Clairity's AI predicts breast cancer risk from a single mammogram

When a radiologist looks at a standard mammogram, they are generally answering a single question: is there a detectable tumor right now? If the image looks clean, traditional tools can't tell you much about what will happen down the road.

Clairity just received FDA De Novo authorization for CLAIRITY BREAST, an AI platform authorized under the name Allix5. It is the first FDA-cleared AI prognostic tool capable of predicting a woman's five-year breast cancer risk using nothing more than a routine screening mammogram image.

Finding clues hidden in plain pixels

Most existing cancer risk calculators lean heavily on questionnaires covering age, lifestyle, and family medical history. But the data reveals a huge blind spot: globally, over 2.3 million women are diagnosed with breast cancer each year, yet 85% of them have zero family history, and roughly half carry no identifiable clinical risk factors.

On top of that, legacy models were built primarily around European Caucasian datasets, meaning their predictive power frequently degrades when applied to racially diverse patient populations.

Clairity's system takes a completely different path. Founded by Dr. Connie Lehman, a breast imaging specialist at Mass General Brigham, the startup trained computer vision models on millions of images to detect sub-visual tissue patterns that signal future vulnerability, even when the scan appears entirely normal to human eyes.

The algorithm was validated across more than 77,000 mammograms collected from five geographically separated screening facilities, ranging from major academic hospitals to freestanding clinics, with all results anchored to actual five-year clinical outcomes. The company plans to roll out the platform across major healthcare networks throughout 2025.

Why it matters

We think this clearance marks a crucial transition for medical AI, shifting the focus from catching visible disease to forecasting invisible risk.

For patients, this changes routine screening from a reactive check-in to proactive prevention. If a standard screening flags an elevated five-year score, a physician can immediately justify supplemental screening like an MRI, catching tumors before they reach late stages without requiring invasive biopsies or expensive genetic tests. Because the score runs on standard mammograms already taken in the clinic, it fits into existing radiology workflows without adding new appointments.

For the broader healthtech space, Clairity just carved out a distinct path in the $63 billion breast cancer prediction market. While most radiology startups have spent the last decade building diagnostic overlays that point out existing lesions, Clairity shows that pixel-level prognosis can secure clear regulatory backing under the De Novo pathway.

The biggest hurdle now will be clinical adoption: health systems and insurers must figure out standard protocols for what to do with a high-risk score on an otherwise normal scan, avoiding unnecessary downstream panic while ensuring supplemental scans get reimbursed.