Frosty or Fresh? Grading Our 2025 RSNA Predictions

Before RSNA 2025, we made six bold predictions about what radiologists would actually be talking about in Chicago. Now, with RSNA 2026 approaching, it’s time to check our work.

Some predictions proved pretty fresh, while others stayed a little frosty. And a few evolved in ways we didn’t anticipate.

Here’s our report card,  and what it tells us about where radiology may be heading next.

1. Imaging moves from standard output to situational judgment.

Grade: Fresh

This one rang true. We predicted imaging would become more personalized, with technology helping radiologists extract insights beyond the original reason for an exam. 

And it was clearly part of the 2025 conversation, with RSNA highlighting opportunistic imaging for cardio and metabolic risk prediction, along with AI-enhanced lesion detection, risk stratification and tumor characterization, as examples of the science advancing more personalized care.

The trend was visible across subspecialties, too. In breast imaging, for example, the program included research examining AI models for long-term breast cancer risk assessment.

What it means for 2026: The question may be shifting from what else can we learn from an image to what do we do with that information? Expect more focus on moving beyond demos to showing these insights at work in real clinical settings, with proof of how they’re being used and the impact they’re having.

2. AI stops selling potential and starts defending its place in the workday.

Grade: Very Fresh

This one played out like we expected it to. RSNA 2025 emphasized the evolution of AI from research to clinical implementation, including applications for stroke triage, cardiac risk stratification and tumor characterization.

But implementation is only part of the story. The bigger shift is toward AI proving its value within the clinical workflow. Not just what the technology can do, but how it fits into the work the radiologist is already doing. 

And the broader adoption numbers make the shift even clearer. According to the American College of Radiology, nearly 90% of surveyed ACR members now use some form of AI in practice, while nearly 80% of FDA-cleared medical AI tools represent radiology use cases.

What it means for 2026: AI has made it into the workday. Now it has to prove that it belongs there. Simply demonstrating that an AI model works isn't enough. With more than 190 sessions including “workflow” in the description, expect more scrutiny around real-world performance, monitoring and integration. 

3. Hardware innovation gets quieter and more relevant.

Grade: Partially Thawed

The “relevant” part was right. Not really the quieter part.

Photon-counting CT was one of RSNA 2025's highlighted trends, with sessions and research spanning functional imaging, radiation dose reduction, image quality and clinical applications across multiple subspecialties.

But the conversation increasingly connected technical advances to tangible clinical benefits. And additional research reinforced that shift. A 2026 study published in Radiology comparing photon-counting CT with conventional CT in patients with lung cancer found 66% lower radiation exposure and 27% lower iodine exposure, along with improvements in image quality and diagnostic confidence.

What it means for 2026: Hardware isn't fading, but expect more emphasis on what new technology enables — better imaging, lower dose and greater efficiency — rather than innovation for innovation's sake.

4. The workforce crisis forces radiology to redesign work, not keep pretending more staffing is coming.

Grade: Fresh

Unfortunately, this rang true. AuntMinnie.com named workforce shortages radiology's biggest threat for the third consecutive year, with increasing imaging volumes coming in second.

RSNA 2025 reflected this with sessions on clinician well-being, advanced practice providers, and new approaches to radiologist-led teams, and research presented at RSNA also illustrated the challenge. 

Using 12 months of overnight CT data from a Level 1 trauma center, researchers found that three radiologists could cover 90% of average hourly demand. But when they accounted for real-world fluctuations in volume, five radiologists were needed to meet demand 85% of the time.

What it means for 2026: Expect capacity to remain one of radiology's defining challenges but with more emphasis on redesigning work around where radiologists' time and expertise create the most value.

5. Sustainability stops being “nice to have” and becomes a business decision.

Grade: Frosty

Sustainability wasn't absent from RSNA 2025, with sessions addressing energy efficiency, lifecycle assessment and ways to reduce radiology's environmental footprint.

And professional societies like ACR and RSNA published a joint sustainability position paper calling on radiology leaders to consider energy, equipment, materials and supply-chain sustainability in operational and purchasing decisions.

But did sustainability become as prominent a business conversation as workforce, AI or workflow? Not yet.

What it means for 2026: As AI adoption grows, its energy and water demands could bring sustainability closer to the business conversation, linking it more directly to infrastructure, efficiency and cost.

6. The radiologist role expands but not by stacking more work on the same shoulders.

Grade: Fresh (with a twist)

We expected radiologists to become more involved in AI governance, enterprise strategy and decisions extending beyond interpretation. That direction held up, but AI adoption has made the responsibility more immediate than we anticipated.

The ACR adopted its first practice parameter for imaging AI in 2026 and has approved exploration and development of AI accreditation — signs that AI oversight is becoming a more formal part of radiology.

What it means for 2026: The conversation will increasingly be about not just what AI can do, but how it’s governed, monitored and built into workflows. Expect big crowds at the “AI Governance, Evaluation and Post-Deployment Monitoring” education session. 

So, how did we do?

Pretty fresh. Going into RSNA 2025, we said radiology was shifting from promise to proof. Heading into RSNA 2026, the question may be less about whether the technology works and more about how it all works together — AI, reporting, imaging, data and the radiologist at the center.

This year may be about putting the pieces together.