The study focused on a demanding clinical setting: CT-guided renal cryoablation, a minimally invasive procedure in which physicians thread thin, needle-like probes into kidney tumors to freeze and destroy cancerous tissue. These procedures require repeated CT imaging—first to plan the approach, then to confirm probe placement, then to monitor the freezing process and verify the result. Because patients with kidney cancer often undergo serial imaging for diagnosis, treatment, and years of follow-up, their cumulative radiation exposure can add up quickly. Reducing the dose of each scan, even modestly, compounds into a meaningful protective benefit over a patient’s entire course of care.

The researchers designed an elegant experiment that sidestepped a common pitfall in dose-reduction research. Traditionally, scientists either simulate low-dose scans by injecting mathematical noise into routine-dose data, or they irradiate volunteers twice at different dose levels to compare the results directly. The first approach is limited by restricted access to raw projection data; the second raises ethical concerns about unnecessary exposure. Instead, the Mayo team exploited a natural feature of the ablation workflow itself. Each patient already receives a routine-dose planning scan and a lower-dose probe placement check within the same hour, on the same scanner, with the same body habitus. By retrospectively collecting both datasets from fifteen patients, the investigators achieved the statistical power of a paired, dual-acquisition design without adding a single extra milligray of radiation.

The numbers behind the comparison are straightforward. The routine-dose planning scans delivered a volumetric CT dose index of 18.9 plus or minus 4.7 milligrays, while the low-dose probe placement scans delivered 9.9 plus or minus 2.5 milligrays—roughly a fifty percent reduction. The routine-dose images were reconstructed with iDose, a Philips iterative reconstruction algorithm that represents the current clinical standard. The low-dose images were processed with Spectral Precise Image, a newly developed AI-based spectral reconstruction technique that this study was the first in the world to evaluate. Both were benchmarked against filtered back projection, the classical reconstruction mathematics that serves as a common reference point for measuring how much noise each modern algorithm removes.

The quantitative results were striking. Measured in the aorta at the level of the twelfth thoracic vertebra, iterative reconstruction reduced image noise by about 23 percent relative to filtered back projection, at both dose levels. The AI-based spectral reconstruction cut noise by roughly 61 to 62 percent—nearly three times the noise suppression of the clinical standard. Most remarkably, the low-dose images processed by AI were actually 32 percent less noisy than the routine-dose images processed with iterative reconstruction, a difference that was highly statistically significant. In other words, the AI algorithm did not merely compensate for the halved dose; it produced a cleaner image than the higher-dose scan.

Noise reduction alone means little if it comes at the cost of blurred anatomy, a known hazard of aggressive denoising that can give tissue a plastic, unnatural texture. To check for this, the researchers plotted line profiles—mathematical cross-sections of signal intensity—across a renal lesion and through the vertebral foramen, testing how well each algorithm preserved edges in both soft tissue and bone. Compared with filtered back projection, neither iterative nor AI-based reconstruction showed any discernible degradation of anatomic structures. Edge sharpness was well preserved for low-contrast and high-contrast features alike, suggesting the AI engine suppresses noise without smearing the fine boundaries radiologists depend on to localize tumors and plan probe trajectories.

The human verdict came from a blinded reader study. Two radiologists independently rated overall image quality for conventional, 50 kiloelectron volt virtual monoenergetic, and electron density results on a five-point Likert scale, without knowing which dose level or reconstruction algorithm had produced each image. For the conventional images, the low-dose AI protocol scored significantly higher than the routine-dose iterative protocol, at 3.6 versus 3.2, a difference the authors described as superior image quality for the reduced-dose condition. For the 50 keV monoenergetic images, the two protocols were statistically indistinguishable, with both scoring near 3.8 to 3.9. Only the electron density maps, which are intended for quantitative measurement rather than direct visual reading, scored low across the board, with radiologists noting that their texture appeared less realistic.

The technical achievement rests on how the AI reconstruction differs from its predecessors. Spectral CT, also called dual-energy CT, acquires x-ray data at high and low energies, enabling the generation of virtual monoenergetic images, electron density maps, and material-specific images such as iodine or calcium maps. But splitting the data by energy amplifies noise, because the increased dimensionality of energy-selective imaging inherently increases signal variance. Earlier AI reconstruction methods tackled noise in the image domain with convolutional neural networks, or in the projection data before filtered back projection, or by converting sinograms directly into images with deep learning. The Spectral Precise Image technique evaluated here is distinctive in that a single AI model can be applied across all available spectral results—conventional, monoenergetic, and electron density—providing uniform noise control across the entire spectral output of the scanner.

The implications extend well beyond kidney ablations. Interventional radiology is a particularly fertile ground for dose reduction because procedures involve many sequential scans, exposing not only patients but also physicians and care teams to scattered radiation. The authors emphasize that their study establishes feasibility rather than clinical readiness: the cohort of fifteen patients at a single center was small, the patient sizes ranged only from about 29 to 38 centimeters in water equivalent diameter, and the study did not assess diagnostic quality, in which the existence and location of lesions are unknown rather than already established. Slices containing the ablation probes were also excluded to keep readers blinded, a necessary compromise the authors acknowledge as a limitation. Future work, they write, will examine whether diagnostic CT can safely adopt the same dose reductions, how performance varies across body habitus, and how much the AI approach outperforms iterative reconstruction when doses are matched.

Still, the central result is hard to ignore: an AI algorithm, applied to scans using half the radiation, produced images that radiologists judged better than the clinical standard. As deep learning reconstruction matures from a promising laboratory concept into a validated clinical tool, the long-standing assumption that image quality must be purchased with radiation dose is beginning to crumble. For the millions of patients who undergo CT guidance each year—and for the cancer survivors who face decades of surveillance imaging—that crumbling cannot come soon enough.

Subject of Research: AI-based spectral CT reconstruction for radiation dose reduction in image-guided renal ablation

Article Title: Feasibility of radiation dose reduction with AI-based spectral reconstruction for computed tomography

Article References: Feasibility of radiation dose reduction with AI-based spectral reconstruction for computed tomography. (n.d.). https://doi.org/10.1007/s44343-025-00021-3

Image Credits: AI Generated

DOI: 10.1007/s44343-025-00021-3

Keywords: computed tomography, AI reconstruction, radiation dose reduction, spectral CT, deep learning, interventional radiology, renal cryoablation, image quality, virtual monoenergetic imaging, denoising, Mayo Clinic, medical imaging

The key to the approach lies in a piece of physics that dates back to 1976, when Robert Alvarez and Albert Macovski showed that X-ray attenuation in the diagnostic energy range can be described as the sum of just two effects: the photoelectric effect and Compton scattering. The photoelectric contribution depends strongly on the atomic number of the material, while the Compton contribution depends mainly on electron density. If attenuation can be measured at several different X-ray energies, the two contributions can be mathematically separated, and from them the material’s mass density and its effective atomic number — a weighted average of the atomic numbers of the elements it contains — can be calculated. This formalism underlies virtually every spectral CT technique in use today, from virtual non-contrast imaging to material decomposition.

What has historically made such calculations difficult in the clinic is that the raw projection data needed for them are locked away inside proprietary scanner software. Manufacturers guard the sinograms, the scanner geometry, the beam-hardening corrections and the spectral detector response functions that a from-scratch reconstruction would require. The German team sidestepped this obstacle elegantly: instead of raw data, they used virtual monoenergetic images, or VMIs, which the scanner’s own software already produces. VMIs simulate what a CT image would look like if the X-ray beam consisted of photons at a single energy, and because the manufacturer has already corrected them for beam hardening, beam quality and detector efficiency, they arrive ready for quantitative analysis. No knowledge of the bow-tie filter or the tube spectrum is needed.

The scanner in question was the Naeotom Alpha from Siemens Healthineers, the first clinically approved photon-counting CT system, released in 2021. Unlike conventional detectors that merely count photons in aggregate, photon-counting detectors register each X-ray photon individually and sort them by energy, providing spectral data intrinsically with every scan and without any additional radiation dose to the patient. From a single acquisition, the researchers generated virtual monoenergetic reconstructions at six energies — 40, 50, 60, 80, 100 and 150 kiloelectronvolts — using the scanner’s Syngo.Via software, and then imported these image sets into their own Python-based analysis pipeline.

The mathematics at the heart of the method is disarmingly compact. Following the Alvarez-Macovski model, the linear attenuation coefficient at each energy is written as a photoelectric term, scaling with density, effective atomic number and an inverse power of the energy, plus a Compton term described by the Klein-Nishina function. Substituting two composite coefficients reduces the problem to a linear system: with attenuation measured at two or more energies, the coefficients can be solved by linear regression, and from them density and effective atomic number follow directly. Using six energies rather than the minimum of two gives the regression extra robustness against the noise that inevitably contaminates real CT data, producing more stable values in the final maps.

Before the algorithm could touch a scanner, its four free parameters had to be calibrated. The team fitted the model to reference attenuation data from the NIST XCOM database for sixteen elements spanning effective atomic numbers from 5, boron, to 20, calcium — precisely the range occupied by the elements that make up human tissue, and a range conveniently free of the abrupt K-edge absorption features that would complicate the fit. The calibration, performed over 40 to 200 keV, yielded parameters remarkably close to those reported by earlier groups, with a coefficient of determination of 0.9996. A verification step then confirmed that the calibrated model reproduces the literature values for density and atomic number with slopes of almost exactly one, small residual deviations being folded in as correction factors.

To test the method on real scanner output, the researchers prepared a series of alcohol-water mixtures — ethanol at 0, 25, 50, 75 and 95 percent by volume, and isopropanol at 0, 10, 35, 60 and 70 percent. These humble mixtures were chosen deliberately: their densities and effective atomic numbers fall squarely in the soft-tissue range of the human body, they are cheap, safe and easy to prepare, and their expected properties can be calculated from published reference data. The tubes were mounted in a custom epoxy-resin phantom nested inside a larger PMMA ring to mimic the scattering conditions of a human torso, and scanned five times at 140 kilovolts so that image noise could be reduced by averaging.

The results were strikingly accurate. Across both mixture series, the measured mass densities deviated from the expected values by an average of just 1.0 percent, while the effective atomic numbers deviated by 0.84 percent. Pure water yielded an effective atomic number of 7.49 against an expected 7.47, and even the most concentrated alcohol solutions tracked their reference values almost perfectly. The maps themselves visually distinguished every concentration in both series. The accuracy is comparable to that reported by earlier image-domain approaches, such as the work of Heismann and colleagues, which required far more complex preprocessing of raw data to achieve deviations of similar magnitude.

The method is not without limits. The Zeff maps proved noisier than the density maps, because Compton scattering depends only weakly on attenuation in this energy range, leaving the atomic-number estimate more vulnerable to image noise, and residual beam-hardening effects produced a gentle gradient across the phantom. The algorithm is also valid only for materials with effective atomic numbers between 5 and 20 and densities up to about 2.0 grams per cubic centimeter — a range that comfortably covers biological tissue, including cortical bone at roughly 1.8 to 2.0, but excludes denser materials such as metal implants. Strong artifacts that distort Hounsfield units, such as severe beam hardening, will propagate into erroneous map values, since the entire method rests on the fidelity of the underlying image data.

Even so, the study lands at a moment of rapid momentum in quantitative spectral imaging. Recent work by Zimmerman and Poludniowski and by Lustermans and colleagues has demonstrated photon-counting CT’s potential for radiotherapy-related material characterization, while other groups are pursuing physics-informed deep-learning approaches to the same problem. Against those data-hungry machine-learning methods, the new algorithm offers a complementary virtue: it is analytically derived, transparently calibrated against tabulated reference data, computationally light, and runs entirely on image data that any clinical workstation can export. The authors suggest it could support applications from radiation therapy planning, where density and atomic number inform dose calculations, to oncology, where tissue characterization may distinguish tumor from benign lesions, or the identification of uric acid stones. Future studies, they note, must validate the approach across different phantoms, protocols, dose levels and realistic tissue compositions — but the demonstration that hospital-ready images alone can yield quantitative chemical maps marks a meaningful step toward making spectral material analysis a routine part of clinical CT.

Subject of Research: Calculation of mass density and effective atomic number maps from virtual monoenergetic photon-counting CT images

Article Title: Computation of the effective atomic number and mass density from virtual monoenergetic photon-counting CT reconstructions

Article References: Horstmeier, S., Thomsen, F. S., & Borggrefe, J. (2026). Computation of the effective atomic number and mass density from virtual monoenergetic photon-counting CT reconstructions. Heliyon, 12(14), Article e45404. https://doi.org/10.1016/j.heliyon.2026.e45404

Image Credits: AI Generated

DOI: Not provided

Keywords: photon-counting CT, effective atomic number, mass density, virtual monoenergetic imaging, spectral CT, Alvarez-Macovski model, material decomposition, radiation therapy, tissue characterization, Heliyon, X-ray attenuation, quantitative imaging