By Beth Strohbusch
Salt Lake City-headquartered Intermountain Health is the nation’s first health system to deploy an AI-centric breast cancer margin assessment solution following its FDA approval in March 2026. The microscopic tissue visualization system, called Claire, is being used for its earliest cases to help surgeons make more informed decisions while the patient is still in surgery.
“Microscopic cancer can be difficult to identify in the operating room. When you’re doing a lumpectomy, there’s no way to know 100% that you have clear margins,” says Teresa Reading, MD, a breast surgeon at Intermountain Health LDS Hospital in Salt Lake City. “This device helps direct my attention to specific areas of concern, giving me more information to guide my decisions during these procedures.”
Surgeons performing breast-conserving surgery have long faced a difficult challenge: remove too little tissue and risk leaving undetected cancer behind or remove too much and potentially affect cosmetic results. Reoperation rates within one year of surgery range from 15% to 21% and translate into 24% higher costs, according to a study published in 2024 in the Annals of Surgical Oncology. “The rates of reoperation after breast cancer surgery have remained high. Continuing efforts to reduce reoperation need more attention,” the study’s researchers concluded.
Reading says many ideas have been pursued to improve margin assessment over the years. “I really think this AI approach is on the right track and will be very helpful,” she says.
The foundational technology relies on optical coherence tomography (OCT), which uses near-infrared light to obtain real-time 3D, volumetric, cross-sectional views of excised tissue. A trained OCT system user in the operating room performs tissue imaging and shares data with the surgeon to direct attention to specific areas of concern.
In its consensus guideline on ductal carcinoma in situ treated with lumpectomy and whole-breast radiation, The Society of Surgical Oncology, American Society of Clinical Oncology, and ASTRO define a 2 mm margin as adequate because it lowers the rate of local recurrence compared with narrower negative margins. Research shows OCT producing microscopic, cross-sectional images of excised tissue to a depth of about 2 mm at roughly 10 times the resolution of ultrasound and specimen radiography and 100 times that of MRI.
Intermountain is using a solution from imaging vendor Perimeter, whose patented concept of wide-field OCT stitches individual images together to scan larger excised tissue specimens. Since receiving FDA clearance for its first OCT solution in 2016, the company has centered its efforts in oncology and specifically in breast cancer. Claire can essentially be viewed as an AI add-on to the company’s core OCT technology. It complements the traditional visualization approach by bringing AI-driven pattern recognition capabilities into the mix, Reading explains.
“During surgery, Claire runs in the background, compares my images to its core dataset, and flags areas of potential concern,” Reading says. “It resembles the approach mammography has used for some time. Having input from both the technician and AI gives me an added level of confidence.”
Promising Results
The Claire trial, presented at the 26th Annual Meeting of the American Society of Breast Surgeons in May 2025, demonstrated that the AI-enabled wide-field OCT system achieved a margin accuracy of 88.1% and a statistically significant reduction in patients with residual cancer postsurgery, compared with the standard of care alone. The trial data demonstrated super-superiority (the lower bound of confidence interval for treatment effect was greater than a predetermined minimal clinically meaningful difference) of its ability to aid surgeons in achieving clear surgical margins.
Claire relies on a proprietary algorithm trained on the vendor’s library of more than two million images of both cancerous and healthy breast tissue. It flags suspicious foci to assist surgical decision-making while the patient remains in the operating room. Reading and a colleague at Intermountain’s American Fork Hospital are the nation’s first physicians to use Claire in the operating room.
“Every breast surgeon wants to give their patient the best possible chance of completing treatment in a single surgery,” she says. “This technology is not perfect, but it’s super-promising and has already been helpful to me.”
Do physicians see a future where this type of AI fully replaces the technician in the operating room? “You’ll always need someone to place the specimen in the machine and do the scanning, but perhaps AI can help improve interpretation,” Reading says.
Pattern recognition is an increasingly relied-upon AI technique in medical imaging. “Claire is still new, but I think the technology will only improve as it gets more data,” Reading adds. “We’ll be tracking take-back data. Ideally, we’ll see a reduction in repeat surgeries.”
New Possibilities
A 33-hospital health system, Intermountain brought this technology into surgical suites via its venture arm. “Reducing unnecessary repeat surgeries is a priority because every additional procedure places a burden on patients, families, surgeons, and the health care system overall,” says Phillip Wood, executive program director for Intermountain Ventures.
Reading believes other surgical specialties might benefit from this approach, especially when the need to freeze larger, more fatty tissue samples delays or precludes intraoperative use cases.
“We look forward to [demonstrating] how AI can be responsibly integrated into surgical care,” Wood says.
OCT is one of several intraoperative margin assessment methods available for breast-conserving surgery. Others, including ultrasound, specimen radiography, frozen section analysis, and cavity shave margins, are considered conventional approaches, according to a 2025 review in the journal Diagnostics. Fluorescence imaging, radiofrequency spectroscopy, and microCT, as well as OCT, are among emerging methods gaining traction.
Since OCT was introduced by Massachusetts Institute of Technology researchers in 1991, the technique has been used widely in ophthalmology, where it’s become a global standard for diagnosing and monitoring retinal diseases and glaucoma. Dermatologists use OCT to image skin cancers and lesions, and, in cardiology, the technique helps physicians guide percutaneous coronary interventions, assess plaque vulnerability, and optimize stent placement.
The American Cancer Society estimates that more than 320,000 new cases of invasive breast cancer will be diagnosed in women this year, and more than 42,000 will die from the disease. Following diagnosis, most women undergo surgery. Many opt for breast-conserving surgery, lumpectomy, where as much of the breast as possible is left intact, rather than mastectomy, where the entire breast is removed.
— Beth Strohbusch of Milwaukee is a freelance writer and a strategic communications professional serving clients in the health care, health IT, tech, education, and nonprofit sectors.