What's New in Photon Counting CT? — July 15, 2026
AI-summarised digest of 6 PubMed articles on Photon Counting CT published in the last 7 days.
Jump to category
What’s New in Photon Counting CT?
July 15, 2026 · 6 articles · 7 research themes · covering July 08, 2026 – July 15, 2026
Overview
This week’s papers cluster around a common theme: extracting more clinically actionable information from advanced imaging—especially photon-counting CT (PCCT/SPCCT/PCD-CT)—by improving quantitative accuracy, optimizing acquisition/reconstruction, and leveraging AI.
On the physics/quantification side, work on photon-counting CT material decomposition and spectral modeling targets more reliable tissue and dose-relevant parameters. One study quantified fundamental limits for estimating proton stopping power ratio and range from PCCT, informing strategies to reduce proton range uncertainty for therapy planning. Another demonstrated scanner-specific power-law spectral modeling to produce voxelwise effective atomic number and electron density, enabling quantitative characterization without relying on conventional dual-energy assumptions.
In cardiology, multiple studies emphasize practical performance gains for detecting clinically subtle findings. A retrospective analysis showed that spectral analysis and reconstruction choices can materially change late iodine enhancement contrast-to-noise ratio, directly affecting myocardial scar detectability. Complementing this, an AI-based CNN denoising approach improved ultrahigh-resolution photon-counting CCTA for non-calcified plaque assessment under varied reconstruction settings. A broader review further contextualized where cardiac CT angiography fits in clinical decision-making—particularly for ruling out coronary artery disease in low-to-intermediate risk patients and expanding beyond luminal stenosis toward plaque characterization.
Finally, outside CT, a prospective study compared an AI-powered compressed-sensing lung MRI protocol against photon-counting detector CT (as reference) for lung nodule detection, evaluating whether a radiation-sparing workflow can approach CT’s screening performance. Together, these results highlight a field moving toward more quantitative, optimized, and AI-enhanced imaging pipelines—aiming to improve diagnostic confidence while managing dose and uncertainty.
Spectral CT for Tissue Characterization (Zeff, Electron Density)
Power law spectral photon-counting CT for quantitative effective atomic number and electron density imaging.
The study developed and evaluated a scanner-specific power-law spectral photon-counting CT method to estimate voxelwise effective atomic number (Zeff) and electron density (ρe) using a commercial spectral photon-counting CT (SPCCT) system. It found that modeling the X-ray attenuation coefficient as a power-law sum of photoelectric, Compton, and empirical correction terms with coefficients fitted to NIST data (Z=6–21; 7–115 keV) enabled quantitative Zeff and ρe imaging in a QRM spectral-CT phantom with soft-tissue-equivalent, hydroxyapatite, and iodine inserts. This is clinically important for improved tissue characterization and material quantification in heterogeneous samples using PCCT without relying on conventional dual-energy assumptions.
Alzaabi M, Behouch A, Tariq B et al. · Physics in medicine and biology · (2026) · View on PubMed ↗
Photon-Counting CT Optimization for Contrast-Enhanced Cardiac Imaging
Image-quality optimization for late iodine enhancement with photon-counting CT: impact of spectral analysis and reconstruction parameters.
This retrospective single-center study evaluated how spectral analysis and reconstruction parameters affect image quality for late iodine enhancement (LIE) on photon-counting detector CT (PCD-CT) in patients undergoing myocardial scar assessment. The key finding was that specific combinations of spectral analysis settings and reconstruction parameters (e.g., reconstruction kernels and slice thickness) meaningfully changed LIE contrast-to-noise ratio (CNR), thereby impacting the detectability of myocardial scar-related LIE. This is significant because it provides practical optimization guidance to improve LIE performance for more reliable myocardial scar identification.
Bruno E, Palmisano A, Pisu F et al. · La Radiologia medica · (2026) · View on PubMed ↗
Cardiac CT Angiography Clinical Evidence & Review
[Computed tomography in cardiology: its necessity and developments].
This article reviewed the necessity and recent developments of computed tomography in cardiology, focusing on cardiac CT angiography (CCTA) and its role in clinical decision-making. It highlighted that CCTA is established for ruling out coronary artery disease in low-to-intermediate pretest probability patients and can support acute evaluation in selected populations while enabling broader coronary atherosclerosis characterization beyond luminal stenosis. The significance lies in consolidating current evidence and technical advances that support CCTA’s expanding diagnostic and plaque characterization capabilities.
Breitbart P, Giokoglu E, Yikit E et al. · Innere Medizin (Heidelberg, Germany) · (2026) · View on PubMed ↗
AI Denoising & Reconstruction for Photon-Counting CT
Deep-learning denoising for ultrahigh-resolution photon-counting detector CT: phantom and in vivo evaluation of non-calcified coronary plaques.
The study investigated convolutional neural network (CNN)-based denoising for ultrahigh-resolution photon-counting detector CT (UHR-PCD) coronary CT angiography (CCTA) to evaluate non-calcified coronary plaques (NCPs). In a dynamic phantom and in vivo patient data, CNN denoising improved image quality for detecting and characterizing non-calcified plaques under varying reconstruction settings (e.g., sharp kernel Bv64, 0.2/0.4 mm slice thickness, quantum iterative reconstruction levels 3/4) compared with reconstructions without denoising. This is significant because it supports lower-noise, higher-resolution plaque assessment on PCD-CT, potentially improving diagnostic performance for coronary disease evaluation.
Hyska S, Hagar MT, Osoria-Velasquez J et al. · The international journal of cardiovascular imaging · (2026) · View on PubMed ↗ · Free PDF ↗
AI-Accelerated MRI vs CT for Lung Nodule Detection
Diagnostic performance of a single breath-hold lung MRI scan with AI-powered compressed sensing for nodule detection in comparison to photon counting detector-CT.
This prospective single-center study compared diagnostic performance of a single breath-hold, AI-powered compressed sensing accelerated 3D-T1w-FFE 3T lung MRI protocol (AI-CS; acceleration factor 9) for lung nodule detection against photon-counting detector CT (PCD-CT) as the reference standard. The key finding was that the MRI approach could detect lung nodules with performance benchmarked against PCD-CT using Lung-RADS scoring and nodule size assessment in 148 healthy adults, with scans performed within 24 hours of PCD-CT. This is significant because it evaluates whether an accelerated AI-CS MRI workflow can serve as a competitive, radiation-sparing screening alternative to PCD-CT for lung nodule detection.
Palmisano A, Piccinni G, Serra D et al. · European radiology · (2026) · View on PubMed ↗ · Free PDF ↗
Proton Therapy Planning with Photon-Counting CT (SPR/Range Estimation)
Optimal dimensionality and fundamental limits of proton stopping power estimation with photon-counting CT material decomposition.
The study assessed fundamental limits and feasibility of estimating proton stopping power ratio (SPR) and proton range from photon-counting CT (PCCT) using Eigentissue decomposition for material decomposition (MD) in the context of proton therapy planning. Key results showed how increasing the dimensionality of MD with PCCT energy measurements affects the achievable accuracy bounds for SPR and range estimation. This is significant because it informs the optimal PCCT/MD strategy to reduce proton range uncertainty and treatment margins in clinical proton therapy.
Larsson K, Näsmark T, Andersson J et al. · Physics in medicine and biology · (2026) · View on PubMed ↗ · Free PDF ↗
Generated automatically on July 15, 2026. Covers PubMed articles published July 08, 2026 – July 15, 2026. Summaries are AI-generated; always consult the original publication for clinical or research decisions.