Contrast-Enhanced Mammography: Physics and QC

Introduction

Contrast-enhanced mammography (CEM) is a dual-energy technique that adds a functional, iodine-based image to conventional mammography by acquiring a low-energy and a high-energy exposure in a single compression and subtracting them. The low-energy image looks like a standard 2D mammogram; the recombined image highlights areas of iodinated contrast uptake, giving the study a vascular, MRI-like character while remaining an x-ray examination.¹²

For a medical physicist, CEM is neither "just mammography" nor a wholly new modality. It is a dual-energy acquisition layered on top of a full-field digital mammography (FFDM) system, and that dual nature is exactly what shapes the quality control (QC), dosimetry, and regulatory picture. The low-energy exposure inherits the entire Mammography Quality Standards Act (MQSA) framework and the ACR Digital Mammography QC program. The recombined image, however, depends on vendor-specific spectra, filters, and subtraction weighting, so its QC follows manufacturer procedures.³⁷

This guide walks through the underlying physics around the iodine K-edge, the dose penalty that comes with the second exposure, the contrast-to-noise ratio (CNR) and figure-of-merit (FOM) metrics that anchor CEM QC, worked calculations a physicist can reproduce, and the regulatory context that governs a CEM program. DRPS supports CEM programs as part of its mammography physics and MQSA and broader medical physics consulting services.

Topic Explanation

What is contrast-enhanced mammography?

CEM is performed after an intravenous injection of iodinated contrast, using a dual-energy acquisition to separate iodine signal from the surrounding fibroglandular and adipose tissue. A typical protocol injects an iodinated agent (commonly around 1.5 mL/kg of a roughly 300 mgI/mL solution), waits approximately two minutes for early enhancement, and then images each breast. Within one compression, the system fires two exposures: a low-energy (LE) exposure whose spectrum sits mostly below the iodine K-edge, and a high-energy (HE) exposure whose spectrum extends above it.¹²

The LE image is diagnostically equivalent to a standard 2D digital mammogram and is read as such. The HE image is not read directly; instead, the two exposures are combined by weighted logarithmic subtraction to produce the recombined (also called iodine or ES, energy-subtracted) image, in which iodine-avid lesions appear bright against a suppressed tissue background.¹² Because both images are acquired in the same compression, they are inherently registered, avoiding much of the motion misregistration that plagued earlier temporal subtraction approaches.

CEM goes by several names in the literature and on vendor platforms — contrast-enhanced spectral mammography (CESM), contrast-enhanced digital mammography (CEDM) — but the physics is the same dual-energy, K-edge-driven approach.² For the standard (non-contrast) side of the program, our guides to mammography quality control under MQSA and mean glandular dose in mammography cover the foundation that CEM builds on.

Why acquire two energies?

Standard mammography struggles when dense fibroglandular tissue overlaps and masks a lesion. Contrast enhancement sidesteps that problem by imaging physiology — the neovascularity of many cancers — rather than relying solely on morphology. The dual-energy trick is what makes the iodine signal separable from tissue: iodine's attenuation changes abruptly at its K-edge, while soft-tissue attenuation varies smoothly with energy. Subtracting a high-energy image (weighted) from a low-energy image cancels most of the smoothly varying tissue background and leaves the K-edge-discontinuous iodine behind.¹²

Key Technical Principles

The iodine K-edge and spectral separation

Iodine (Z = 53) has a K-absorption edge at 33.2 keV. Just above this energy, iodine's mass attenuation coefficient rises sharply, so photons above the K-edge are absorbed far more strongly by iodine than photons just below it. CEM is engineered around that discontinuity.¹²

• The low-energy spectrum is a conventional mammographic beam (for example, ~26–32 kVp with Rh or Ag targets/filters) whose photons sit largely below 33.2 keV.
• The high-energy spectrum uses a higher tube potential (typically ~45–49 kVp) plus additional filtration (copper, titanium, or thicker rhodium) to push a meaningful fraction of the beam above the K-edge and harden it.

The table below summarizes the two acquisitions.

Weighted logarithmic subtraction

The recombined signal is formed by a weighted logarithmic subtraction of the two exposures. In simplified form, for measured signals $I_{LE}$ and $I_{HE}$ at a pixel:

$$S_{\text{recomb}}=\ln \left(I_{LE}\right)-w\cdot \ln \left(I_{HE}\right)$$

The weighting factor $w$ is chosen so that the difference in the logarithmic attenuation of adipose and glandular tissue cancels, i.e. the background becomes approximately flat:

$$w=\frac{{\mu }_{g,LE}-{\mu }_{a,LE}}{{\mu }_{g,HE}-{\mu }_{a,HE}}$$

where ${\mu }_g$ and ${\mu }_a$ are the effective linear attenuation coefficients of glandular and adipose tissue at the two energies. With the background suppressed, the residual signal is dominated by the K-edge attenuation of iodine, which does not cancel because its energy dependence differs from that of soft tissue.¹² Vendors implement proprietary variants of this recombination, which is precisely why the recombined-image QC is manufacturer-specific.⁷

Image-quality metrics: CNR and figure of merit

The workhorse metric for CEM performance is the iodine contrast-to-noise ratio in the recombined image. For an iodine insert of mean signal ${\bar{S}}_I$ against background ${\bar{S}}_{bg}$ with background noise ${\sigma }_{bg}$:

$$\mathrm{CNR}=\frac{{\bar{S}}_I-{\bar{S}}_{bg}}{{\sigma }_{bg}}$$

Because CNR can always be improved by increasing dose, physicists normalize to dose using a figure of merit:

$$\mathrm{FOM}=\frac{{\mathrm{CNR}}^2}{D_g}$$

where $D_g$ is the mean glandular dose of the study. FOM rewards spectra and technique factors that deliver iodine contrast efficiently rather than simply by adding dose, and it is the natural quantity to optimize when comparing target/filter and kVp combinations.¹⁴ Phantom studies consistently show CNR increasing with iodine concentration and decreasing with breast thickness — behavior that a CEM phantom with graded iodine inserts is designed to verify.³⁴

Mean glandular dose in CEM

CEM's dose is the sum of the LE and HE contributions:

$$D_{g,\text{view}}=D_{g,LE}+D_{g,HE}$$

As a representative worked example, suppose a single view delivers $D_{g,LE}\approx 0.8\text{mGy}$ and $D_{g,HE}\approx 0.4\text{mGy}$, so:

$$D_{g,\text{view}}\approx 0.8+0.4=1.2\text{mGy}$$

A bilateral two-view study (four views) then delivers approximately:

$$D_{g,\text{study}}\approx 4\times 1.2\approx 4.9\text{mGy}$$

That figure is on the order of 30% higher than a comparable two-view digital mammogram, consistent with the added HE exposure.⁴ Phantom measurements across vendors show per-view average glandular doses that vary widely with thickness — from well under 1 mGy for thin phantoms to several mGy for thick, dense configurations — and the HE exposure accounting for roughly 24–42% of the total, depending on system and thickness.⁴ The individual values above are illustrative; a facility must measure its own MGD as part of the physicist survey rather than adopt a generic number.

Clinical Impact

CEM's value proposition is functional imaging at the accessibility and cost of mammography. By showing enhancement, CEM improves lesion conspicuity in dense breasts, helps characterize findings seen on standard mammography or ultrasound, and supports staging and problem-solving in the diagnostic setting. Recent syntheses of the literature position CEM as approaching the sensitivity of contrast-enhanced breast MRI for many indications, while being faster and more available — with the caveat that its optimal role relative to MRI, tomosynthesis, and ultrasound is still being defined.⁵

For QC purposes, the clinical stakes cut two ways. First, because the LE image is read as a diagnostic mammogram, any degradation in LE image quality directly affects standard interpretation — so CEM does not relax the ACR Digital Mammography QC obligations, it adds to them.³⁷ Second, the recombined image drives contrast-based decisions, so instability in the dual-energy chain (spectral drift, detector lag between LE and HE, recombination artifacts) can mimic or mask enhancement. A physicist evaluating CEM has to protect both the morphologic (LE) and functional (recombined) information streams. For the tomosynthesis comparison many facilities weigh alongside CEM, see our digital breast tomosynthesis QC guide.

A specific CEM artifact worth flagging is detector lag / ghosting between the paired exposures: residual signal from the LE exposure can contaminate the HE exposure (or vice versa), producing recombination artifacts that look like spurious enhancement or edge halos. Because the two exposures occur milliseconds to a couple of seconds apart, lag behavior is a legitimate QC concern unique to the dual-energy workflow, and it connects to the same detector physics discussed in digital radiography lag and ghosting QC.

Practical Optimization Tips

A defensible CEM QC program layers CEM-specific tests on top of the standard mammography QC foundation.

1. Keep the low-energy image fully MQSA-compliant

The LE image is a diagnostic mammogram. All routine ACR Digital Mammography QC tests — phantom image quality, signal-to-noise and CNR, AEC performance and reproducibility, artifact evaluation, and the annual physicist survey — apply unchanged.³⁷ Do not let "CEM QC" become a reason to under-test the 2D chain.

2. Verify the recombined image with a CEM phantom

Use a dedicated CEM phantom containing inserts of known iodine areal concentration (for example, spanning roughly 0.5–2 mgI/cm²) embedded in breast-equivalent material at multiple thicknesses. Measure:

• Iodine CNR at each concentration and thickness.
• Linearity of recombined signal versus iodine concentration (a strong linear response is expected and is a sensitive indicator of recombination health).⁴
• FOM (CNR²/MGD) to track dose efficiency over time.

3. Trend, don't just pass/fail

Establish baselines at acceptance and trend CNR, linearity slope, and MGD. A gradual CNR decline or linearity drift at fixed technique often signals spectral or detector drift before it becomes visible clinically.

4. Test AEC behavior at both energies

CEM AEC must select appropriate LE and HE techniques across thickness and density. Confirm that the AEC produces adequate iodine CNR without runaway dose at thick/dense settings, where the HE contribution grows.⁴

5. Evaluate lag/ghosting and recombination artifacts

Include a check for residual-signal ghosting between paired exposures and for recombination artifacts (skin-line halos, motion residue). These are dual-energy-specific and are not captured by standard 2D QC.

6. Document contrast workflow and safety

CEM involves an iodinated contrast injection, so the program must integrate contrast screening (renal function, allergy history), a contrast reaction plan, and staff training — clinical-safety elements a physicist should confirm exist even though they fall outside the imaging chain.

Common pitfalls

• Assuming ACR covers everything. The recombined image is manufacturer-QC territory.⁷
• Adopting a generic MGD. Measure facility- and thickness-specific dose.⁴
• Ignoring the LE image. It is still a diagnostic mammogram under MQSA.³
• Skipping linearity. CNR alone can look acceptable while the concentration response drifts.⁴

Regulatory Considerations

CEM sits at the intersection of MQSA device regulation, drug regulation for the contrast agent, and manufacturer QC for the dual-energy application. Getting the framework right matters for both compliance and defensibility.

• MQSA / 21 CFR Part 900. CEM is performed on an FDA-approved FFDM system. That unit — and the low-energy imaging it produces — is subject to the Mammography Quality Standards Act, including facility certification, accreditation, and the annual survey by a qualified medical physicist.⁶⁷ The 2023 MQSA Final Rule, with enforcement beginning September 10, 2024, modernized the regulations (including breast-density reporting) but did not create a separate CEM accreditation module; CEM's 2D component is handled through the existing digital mammography pathway.⁶
• Manufacturer QC for the contrast application. The FDA's position is that facilities should follow the manufacturer's QC procedures for contrast-enhancement (recombined) imaging, since the dual-energy acquisition and recombination are vendor-specific and are not covered by the ACR Digital Mammography QC Manual.⁷
• The contrast agent is a drug. Intravenous iodinated contrast is regulated as a pharmaceutical; its administration falls under the facility's medical and pharmacy governance, contrast-reaction protocols, and (for many CEM indications) off-label-use considerations rather than MQSA.
• X-ray machine and state oversight. The mammography unit is also a radiation-producing device subject to FDA performance standards (21 CFR 1020.30) and to state radiation-control programs. For imaging (x-ray) devices, jurisdiction is FDA plus the state — distinct from the NRC's authority over byproduct material.

Because DRPS serves facilities across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware, the practical guidance is to confirm the state radiation-control requirements alongside the federal MQSA obligations, and to keep the physicist's CEM survey documentation aligned with both the ACR Digital Mammography QC Manual (LE image) and the vendor QC manual (recombined image).³⁷ Facilities pursuing or maintaining accreditation should fold CEM into their broader accreditation support planning; see also our overview of ACR accreditation physics requirements.

Frequently Asked Questions (FAQs)

What is contrast-enhanced mammography (CEM)?

CEM is a dual-energy mammographic technique performed after an intravenous iodinated contrast injection. Each view acquires a low-energy exposure that resembles a standard 2D mammogram and a high-energy exposure above the iodine K-edge. Weighted subtraction of the two produces a recombined image that shows iodine uptake, similar in concept to the enhancement seen on breast MRI.

How much more dose does CEM deliver than a standard mammogram?

The added high-energy exposure raises the mean glandular dose. Published work generally places a full bilateral two-view CEM study on the order of about 30% higher mean glandular dose than the equivalent standard digital mammogram, with the high-energy exposure contributing roughly a quarter to two-fifths of the total per view depending on breast thickness and system. Facility-specific values should be measured, not assumed.

Does the ACR Digital Mammography QC Manual cover CEM?

The ACR Digital Mammography QC Manual governs the low-energy (2D) image, which is acquired the same way as a standard digital mammogram. The FDA position is that facilities should follow the manufacturer's quality control procedures for the contrast-enhancement (recombined) application, because the dual-energy acquisition and recombination are vendor-specific.

Is CEM regulated under MQSA?

Yes, in part. CEM is performed on an FDA-approved full-field digital mammography system, so the unit and its low-energy imaging fall under the Mammography Quality Standards Act (21 CFR Part 900), including the annual medical physicist survey. The iodinated contrast agent is a separately regulated drug, and the recombined-image QC follows manufacturer procedures rather than a dedicated MQSA CEM standard.

What metrics does a physicist use to evaluate CEM image quality?

The core quantitative metrics are iodine contrast-to-noise ratio (CNR) in the recombined image, the linearity of measured signal with iodine concentration, and a figure of merit that normalizes CNR-squared to dose. These are measured with a dedicated CEM phantom containing known iodine-concentration inserts at several phantom thicknesses.

Why is the iodine K-edge important in CEM?

Iodine has a K-edge at 33.2 keV, where its x-ray attenuation jumps sharply. CEM exploits this by acquiring one image with most photons below the K-edge and one with a substantial fraction above it, then subtracting. The differential attenuation of iodine across the K-edge is what makes iodine stand out while glandular and adipose background largely cancels.

Can CEM replace breast MRI?

CEM and MRI both rely on contrast enhancement and neovascularity, and CEM is faster, lower cost, and more accessible. Current evidence positions CEM as a strong problem-solving and workup tool and a reasonable MRI alternative when MRI is contraindicated or unavailable, but the best use of CEM relative to MRI, ultrasound, and tomosynthesis remains an active clinical question rather than a settled substitution.

Key Takeaways

• CEM is dual-energy imaging around the iodine K-edge (33.2 keV). A low-energy image (read as a standard mammogram) and a high-energy image are combined by weighted subtraction into an iodine-only recombined image.¹²
• The dose penalty is real but modest. A bilateral two-view CEM study runs on the order of ~30% higher mean glandular dose than standard digital mammography, driven by the added HE exposure.⁴
• QC is layered. The LE image stays fully under ACR Digital Mammography QC and MQSA; the recombined image follows manufacturer QC.³⁷
• Measure CNR, linearity, and FOM. A CEM phantom with graded iodine inserts is the core tool; trend the results rather than only checking pass/fail.⁴
• Watch dual-energy-specific artifacts. Detector lag/ghosting and recombination artifacts can mimic or mask enhancement and are not caught by 2D QC.
• Know the regulatory split. MQSA governs the unit and 2D image; the contrast agent is a drug; the recombined-image QC is vendor-specified.⁶⁷

Conclusion

Contrast-enhanced mammography is one of the most practical ways to bring functional, enhancement-based breast imaging into a mammography suite without the cost and access barriers of MRI. But its dual-energy nature means the physics — spectral separation across the iodine K-edge, weighted logarithmic recombination, and a measurable dose penalty — has direct QC consequences. A strong CEM program keeps the low-energy image fully compliant with MQSA and the ACR Digital Mammography QC Manual, adds recombined-image testing per the manufacturer's procedures, and trends CNR, linearity, FOM, and dose over time. Treated that way, CEM is a well-characterized, defensible addition to the diagnostic toolkit rather than an under-tested bolt-on.

How DRPS Can Help

Diagnostic Radiation Physics Services helps breast imaging facilities implement and maintain defensible CEM programs. That support includes acceptance and annual mammography physics surveys under MQSA, CEM-specific recombined-image testing (iodine CNR, linearity, and figure of merit), mean-glandular-dose measurement, AEC evaluation across the dual-energy chain, artifact and lag assessment, and accreditation support that keeps ACR Digital Mammography QC and manufacturer CEM QC aligned.

DRPS serves facilities across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware. See our service locations or contact us to discuss a CEM QC program.

Related Resources

• Mammography quality control under MQSA
• Mean glandular dose in mammography
• Digital breast tomosynthesis QC
• Digital radiography lag and ghosting QC
• ACR accreditation physics requirements
• Mammography physics and MQSA services
• Accreditation support

References

1. Covington MF, Salmon S, Weaver BD, Fajardo LL. State-of-the-art for contrast-enhanced mammography. Br J Radiol. 2024;97(1156):695-704. doi:10.1093/bjr/tqae017. doi.org
2. Fallenberg EM. Contrast-enhanced mammography. Radiologe. 2021;61(2):177-182. doi:10.1007/s00117-021-00805-7. doi.org
3. American College of Radiology. Digital Mammography Quality Control Manual. Reston, VA: ACR. acr.org
4. Bruschi G, Ricciardi V, De Marco P, Origgi D. Phantom-based comparative analysis of contrast-enhanced mammography systems: image quality and performance evaluation. J Appl Clin Med Phys. 2025;26(7):e70163. doi:10.1002/acm2.70163. doi.org
5. Niroshani S, Nakamura T, Michiru N, Negishi T. Evaluation of exposure factors of dual-energy contrast-enhanced mammography to optimize radiation dose with improved image quality. Acta Radiol Open. 2022;11(8):20584601221117251. doi:10.1177/20584601221117251. doi.org
6. U.S. Food and Drug Administration. Mammography Quality Standards Act (MQSA) and MQSA Program; Final Rule to Amend MQSA (enforcement began September 10, 2024). fda.gov
7. U.S. Food and Drug Administration. Mammography Quality Standards; Policy Guidance and QC for Digital Mammography and Advanced Applications (facilities follow manufacturer QC for contrast-enhancement applications). fda.gov
8. Endarko, Celina FM, Gani MRA. Analysis of dual-energy mammography subtraction technique for the dose and image quality evaluation using 3D-printed breast phantom. Phys Eng Sci Med. 2023;46(4):1693-1701. doi:10.1007/s13246-023-01330-8. doi.org
9. Heck L, Dierolf M, Jud C, et al. Contrast-enhanced spectral mammography with a compact synchrotron source. PLoS One. 2019;14(10):e0222816. doi:10.1371/journal.pone.0222816. doi.org
10. Jakubiak RR, Gamba HR, Neves EB, Peixoto JE. Image quality, threshold contrast and mean glandular dose in CR mammography. Phys Med Biol. 2013;58(18):6565-6583. doi:10.1088/0031-9155/58/18/6565. doi.org