What’s New in Photon Counting CT?

August 11, 2026 · 5 articles · 5 research themes · covering August 04, 2026 – August 11, 2026

Overview

Across these recent studies and updates, photon-counting CT (PCCT) emerges as a rapidly maturing platform spanning clinical imaging, workflow integration, and core reconstruction physics. Clinically, PCCT-derived iodine mapping for late iodine enhancement in coronary CTA work-ups shows improved quantitative image quality and reader confidence for ischemic-type myocardial scar assessment, suggesting a more reliable route to ischemic scar imaging than conventional approaches.

On the engineering and methods side, multiple papers address how PCCT performance depends on the full imaging chain. Cadaveric-head experiments quantify how smaller detector pixels influence “bucket effects” and signal-to-noise behavior under iterative reconstruction across dose levels, clarifying reconstruction–detector interactions that can guide ultra-high-resolution protocol optimization. Complementing this, an end-to-end differentiable PCCT framework makes material decomposition (maximum-likelihood estimation) trainable within the imaging pipeline, enabling cross-domain learning and potentially better spectral/quantitative performance by optimizing upstream models with image-domain quantitative fidelity.

Finally, the translational theme is reinforced by AI-enabled and regional technology perspectives. An AI routing study uses CT request data to automatically determine eligibility for PCCT versus conventional EID CT, aiming to remove the practical burden of manual triage and ensure appropriate patient selection. A narrative update on French advances highlights active translation of spectral CT (including photon-counting CT) and AI into cancer diagnosis and interventional workflows—together pointing to an ecosystem where improved hardware, smarter reconstruction, and decision-support tools are converging to accelerate clinical adoption.

---

Photon-Counting CT for Myocardial/Ischemic Scar Imaging

Photon-counting CT iodine maps for late iodine enhancement: a retrospective feasibility study of image quality and reader confidence.

This retrospective single-center study evaluated photon-counting detector CT (PCD-CT) iodine maps for late iodine enhancement (LIE) in patients undergoing coronary CTA work-up for advanced coronary artery disease, with LIE acquired 5 minutes after contrast injection (1.5 mL/kg, 350 mg/mL iodine). The study found that PCD-CT-derived iodine maps provided quantitative image quality and improved reader confidence for ischemic-type LIE compared with limitations of conventional LIE CT contrast. These results support photon-counting CT as a more reliable approach for myocardial scar assessment, potentially reducing validation barriers for LIE-based ischemic scar imaging.

Hartung V, Gruschwitz P, Serfling JK et al. · European radiology experimental · (2026) · View on PubMed ↗ · Free PDF ↗

---

Spectral CT Technology Translation & Policy/Regional Updates

Imaging in France: 2026 Update.

This narrative update reviewed recent French advances in diagnostic imaging and interventional radiology, with emphasis on research directions including dual-energy and photon-counting computed tomography and emerging artificial intelligence applications. The key finding is that French research groups are actively translating innovations in spectral CT (including photon-counting CT) and AI into improved cancer diagnosis and interventional workflows. The article’s significance is to contextualize ongoing technology development that may accelerate clinical adoption of photon-counting CT and AI-driven imaging in France.

Talabard MP, Greffier J, Calame P et al. · Canadian Association of Radiologists journal = Journal l’Association canadienne des radiologistes · (2026) · View on PubMed ↗

---

AI Decision Support for CT Modality/Protocol Routing

Automatic Patient Eligibility for Photon-counting CT Using Discriminative and Generative AI Models in Neuroradiology.

This retrospective neuroradiology study used Spanish-language CT request data to automatically determine patient eligibility/routing for photon-counting CT (PCCT) versus conventional energy-integrating detector (EID) CT using discriminative and generative AI models. The key finding was that AI-based decision support can automate selection of PCCT when appropriate, addressing the impracticality of manual routing given the large number of images produced by PCCT. Clinically, this could improve workflow efficiency and ensure patients receive the most suitable CT technology without relying on radiologist/PACS manual triage.

Martín-Noguerol T, López-Úbeda P, Escartín J et al. · Academic radiology · (2026) · View on PubMed ↗

---

Detector Pixel Size, Bucket Effects & Iterative Reconstruction Physics

Small-pixel Acquisition Unmasks “Bucket Effects” of Iterative Reconstruction in Ultra-High-Resolution Photon-Counting CT.

This cadaveric-head study investigated how small detector pixels affect “bucket effects” and image quality under iterative reconstruction in ultra-high-resolution photon-counting CT (UHR PCCT), comparing UHR (120×0.2 mm) versus standard collimation (144×0.4 mm) across multiple dose levels (0.08–10 mGy). The key finding was that smaller pixel acquisition altered signal-to-noise behavior and subjective image quality, helping unmask and characterize reconstruction-related bucket effects. Scientifically, the results clarify how reconstruction kernels and iterative reconstruction interact with detector pixel size, informing optimization of UHR PCCT protocols.

Huflage H, Wech T, Patzer TS et al. · Academic radiology · (2026) · View on PubMed ↗ · Free PDF ↗

---

End-to-End Differentiable Photon-Counting CT Reconstruction/Material Decomposition

End-to-End Differentiable Photon Counting CT.

This work developed an end-to-end differentiable photon-counting CT framework by making the material decomposition step (maximum-likelihood estimation) differentiable and inserting it as a layer into the PCCT imaging chain. The key finding is that differentiable material decomposition enables cross-domain learning and end-to-end optimization of upstream models using quantitative information in the image domain. This is significant because it can improve spectral/quantitative PCCT performance by allowing training that directly accounts for the full imaging pipeline rather than treating reconstruction and decomposition as fixed steps.

Wang S, Yang Y, Lee J et al. · IEEE transactions on medical imaging · (2026) · View on PubMed ↗

---

---

Generated automatically on August 11, 2026. Covers PubMed articles published August 04, 2026 – August 11, 2026. Summaries are AI-generated; always consult the original publication for clinical or research decisions.