The FDA has granted premarket approval to Claire, an AI-powered imaging system that assesses tumor margins during breast cancer surgery in real time. It’s the first AI-enabled device cleared for this purpose in the United States.
What the Device Does
Claire uses optical coherence tomography combined with AI trained on over 2 million breast tissue images to visualize surgical margins during lumpectomy. The system produces images at 10 times the resolution of standard X-ray or ultrasound, with 2mm imaging depth, the clinically relevant margin width for breast cancer assessment.
During surgery, the AI identifies suspicious areas on excised tissue. If it detects potential residual cancer, the surgeon can remove additional tissue before closing, rather than waiting up to 10 days for pathology results.
The Clinical Trial
In a study of 206 patients, Claire demonstrated an overall accuracy of 88.1%. The numbers tell the story of where it helped:
- 17% of patients had residual disease margins after standard care alone
- Claire correctly detected residual disease in 40% of patients when used after standard evaluation
- 20% of patients were fully cleared of all residual disease thanks to Claire-guided shaves
- 6 patients had disease-confirmed margins that both standard care and histopathology missed
The results met the prespecified “super-superiority” performance goal with statistical significance (P = .0050).
What This Means
About 300,000 breast-conserving surgeries happen in the U.S. each year. Roughly 20% require repeat operations because pathology later reveals positive margins, meaning cancer cells extend to the edge of the removed tissue.
Each reoperation means another round of anesthesia, surgery, recovery, and waiting. Beyond the physical burden, there’s the psychological toll of learning the first surgery wasn’t enough.
Claire addresses this by moving margin assessment from the pathology lab into the operating room. If it works as advertised in broader deployment, fewer patients would need that second surgery.
The Fine Print
The FDA approval includes a predetermined change control plan, allowing Perimeter to implement planned AI enhancements without additional regulatory review. This reflects a newer FDA approach to AI devices: acknowledging that the algorithms should improve over time while maintaining oversight of the improvement pathway.
The company plans a nationwide launch in coming weeks. Real-world performance across diverse surgical teams and patient populations will be the true test.
Whether Claire becomes standard of care depends on practical questions yet to be answered: cost, operating room workflow integration, surgeon learning curves, and reimbursement. An 88% accuracy rate is strong but not perfect. Surgeons will need to calibrate how much they trust the AI’s guidance versus their own judgment and existing tools.
Still, for a condition where the alternative is often a second surgery, even incremental improvement matters.