RSNA 2025: United Imaging Intelligence Puts AI Agents to Work From Scan to Report-and Into the Exam Room

At RSNA 2025, AI agents stepped from single tasks to partners, from uAI Insight to robotic ultrasound. The gains to watch are throughput and consistency, with humans in the loop.

Categorized in: AI News Healthcare
Published on: Dec 18, 2025
RSNA 2025: United Imaging Intelligence Puts AI Agents to Work From Scan to Report-and Into the Exam Room

RSNA 2025: AI Agents Move From Tasks to Clinical Partnership

December 12, 2025 - At RSNA 2025, United Imaging Intelligence (UII) put a clear prompt in front of radiology: "Agent Radiologist: Are We There Yet?" The focus was practical-AI agents that can understand, reason, and act across the care pathway to make precision care more accessible and to close gaps across sites and populations.

The conversation wasn't about hype. It was about redefining day-to-day workflows, tightening reporting consistency, and connecting clinical reasoning to action in real settings.

Imaging-to-Report: uAI Insight Steps Into the Reading Room

uAI Insight made its North American debut as an Imaging-to-Report agent that brings clinical reasoning into the reporting process. Built on multimodal medical models across language, vision, and speech, it correlates findings, flags key patterns, and generates structured reports with clear statements and supporting evidence.

Within the suite, the uAI Agent for Chest CT Reporting detected 73 thoracic diseases from a single scan with an average AUC of 95%. The uAI Agent for Brain MRI Reporting identified 47 neurological conditions from one scan. Together, they point to a shift from single-task tools to systems that support end-to-end diagnostic thinking.

Ultrasound, Reconsidered: Embodied AI With Tactile Feedback

UII also introduced the uAI Agent for Ultrasound-an embodied AI system with a robotic arm that performs autonomous scanning and assists with diagnosis. It interprets visual, tactile, and pressure cues in real time, adjusting probe position and force with precision to support comfort and consistency.

A Unified Ecosystem: From Patient Intake to Hospital Operations

Across the booth, AI agents worked as a team. uAI Avatar acted as an intelligent medical assistant for pre-diagnostic intake through natural conversation, while the uAI Agent for Hospital Management tracked imaging and reporting quality, evaluated equipment performance, and turned operational data into actions clinicians and administrators can use.

UII's broader concept-the "Digitelligent Hospital"-brings these pieces together. Built on multimodal medical large models refined through clinical collaborations and supported by United Imaging's technology portfolio, the approach pairs agent reasoning with embodied systems to handle real-world tasks and natural interactions.

The result is a hospital that learns, adapts, and improves across every stage of care.

What This Means for Healthcare Teams

For radiology and enterprise leaders, the message is straightforward: AI agents are moving from add-ons to infrastructure. The gains to watch are throughput, report quality, standardization, and reduced documentation drag-without losing clinician oversight.

  • Start with narrow, high-value workflows (e.g., chest CT follow-up) and define clear KPIs for time-to-report, quality, and safety.
  • Validate on local data before scaling; monitor performance drift and bias over time.
  • Integrate outputs into PACS/RIS/EMR via existing standards to avoid workflow friction.
  • Keep a human-in-the-loop for exceptions and critical findings; document escalation paths.
  • Train radiologists, technologists, and sonographers to collaborate with agents and override when needed.
  • Address governance early-privacy, cybersecurity, audit trails, and change management.

What to Watch Next

Interoperability across scanners and vendors. Generalization across sites and patient populations. Measurable ROI in time, quality, and patient experience. And careful attention to fairness so benefits reach low-resource settings, not just flagship centers.

Regulatory and Availability

Products and features referenced may not be available in all countries, and future availability cannot be guaranteed. Not all AI applications presented are CE-marked or FDA-cleared.

For regulatory context on AI/ML-based software as a medical device, see the FDA's guidance and updates: FDA: AI/ML Software as a Medical Device.

Learn More and Skill Up

If your organization is building an AI upskilling plan by role, explore curated options here: Complete AI Training - Courses by Job.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)
Advertisement
Stream Watch Guide