How AI Is Redefining Breast Cancer Risk

Big news came out of the Radiological Society of North America annual meeting recently, where researchers unveiled an AI model that outperformed traditional methods in predicting breast cancer risk up to five years in advance.

This image-only AI tool, Clairity Breast, was trained using more than 420,000 mammograms from facilities in the U.S., Europe and South America. By analyzing mammograms from women who later developed breast cancer and those who didn’t, the model learned to detect subtle patterns invisible to the human eye.

(Patterns that even experienced radiologists might miss.)

What makes this breakthrough especially important is that it surpasses breast density as a risk factor. Traditionally, density has been used to help estimate a woman's risk, but in the study, women with high AI-predicted risk had over a 4x higher likelihood of developing cancer compared to those in the low-risk group.

This matters, because breast density isn’t always predictive — and it’s not always shared.

Too many women don’t learn about their breast density until after a diagnosis. Some never learn it at all, depending on their state’s notification laws or how their provider reports it. But this new model offers a different kind of promise: risk scoring based purely on the image data, not self-reported history or questionnaires. That means it could work quietly in the background, catching red flags earlier and with fewer blind spots.

Shifting the Screening Conversation

This development signals a broader shift toward personalized and earlier screening — a sharp departure from the rigid age-based guidelines that dominate insurance coverage and clinical pathways.

As more young women are diagnosed with breast cancer — many without a family history or known genetic mutation — the one-size-fits-all approach of “start at 40” looks increasingly outdated. The hope is that, in the future, a woman could have a baseline mammogram around age 30. From there, her AI risk score would determine whether she should enter an earlier or more frequent screening track.

And it won’t just be about more mammograms. It’s about the right ones.

With this kind of model, radiologists and clinicians could proactively tailor when to screen, how often, and with which technology (2D, 3D, ultrasound, MRI). That reduces unnecessary scans for low-risk women and ensures high-risk patients don’t fall through the cracks due to insurance timelines or guideline inertia.

Why Advocacy Must Catch Up

Right now, insurance companies and federal guidelines still cling to decades-old data. Most women under 40 have to fight to get covered for a mammogram unless they already have cancer — which defeats the entire purpose of early detection.

This research opens the door for a policy push: not just to cover risk-based screening, but to fund and integrate these AI tools across health systems, especially those serving underserved or rural communities. Otherwise, the same populations that already face delays in diagnosis could be left behind again, even with better technology in hand.

The bottom line?

The risk for breast cancer doesn’t start at 40. And now, neither should the conversation.