I just got back from SIIM in Pittsburgh and I am reminded that the informatics community has the good fortune to see what is coming before it arrives.

We have tracked this together.

Film to PACS felt disruptive until it expanded what radiologists could offer. Voice recognition felt like a loss of craft until it became invisible infrastructure. Narrow AI algorithms were the state of the art until they weren’t.

The research agenda at SIIM 2026 looked like a field that is no longer asking whether foundation models belong in radiology. It is asking how to deploy them well.

Foundation models have moved from research curiosity to practical deployment much faster than many expected. I wanted to consolidate what I’ve learned into a short guide for radiologists trying to understand where this is heading.

The result is A Radiologist’s Introduction to Foundation Models: The Best Time in History to Be a Radiologist, now available on Amazon.

It is written for practicing radiologists, residents, fellows, and imaging leaders who want to understand what foundation models and VLMs actually mean for clinical practice.

The book covers the shift from narrow algorithms to foundation models, how these systems are evaluated and governed, and why I believe this is the most capable moment in the history of the specialty.

To celebrate the launch, the Kindle edition will be available free on Amazon through June 18, 2026. No code needed.

Link to book on Amazon: https://a.co/d/0cn7zUuM

As in years prior, I’ll have paperback copies with me at RSNA, so feel free to flag me down at McCormick.

The future of radiology will not be defined despite what radiologists bring to AI. It will be defined because of it.