By Willie Foerstner

The RSNA 2025 Annual Meeting delivered one of the most transformative demonstrations of progress in modern medical imaging, showcasing sweeping advancements in artificial intelligence, photon-counting CT, next-generation MR, advanced ultrasound, enterprise workflow ecosystems, and a decisive leap forward in PET and nuclear medicine.

This year’s theme, Imaging the Individual, reflected a profound evolution in radiology: the transition from isolated image acquisition toward fully integrated, intelligent diagnostic and therapeutic pathways designed around each patient’s biology, risk profile, and clinical trajectory.
Artificial intelligence becomes radiology’s operating system
AI at RSNA 2025 was no longer framed as an accessory or add-on. Instead, it emerged as the foundational operating layer of radiology, connecting scheduling, acquisition, reconstruction, interpretation, reporting, and downstream clinical decision-making. The future on display was unmistakably AI-first, multimodal, agentic, continuously learning (under controlled regulatory frameworks), and enterprise-integrated.

Microsoft set the enterprise tone with the expansion of Dragon Copilot, a multimodal, generative, and agentic AI embedded within PowerScribe One. Radiologists experienced automated impression drafting, real-time prior summarization, anatomy recognition, billing optimization, and instant clinical reasoning — transforming reporting into a dynamic conversation between clinician and machine.

Across the show floor, Siemens Healthineers, GE HealthCare, Philips, Canon Medical Systems, Hologic, Samsung, and United Imaging demonstrated that AI is no longer bolted onto systems but architected directly into detectors, reconstruction pipelines, user interfaces, and cloud platforms. Entire imaging suites were redesigned around AI-guided acquisition, intelligent protocoling, motion correction, dose optimization, and technologist navigation.

RSNA’s AI Pavilion featured hundreds of innovators. Among the most influential were DeepHealth, Flywheel, Annalise.ai, Aidence, Subtle Medical, Lunit, Contextflow, Qure.ai, HeartFlow, and Perspectum, showcasing systems capable of multi-organ analysis, automated staging, autonomous triage, risk stratification and, in research environments, controlled self-learning. Together, these platforms signaled a radiology ecosystem capable of continuous improvement at scale.

Siemens Healthineers: Photon-counting leadership, AI pathways, and strategic reinvention
Siemens Healthineers delivered one of the strongest strategic statements of RSNA 2025. Throughout the meeting, executives, engineers, and clinical collaborators consistently emphasized that photon-counting CT is the future of diagnostic imaging, offering unmatched spectral fidelity, material separation, and noise performance that fundamentally changes how disease is detected and characterized.

The company highlighted its expanding photon-counting ecosystem alongside Optiq AI, an advanced imaging chain delivering real-time AI enhancement for interventional radiology with exceptional clarity at low radiation dose. Siemens also continued its push into MRI-guided therapy with the MAGNETOM Free.XL, pairing MRI accessibility with intelligent acquisition engines to expand procedural and interventional use cases.

A pivotal corporate milestone was announced: Siemens Healthineers is now operating as a fully stand-alone company following its corporate divestiture, granting it increased strategic freedom to accelerate investment across imaging, oncology, and digital platforms.

One of the most consequential developments occurred off the show floor in Forchheim, Germany, where Siemens opened its High Energy Photonics Center, a highly digitalized, nearly CO₂-neutral production facility for high-performance X-ray tubes and precision imaging components. This represents the largest single new-construction investment in the company’s history and establishes a new global benchmark for imaging manufacturing quality.

GE HealthCare: Deep silicon photon-counting CT and the digital radiology ecosystem
GE HealthCare introduced one of RSNA’s most anticipated breakthroughs with Deep Silicon photon-counting CT, marking GE’s formal entry into next-generation detector physics. The Deep Silicon architecture delivers superior spectral resolution, reduced noise, and enhanced visualization of subtle pathology—addressing long-standing limitations of conventional CT.

GE complemented this hardware innovation with a comprehensive digital ecosystem strategy focused on AI-powered patient positioning, autonomous protocoling, predictive maintenance, workflow forecasting, and diagnostic copilots integrated across enterprise imaging networks. The company’s messaging balanced physics innovation with a strong emphasis on radiologist experience, throughput, and operational efficiency.

Canon Medical Systems: AI-driven reconstruction and practical photon-counting direction
Canon Medical Systems reinforced its commitment to accessible, AI-enhanced imaging. Its deep-learning reconstruction platform, once limited to flagship systems, now spans the CT and MR portfolio, delivering high-fidelity imaging, reduced variability, and consistent quality even in high-throughput environments.

Canon’s RSNA demonstrations focused heavily on technologist empowerment, showcasing intelligent protocol guidance and automated optimization designed to reduce manual workload while preserving diagnostic integrity. Canon also previewed its photon-counting CT direction in a research and innovation setting, signaling a clear trajectory toward spectral imaging that prioritizes workflow practicality, quantitative accuracy, and clinical usability.

Philips: Intelligent visualization and multimodality interoperability
Philips advanced the enterprise imaging conversation with Advanced Visualization Workspace (AVW) 16, a vendor-neutral, cloud-enabled platform integrating AI across cardiology, oncology, neuro, and vascular imaging. Philips’ emphasis on interoperability and cross-site collaboration resonated strongly with health systems seeking unified diagnostic environments across modalities and vendors.

Hologic: Population-scale breast imaging and AI-driven screening
Hologic expanded its leadership in breast imaging, showcasing advances in digital breast tomosynthesis, AI-assisted cancer detection, and workflow automation designed for population-scale screening programs. With millions of annual exams at stake, Hologic demonstrated how AI can materially improve early detection, reading efficiency, and screening consistency at national and global scale.

United Imaging: Democratizing high-end imaging with native AI
United Imaging presented a rapidly expanding portfolio across CT, MR, PET, and ultrasound, unified by a philosophy of native AI integration. Its uSONIQUE ultrasound platform drew attention for AI-based optimization that adapts to patient anatomy while automating routine acquisition steps. The company’s message centered on broad access to high-end imaging, extending advanced capabilities beyond major academic centers.

Samsung: Ultrasound precision through AI-augmented acquisition
Samsung continued its momentum in ultrasound with the debut of the R20 Ultrasound System, integrating AI-driven measurement automation, guided acquisition, and operator fatigue reduction. The R20 reinforced ultrasound’s growing role as a high-accuracy, high-throughput frontline diagnostic modality.

National imaging providers: Scaling innovation into clinical reality
Large national imaging organizations emerged as essential technology integrators and validation engines.

RadNet highlighted continued integration of DeepHealth, streamlining interpretation and improving consistency while thoughtfully managing radiologist and technologist workload across its nationwide footprint.

SimonMed, Rayus Radiology, CDI (Center for Diagnostic Imaging), and Shared Health Services demonstrated how AI-guided workflows, PET expansion, and distributed imaging models bring advanced diagnostics into community and mobile settings.

Ackumin added momentum to molecular diagnostics, expanding PET/CT and radiopharmaceutical capabilities with AI-driven quantification for oncology and neurology, positioning itself at the intersection of detection, therapy planning, and theranostics.

Looking Ahead: Precision imaging becomes standard of care
RSNA 2025 marked the moment radiology fully embraced its next era. AI is now the operating system. Photon-counting CT is the emerging gold standard. PET and molecular imaging are central to oncology and neurology. Enterprise platforms unify data, workflow, and intelligence. National providers scale innovation to millions of patients.

What once felt futuristic is now operational reality. RSNA 2025 will be remembered as the inflection point where radiology transitioned decisively into precision imaging, intelligent diagnostics, and patient-centric care pathways, launching the field into a profoundly new chapter.

About the author: Willie Foerstner is healthcare correspondent at PMAC Capital Reports. Willie leverages his unique blend of clinical insight and capital expertise to report on how innovation is funded, developed, validated, and delivered to patients. From advanced imaging technologies — including photon-counting CT, PET/MR, and myocardial perfusion imaging — to cyclotron-based isotope production, theranostics, and next-generation radiopharmaceutical trials, he offers perspective that bridges financial strategies, regulatory pathways, and patient outcomes.

Artificial Intelligence Homepage