First AI-Enabled Imaging Device for Breast Cancer Surgery Granted FDA Premarket Approval
The U.S. Food and Drug Administration (FDA) has granted premarket approval to Claire, an AI-enabled imaging device designed for assessment of intraoperative breast cancer margins. The device, which was previously called the Perimeter OCT B-Series with ImgAssist AI 2.0, had received a prior Breakthrough Device designation from the FDA.
Claire assists surgeons in detecting difficult-to-see tumors during breast-conserving surgeries to reduce the need for re-operations due to residual disease. The device uses wide-field optical coherence tomography, for 10 times higher resolution than standard X-ray and ultrasound at only 2 mm imaging depth, plus a proprietary AI detection algorithm that was trained on over 2 million breast tissue images to evaluate the sub-surface microstructure of tissues for residual disease. This allows surgeons to determine during the surgery if more tissue should be removed before the procedure is completed.
"Despite progress in breast cancer treatment, intra-operative margin assessment remains challenging, often leading to excess removal of healthy tissue, re-operations, and anxiety for patients as they await pathology results," said Alastair Thompson, BScHons, MBChB, MD, FRCSEd, FACS, Surgeon and Professor, Section Chief of Breast Surgery, Olga Keith Wiess Chair of Surgery at Baylor College of Medicine, Breast Cancer Program Leader at the Dan L Duncan Comprehensive Cancer Center, in a statement. He was the primary principal investigator of the pivotal trial that supported Claire's premarket approval application. "Claire has the potential to become a new standard in breast surgical care, helping reduce re-excisions while improving patient outcomes."
Each procedure performed with the device is also used towards further improvements of the AI-assisted device for even better outcomes for patients.
As part of the premarket approval, the FDA authorized a predetermined change control plan for planned AI enhancements to Claire without further FDA review.
Supporting Data
Findings from Claire’s pivotal trial were presented at the 2025 American Society of Breast Surgeons Annual Meeting (Abstract 2038296). The prospective, multicenter, randomized, controlled clinical trial explored the use of Claire to address positive pathologic margins after standard intraoperative assessments during lumpectomy vs standard intraoperative assessments alone. The primary endpoint was the occurrence of at least one unaddressed positive margin per subject in the device arm to assess effectiveness.
A total of 613 women with biopsy-confirmed breast carcinoma were enrolled in the study between January 25, 2022, and September 23, 2024; 208 women were randomly assigned to the device arm, and 206 were evaluable for effectiveness analyses.
Standard-of-care assessments evaluated 1,735 margins and left 56 margins positive and unaddressed in 35 patients (17%) in the device arm. The AI-assisted device evaluated 1,230 margins and determined that 115 margins needed further cutting (9.3%). Claire correctly detected residual disease in 14 of 35 patients (40%), leading to seven of these patients (20%) being fully cleared of all their residual disease. Only 28 patients (13.5%) were left with unaddressed positive margins (P = .005).
The average total lumpectomy tissue volume excised was 74 cm3 in the device arm—76.4% of that volume was from the primary lumpectomies, 19.9% from standard-of-care additional cuts, and 3.8% from device-directed shaves.
Overall margin accuracy was 88.1% and the study met its super-superiority performance goal.
No unanticipated device-related or serious adverse events were reported in the study.
DISCLOSURE: For disclosures from the study authors, visit breastsurgeons.org.
ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.
Performance of a convolutional neural network in determining differentiation levels of cutaneous squamous cell carcinomas was on par with that of experienced dermatologists, according to the results of a recent study published in JAAD International.
“This type of cancer, which is a result of mutations of the most common cell type in the top layer of the skin, is strongly linked to accumulated [ultraviolet] radiation over time. It develops in sun-exposed areas, often on skin already showing signs of sun damage, with rough scaly patches, uneven pigmentation, and decreased elasticity,” stated lead researcher Sam Polesie, MD, PhD, Associate Professor of Dermatology and Venereology at the University of Gothenburg and Practicing Dermatologist at Sahlgrenska University Hospital, both in Gothenburg, Sweden.