (140 vs 159 keV) demands fine energy resolution or strong spillover correction.
• TEW is fast and model-free; Monte Carlo is accurate. The triple-energy-window estimate subtracts a trapezoidal scatter approximation; model-based reconstruction simulates downscatter from the higher-energy isotope.
• Activity ratio is a design choice. Set it so the weaker-window isotope keeps adequate statistics after downscatter.
• Validate on a phantom, then re-validate after changes. Clinical images cannot prove quantitative accuracy; a known dual-isotope phantom can.
• The clinical landscape is shifting. Dual-isotope myocardial perfusion has largely given way to low-dose technetium imaging, while parathyroid subtraction and quantitative theranostic dual-isotope imaging remain important.

Conclusion

Dual-isotope SPECT trades a harder physics problem for a real clinical advantage: two co-registered tracer images from one acquisition. The advantage is only real if the crosstalk is handled. That means choosing isotopes whose photopeaks the camera can separate, setting energy windows that match the correction, selecting an activity ratio that protects the weaker isotope, and validating the whole chain on a phantom before any patient is scanned.

The role of the medical physicist is to own that chain end to end — to decide whether a triple-energy-window subtraction is enough or a model-based downscatter correction is required, to document the windows and the validation, and to re-check the protocol whenever the hardware or reconstruction changes. Done that way, a dual-isotope study is quantitative. Done casually, it is two contaminated images dressed up as one clever acquisition.

How DRPS Can Help

Diagnostic Radiation Physics Services helps nuclear medicine and cardiology programs design, validate, and document dual-isotope and quantitative SPECT protocols. That includes energy-window and collimator selection, crosstalk-correction validation with dual-isotope phantoms, camera energy-resolution and uniformity QC, quantitative SPECT/CT calibration, and alignment with SNMMI/ASNC guidance and accreditation requirements, delivered through our PET/CT and nuclear medicine physics and medical physicist 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 dual-isotope protocol is a physics commitment, not just a clinical preference — and it should be set up and verified like one.

Related Resources

• Common PET and radiopharmaceutical-therapy isotopes
• SPECT scatter correction
• SPECT/CT attenuation correction
• Quantitative SPECT/CT calibration
• Cardiac SPECT myocardial perfusion quality control
• Gamma camera energy resolution QC
• PET/CT and nuclear medicine physics services
• Medical physicist consulting

References

1. Okudan B, Smitherman TC. The value and throughput of rest Thallium-201/stress Technetium-99m sestamibi dual-isotope myocardial SPECT. Anadolu Kardiyol Derg. 2004;4(2):161-168. PubMed
2. Dorbala S, Ananthasubramaniam K, Armstrong IS, et al. Single Photon Emission Computed Tomography (SPECT) Myocardial Perfusion Imaging Guidelines: Instrumentation, Acquisition, Processing, and Interpretation. J Nucl Cardiol. 2018;25(5):1784-1846. doi:10.1007/s12350-018-1283-y. doi.org
3. International Atomic Energy Agency. Nuclear Medicine Physics: A Handbook for Teachers and Students. Vienna: IAEA; 2014. iaea.org
4. Takeuchi W, Suzuki A, Shiga T, et al. Simultaneous Tc-99m and I-123 dual-radionuclide imaging with a solid-state detector-based brain-SPECT system and energy-based scatter correction. EJNMMI Phys. 2016;3(1):10. doi:10.1186/s40658-016-0147-2. doi.org
5. Kangasmaa T, Kuikka J, Sohlberg A. Optimisation of simultaneous Tl-201/Tc-99m dual isotope reconstruction with Monte-Carlo-based scatter correction. Int J Mol Imaging. 2012;2012:695632. doi:10.1155/2012/695632. doi.org
6. van Rooij R, Braat AJAT, de Jong HWAM, Lam MGEH. Simultaneous Ho-166/Tc-99m dual-isotope SPECT with Monte Carlo-based downscatter correction for automatic liver dosimetry in radioembolization. EJNMMI Phys. 2020;7(1):13. doi:10.1186/s40658-020-0280-9. doi.org
7. McBiles M, Lambert AT, Cote MG, Kim SY. Sestamibi parathyroid imaging. Semin Nucl Med. 1995;25(3):221-234. doi:10.1016/s0001-2998(95)80012-3. doi.org
8. Kushchayeva YS, Tella SH, Kushchayev SV, Van Nostrand D, Kulkarni K. Comparison of hyperparathyroidism types and utility of dual radiopharmaceutical acquisition with Tc-99m sestamibi and I-123 for localization of rapid washout parathyroid adenomas. Osteoporos Int. 2019;30(5):1051-1057. doi:10.1007/s00198-019-04846-6. doi.org
9. Ogawa K, Harata Y, Ichihara T, Kubo A, Hashimoto S. A practical method for position-dependent Compton-scatter correction in single photon emission CT. IEEE Trans Med Imaging. 1991;10(3):408-412. doi:10.1109/42.97591. doi.org
10. Hachamovitch R. Clinical application of rest thallium-201/stress technetium-99m sestamibi dual isotope myocardial perfusion single-photon emission computed tomography. Cardiol Rev. 1999;7(2):83-91. doi:10.1097/00045415-199903000-00011. doi.org
11. Williams KA, Hill KA, Sheridan CM. Noncardiac findings on dual-isotope myocardial perfusion SPECT. J Nucl Cardiol. 2003;10(4):395-402. doi:10.1016/s1071-3581(03)00524-5. doi.org
12. U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. ecfr.gov
13. International Atomic Energy Agency. Quality Assurance for SPECT Systems. IAEA Human Health Series No. 9. Vienna: IAEA; 2009. iaea.org
14. National Electrical Manufacturers Association. NEMA Standards Publication NU 1: Performance Measurements of Gamma Cameras. Rosslyn, VA: NEMA. nema.org