Photon-Counting CT: Image Quality and Dose

Photon-counting detector CT (PCD-CT) counts and energy-sorts individual X-ray photons instead of integrating a summed signal, which lowers electronic noise, improves spatial resolution, strengthens iodine contrast, and makes spectral data available on every scan — often with a radiation-dose benefit. Realizing those benefits, however, requires acceptance testing, protocol design, and quality control that account for physics conventional CT never had. ¹²

Since the first photon-counting CT system was cleared for clinical use in 2021, the technology has moved from research benches into routine radiology. ³⁹ For imaging facilities, that raises practical questions: what actually changes at the detector, where does image quality improve, how much dose can be saved, and what does the medical physicist need to verify before the scanner goes live. This guide answers those questions and connects them to CT physics testing and protocol optimization.

Introduction

For its entire clinical history, CT has relied on energy-integrating detectors (EIDs). In an EID, X-rays strike a scintillator, are converted to visible light, and that light is measured by a photodiode. The signal from every photon in a measurement interval is summed together — so a low-energy photon and a high-energy photon contribute in proportion to their energy, and information about individual photons is lost. Reflective septa between detector elements prevent optical cross-talk but occupy space and reduce geometric dose efficiency. ¹²

A photon-counting detector (PCD) works differently. A semiconductor such as cadmium telluride (CdTe) or cadmium zinc telluride (CZT) directly converts each X-ray into a cloud of electron-hole pairs, producing an electrical pulse whose height is proportional to the photon's energy. Fast readout electronics count each pulse and, using one or more energy thresholds, sort photons into energy bins. ¹³⁸ This one-step, direct-conversion process removes the scintillator, removes the need for reflective septa, and — critically — allows the system to reject electronic noise by setting the lowest threshold above the noise floor. ²³

The consequences are substantial: essentially no electronic noise, higher achievable spatial resolution, improved iodine signal, and spectral information available for every acquisition. ¹²⁴ But "different detector" also means "different QC." A medical physicist evaluating a PCD-CT system cannot simply reuse an EID checklist unchanged. This guide walks through the physics, the measurable image-quality and dose effects, the clinical value, practical optimization, and the regulatory and accreditation context.

Topic Explanation

What is photon-counting CT?

Photon-counting CT is a form of computed tomography in which the detector counts individual X-ray photons and measures each photon's energy, rather than integrating the total energy deposited over a readout interval. Because energy is recorded photon-by-photon, the scanner produces spectral (multi-energy) data intrinsically, not as a special mode. ¹⁶

Key terms used throughout this guide:

Energy-integrating detector (EID) — the conventional scintillator-plus-photodiode detector that measures a summed, energy-weighted signal.
Photon-counting detector (PCD) — a direct-conversion semiconductor detector that counts individual photons and bins them by energy.
Energy threshold / energy bin — an electronic level that separates counted photons into low- and high-energy groups; multiple thresholds create multiple bins.
Virtual monoenergetic image (VMI) — a reconstruction representing the image at a single selected photon energy (keV), derived from spectral data.
Ultra-high-resolution (UHR) mode — an acquisition/reconstruction mode that exploits the small effective detector-element size of PCDs to resolve fine detail.

Why the detector change matters

In an EID, two physical factors limit performance. First, electronic noise adds to the signal, which is most damaging at low dose and in large patients where the transmitted signal is weak. Second, the summed signal weights photons by energy, so lower-energy photons — which carry more soft-tissue and iodine contrast — contribute less than their information content would justify. ¹²

A PCD addresses both. By counting photons above a noise-rejection threshold, it eliminates electronic-noise contribution to the image. By counting each photon roughly equally (rather than energy-weighting), it recovers contrast that an EID underweights, improving the iodine contrast-to-noise ratio. And because detector elements can be made small without optical-septa penalties, the intrinsic spatial resolution improves. ¹²⁴ For background on how resolution and low-contrast detectability are quantified on any CT system, see our guide to CT image quality: MTF and low-contrast detectability.

Key Technical Principles

Contrast-to-noise ratio and the dose relationship

Image quality in CT is often summarized by the contrast-to-noise ratio (CNR), the separation between two materials relative to image noise:

CNR = \frac{∣ μ _{A} - μ _{B} ∣}{σ}

where $μ_{A}$ and $μ_{B}$ are the mean CT numbers of two regions and $σ$ is the noise standard deviation. In the quantum-noise-limited regime, noise scales inversely with the square root of the number of detected photons, and therefore with dose $D$ :

σ \propto \frac{1}{D} \Rightarrow D \propto \frac{1}{σ ^{2}} \propto \frac{1}{CNR ^{2}}

This inverse-square relationship is why even a modest CNR improvement can translate into a substantial dose reduction. If a photon-counting system achieves the same task at a higher CNR, the dose required to merely match the conventional-detector CNR falls as the square of the CNR ratio: ¹²

\frac{D _{PCD}}{D _{EID}} \approx (\frac{CNR _{EID}}{CNR _{PCD}})^{2}

Worked illustration. In a clinical non-contrast chest comparison, a photon-counting system produced a CNR of 59.2 versus 53.3 on an energy-integrating system. ⁵ Treating the exam as noise-limited, the dose needed for the PCD to reproduce the EID's CNR would be approximately:

D \approx D_{PCD} (\frac{53.3}{59.2})^{2} = D_{PCD} \times 0.81

i.e., roughly a further 19% below the already-reduced PCD dose. This is an idealized, single-factor estimate — real protocols involve reconstruction, patient size, and task — but it explains the direction and magnitude of the benefit. In that same study the measured median CTDIvol was 4.71 mGy for the photon-counting scanner versus 7.80 mGy for the energy-integrating scanner, with median DLP of 182.0 versus 262.6 mGy·cm, a dose reduction of roughly 40% while CNR improved. ⁵

Spectral separation and virtual monoenergetic images

Because a PCD sorts photons into energy bins, every scan carries two (or more) spectral measurements. Material decomposition then expresses the image as a combination of basis materials (for example, water and iodine), enabling iodine maps, virtual non-contrast images, and virtual monoenergetic images (VMI) at a chosen energy $E$ . ¹⁶⁷ Low-keV VMI (around 40–50 keV) sit near the iodine K-edge (33.2 keV) and boost iodine contrast; high-keV VMI (around 100–190 keV) suppress beam-hardening and metal artifact. The spectral framework is the same one described for dual-energy systems in AAPM Task Group 291, but a photon-counting detector delivers it on a single, always-on acquisition. ⁶

Spatial resolution

Effective detector-element size, focal spot, and reconstruction all limit spatial resolution. PCDs can be built with small sub-pixels and without inter-pixel septa, so the detector contributes less blur. Reported systems reach substantially finer high-contrast resolution than conventional CT, which is measured by the modulation transfer function (MTF) and is most useful for bone, temporal-bone, lung, and calcified-vessel tasks. ¹⁴⁸

EID versus PCD at a glance

Characteristic	Energy-integrating detector (EID)	Photon-counting detector (PCD)
Conversion	X-ray → light (scintillator) → charge	X-ray → charge directly (semiconductor)
Detector material	Scintillator (e.g., GOS) + photodiode	CdTe or CZT semiconductor
Signal	Energy-weighted sum over interval	Individual photon counts, energy-binned
Electronic noise	Contributes, worst at low dose/large patients	Rejected below the lowest threshold
Reflective septa	Required (reduce fill factor)	Not required
Spectral data	Only in a dedicated dual-energy mode	Intrinsic on every acquisition
Spatial resolution	Limited by element size and septa	Higher; small effective element size
Iodine contrast	Underweighted low-energy photons	Improved iodine CNR

Representative values in the text (CTDIvol, DLP, CNR) are drawn from published clinical comparisons and are not a substitute for site-specific acceptance testing and protocol validation on your own system. ¹⁵

Clinical Impact

The measurable physics advantages of photon-counting CT translate into clinical value where fine detail, low dose, or spectral characterization matters. ¹⁴

Cardiac and coronary imaging. Higher spatial resolution reduces calcium "blooming," improving assessment of the lumen in heavily calcified or stented coronary arteries — a long-standing limitation of conventional coronary CT angiography. Spectral data additionally supports plaque and myocardial characterization. ¹
Lung and thoracic imaging. UHR imaging sharpens visualization of fine parenchymal and airway structures, and the dose efficiency supports low-dose follow-up. Clinical chest comparisons have shown dose reductions with preserved image quality. ⁵⁴
Musculoskeletal and temporal bone. The resolution gain benefits fine bony anatomy such as the temporal bone and small trabecular structure. ¹
Reduced iodinated contrast. Because low-keV VMI amplify iodine signal, some protocols can maintain diagnostic enhancement with reduced contrast volume — valuable for patients with impaired renal function. ¹
Oncologic and quantitative imaging. Always-on spectral data supports iodine quantification and virtual non-contrast reconstructions, potentially consolidating multiphase protocols. ⁶

These benefits are real but application- and protocol-dependent. A facility that runs a photon-counting scanner on defaults will not necessarily capture them. Deliberate protocol design — informed by CT protocol optimization — is what converts detector physics into clinical and dose benefit.

Practical Optimization Tips

Design protocols around the new degrees of freedom

Photon-counting CT adds parameters an EID protocol never had: energy-threshold selection, spectral reconstructions, VMI keV level, and UHR modes with matched sharp kernels and thinner slices. Each has an image-quality and dose implication.

Match kernel, slice thickness, and matrix to the task. UHR modes only pay off when reconstruction preserves resolution; pairing UHR data with a smooth kernel wastes the benefit, while an overly sharp kernel raises noise.
Choose VMI keV by purpose. Low keV (≈40–55) for iodine-dependent tasks; higher keV to suppress metal and beam-hardening artifact. See our related discussion of metal artifact reduction in CT.
Re-baseline dose indices. Do not assume EID reference levels. Re-establish CTDIvol and DLP expectations for each protocol on the new system and compare against diagnostic reference levels. Our overview of CTDIvol and DLP dose metrics explains what these indices do and do not represent.
Use size-specific thinking. Dose and image quality vary with patient size; interpret indices with size-specific dose estimates as described in our SSDE guide.
Leverage denoising carefully. Deep-learning and iterative reconstruction can further reduce noise on PCD-CT, as shown for CZT systems, but any denoising must be validated so it does not remove low-contrast detail. ⁷

Common pitfalls to avoid

Reusing EID protocols unchanged. Default translation of old protocols leaves dose and image-quality benefit on the table.
Ignoring CT number accuracy across keV. Spectral reconstructions must be checked for CT number accuracy at multiple energies, not just at a single conventional setting. See CT number (HU) calibration QC.
Over-sharpening. Chasing the highest resolution kernel can raise noise beyond diagnostic benefit.
Skipping UHR-specific QC. Standard resolution phantoms may not fully characterize the UHR regime.
Assuming dose savings without measurement. The dose benefit must be measured per protocol, not presumed.

Regulatory Considerations

A photon-counting CT scanner is regulated as a radiation-producing device and must meet the same federal and state requirements as any diagnostic CT, while its novel modes require careful acceptance testing and accreditation review. ¹⁰¹¹

Key frameworks:

FDA and 21 CFR 1020.33. CT systems are subject to FDA performance standards for diagnostic X-ray equipment, and clinical photon-counting systems have entered practice through FDA clearance. The first photon-counting CT was cleared in 2021. ⁹ X-ray-producing machines are regulated by the FDA together with state radiation-control programs, distinct from the NRC framework that governs radioactive material.
ACR–AAPM CT Technical Standard. Ongoing medical physics performance monitoring — including CTDIvol accuracy, image quality, CT number accuracy, and dose indices — follows the ACR–AAPM technical standard for CT, which the physicist adapts to include spectral and UHR evaluation. ¹¹
AAPM Task Group 291 (multi-energy CT). Provides the physical framework for spectral/multi-energy CT that underlies photon-counting spectral outputs, including material decomposition and dose considerations. ⁶
Dose indices and SSDE. CTDIvol and DLP are displayed and recorded; size-specific dose estimation (AAPM Report No. 204) supports patient-level interpretation. ¹⁰
Accreditation. ACR CT accreditation requires phantom image quality and dose within limits; a new modality warrants early coordination so novel modes are tested against accreditation criteria. Our accreditation support service helps map these requirements to the scanner's capabilities.

For X-ray CT the licensing authority is the state radiation-control program (with FDA device oversight), not the NRC. Facilities in DRPS service areas — Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware — should confirm state registration, physicist survey, and accreditation requirements before clinical launch.

Frequently Asked Questions (FAQs)

What is photon-counting CT?

Photon-counting CT (PCD-CT) uses semiconductor detectors, typically cadmium telluride or cadmium zinc telluride, that convert each X-ray directly into an electrical pulse and count individual photons while measuring their energy. Conventional CT instead uses energy-integrating detectors that turn X-rays into visible light in a scintillator and measure a summed signal, discarding per-photon energy information.

Does photon-counting CT lower radiation dose?

It can. Because photon-counting detectors have essentially no electronic noise and higher dose efficiency, they can maintain image quality at lower dose, or improve image quality at the same dose. Published clinical comparisons have shown meaningful CTDIvol and DLP reductions with preserved or improved contrast-to-noise ratio, but the achievable reduction depends on the exam, patient size, and how the protocol is configured.

How is photon-counting CT different from dual-energy CT?

Dual-energy CT obtains two energy measurements using two tube potentials, fast kV switching, dual sources, or dual-layer detectors. A photon-counting detector sorts every detected photon into energy bins on a single acquisition, so spectral information is available for every scan without a separate dual-energy mode, and usually with more than two energy thresholds.

Do photon-counting CT scanners need different QC than conventional CT?

The core acceptance-testing framework still applies — CTDIvol accuracy, image quality, spatial resolution, CT number accuracy, and dose indices — but the physicist must also evaluate ultra-high-resolution modes, energy-bin and virtual monoenergetic behavior, CT number accuracy across keV levels, and reconstruction options that do not exist on energy-integrating systems.

What are virtual monoenergetic images?

Virtual monoenergetic images (VMI) are reconstructions that estimate what the scan would look like at a single chosen photon energy, expressed in kiloelectronvolts (keV). Low-keV VMI boost iodine contrast; high-keV VMI reduce beam-hardening and metal artifact. Because photon-counting CT is inherently spectral, VMI can be generated from any acquisition.

Is photon-counting CT better for coronary and lung imaging?

The improved spatial resolution is particularly valuable where fine detail matters, such as imaging heavily calcified coronary arteries, the lung parenchyma, temporal bone, and small vessels. Spatial resolution and reduced blooming can improve visualization, but clinical value should be confirmed for each application and protocol rather than assumed.

Should a facility get a physicist involved before buying a photon-counting CT?

Yes. A medical physicist should support acceptance testing, protocol optimization, dose-index review, spectral and ultra-high-resolution QC, and accreditation. Because the technology is new, early physicist involvement helps a facility realize the dose and image-quality benefits rather than simply running default settings.

Key Takeaways

Direct conversion changes the physics. PCDs convert X-rays straight to charge and count photons, removing the scintillator, reflective septa, and electronic-noise penalty of energy-integrating detectors.
Image quality improves on several axes at once. Lower electronic noise, higher spatial resolution, and stronger iodine contrast can be realized together, with spectral data on every scan.
Dose can fall or image quality can rise — measure it. Clinical comparisons show ~40% CTDIvol/DLP reductions with preserved or improved CNR, but the benefit is protocol- and task-dependent and must be verified on site.
Spectral is always on. Virtual monoenergetic images, iodine maps, and virtual non-contrast reconstructions come from every acquisition rather than a special mode.
QC must expand. Acceptance testing should add UHR, energy-bin, VMI, and multi-keV CT number checks to the standard CT battery.
Regulation is the state/FDA X-ray framework. CT is FDA- and state-regulated; new modes warrant early physics and accreditation planning.

Conclusion

Photon-counting CT is the most significant change to the CT detector in decades. By counting and energy-resolving individual photons, it addresses two fundamental limits of energy-integrating detectors — electronic noise and energy weighting — and delivers higher spatial resolution and intrinsic spectral imaging as part of the same advance. The clinical payoff is real for calcified coronary arteries, lung and bone detail, contrast reduction, and quantitative tasks.

But the technology does not optimize itself. Capturing the dose and image-quality benefits depends on protocol design built around new degrees of freedom, on acceptance testing that verifies the novel modes, and on dose-index baselines re-established for the new system. Facilities that pair the hardware with disciplined medical physics support are the ones that turn detector physics into measurable clinical value.

How DRPS Can Help

Diagnostic Radiation Physics Services supports facilities adopting photon-counting and advanced CT with acceptance testing, protocol optimization, dose-index review, spectral and ultra-high-resolution QC, and accreditation support. Our board-certified medical physicists provide CT physics testing, medical physics consulting, and accreditation support so that new technology is verified, optimized, and defensible.

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

A strong CT physics program does more than pass accreditation — it makes sure the scanner you bought is delivering the image quality and dose performance you paid for.

Related Resources

References

Willemink MJ, Persson M, Pourmorteza A, Pelc NJ, Fleischmann D. Photon-counting CT: technical principles and clinical prospects. Radiology. 2018;289(2):293-312. doi:10.1148/radiol.2018172656. PubMed
Danielsson M, Persson M, Sjölin M. Photon-counting x-ray detectors for CT. Phys Med Biol. 2021;66(3):03TR01. doi:10.1088/1361-6560/abc5a5. PubMed
McCollough CH, Rajendran K, Baffour FI, et al. Clinical applications of photon counting detector CT. Eur Radiol. 2023;33(8):5309-5320. doi:10.1007/s00330-023-09596-y. PubMed
Sartoretti T, Wildberger JE, Flohr T, Alkadhi H. Photon-counting detector CT: early clinical experience review. Br J Radiol. 2023;96(1147):20220544. doi:10.1259/bjr.20220544. PubMed
Donuru A, Araki T, Dako F, et al. Photon-counting detector CT allows significant reduction in radiation dose while maintaining image quality and noise on non-contrast chest CT. Eur J Radiol Open. 2023;11:100538. doi:10.1016/j.ejro.2023.100538. PubMed
McCollough CH, Boedeker K, Cody D, et al. Principles and applications of multienergy CT: Report of AAPM Task Group 291. Med Phys. 2020;47(7):e881-e912. doi:10.1002/mp.14157. PubMed
Sasaki T, Kuno H, Nomura K, et al. CZT-based photon-counting-detector CT with deep-learning reconstruction: image quality and diagnostic confidence for lung tumor assessment. Jpn J Radiol. 2025;43(7):1132-1144. doi:10.1007/s11604-025-01759-9. PubMed
Layer YC, Kravchenko D, Dell T, Kütting D. CT technology: photon-counting detector computed tomography. Radiologie (Heidelb). 2023;63(7):497-506. doi:10.1007/s00117-023-01166-z. PubMed
Siemens Healthineers. Siemens Healthineers announces FDA clearance of NAEOTOM Alpha, the world's first photon-counting CT. Press release; September 30, 2021. siemens-healthineers.com
American Association of Physicists in Medicine. Size-Specific Dose Estimates (SSDE) in Pediatric and Adult Body CT Examinations. AAPM Report No. 204. College Park, MD: AAPM; 2011. aapm.org
American College of Radiology, American Association of Physicists in Medicine. ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of Computed Tomography (CT) Equipment. acr.org