Breast Cancer Locator Noninferior to Wire Localization for Lumpectomy Margins

Use of the Breast Cancer Locator (BCL), an investigational supine MRI–based guidance system, was noninferior to standard wire localization (WL) and resulted in a 32% relative reduction in the positive margin rate (PMR) among patients undergoing breast-conserving surgery (BCS) for nonpalpable invasive breast cancer or ductal carcinoma in situ (DCIS), according to findings from a phase 3 trial (NCT04397185) published in the Annals of Surgical Oncology.

In the modified intent-to-treat population, the PMR was 14.4% (30/209) in the BCL group compared with 21.1% (44/209) in the WL group. In the primary analysis, which was stratified by tumor type on core biopsy, the treatment difference (TD) was –7.1% (95.4% CI, –14.3% to 0.1%), with a noninferiority P value of .011. However, the 1-sided P value for superiority (P = .025) slightly exceeded the prespecified significance level of .023, meaning superiority was not formally met in the primary analysis.

In a prespecified sensitivity analysis stratified by final tumor type on pathology, the TD was –7.5% (95.4% CI, –14.7% to –0.3%), and the PMR was significantly lower with BCL (superiority P = .019). Fewer patients in the BCL group underwent re-excision (12.9% vs 19.6%), a relative decrease of 34% (TD, –6.7%; 95% CI, –13.7% to 0.3%).

“BCL was noninferior to WL and superior in analysis stratified by final tumor type,” the study authors wrote. “Clinically meaningful relative reductions in PMR and re-excision rates were observed.”

Trial Design

The international, prospective, randomized trial was conducted at 23 sites in the US, Canada, the United Kingdom, and Austria, including both major academic medical centers and community hospitals. A total of 44 surgeons, a mix of general surgeons and breast specialists, participated. Between December 23, 2020, and September 9, 2025, 480 participants were enrolled; 23 were assigned to a roll-in population at sites without prior BCL experience, and 457 were randomly assigned 1:1 to BCL (n = 228) or WL (n = 229). Overall, 418 randomly assigned patients underwent surgery, with 209 in each group. Randomization was stratified by histologic diagnosis on core biopsy (invasive cancer vs DCIS) and by treating center.

Eligible patients were adult women with nonpalpable invasive breast cancer or DCIS on core biopsy who chose BCS and had an enhancing lesion of at least 8 mm on prone MRI. Patients with multifocal tumors were eligible, whereas those with multicentric disease, contraindications to MRI, pregnancy, prior neoadjuvant chemotherapy, or a need for more than 2 localization wires were excluded. The median age of randomly assigned patients was 62 years.

Patients in the BCL group underwent supine breast MRI, from which a patient-specific, 3D-printed, bra-like plastic form was fabricated. After anesthesia induction, surgeons placed the form on the breast, marked the projected tumor edges on the skin, and deployed bracketing wires through ports to delineate 1-cm margins. Lumpectomy was then guided by the wires, the skin markings, and an interactive 3D image of the tumor known as the Visualizer. Patients in the WL group had 1 to 2 wires placed before surgery. The primary end point was PMR, with a noninferiority margin of 2%, and pathologists were blinded to treatment arm.

Subgroup and Secondary Findings

The benefit appeared most pronounced in tumor types associated with difficult margins. Among patients with DCIS on final pathology, the PMR was 20.6% (7/34) with BCL vs 53.3% (16/30) with WL (TD, –32.8%; 95% CI, –52.1% to –9.1%), a 61% relative reduction. In invasive lobular carcinoma, the PMR was 9.1% (4/44) vs 25.6% (10/39), respectively (TD,

–16.6%; 95% CI, –32.9% to –0.2%), a 64% relative reduction. Among patients with extremely dense breasts, the PMR was 6.7% with BCL vs 55.6% with WL (TD, –48.9%; 95% CI, –75.0% to –11.9%). Subgroup analyses were not adjusted for multiplicity.

Shave excisions were performed less often with BCL (53% vs 63%; TD, –10.0%; 95% CI, –19.5% to –0.6%), and the mean number of shaves per patient was lower (1.7 vs 2.1). Total specimen volume was higher with BCL, at 68 mL vs 49 mL (ratio of means, 1.38; 95.4% CI, 1.23-1.55); investigators noted that this difference represented approximately 2.2% of an average total breast volume of 850 mL.

Safety and Operative Time

Adverse device effects occurred in 25.6% of patients in the BCL group and 28.3% of those in the WL group. The most common events with BCL vs WL included positive margins (13.7% vs 21.2%), breast infections (3.1% vs 1.9%), close margins leading to re-excision (0.4% vs 2%), and contact dermatitis attributed to Tegaderm dressing (6.2% vs 0%), which resolved in all cases after the dressing was removed. One pneumothorax in the BCL group was attributed to direct injection of anesthetic into the chest wall rather than to a BCL wire. Serious adverse device effects, which included all positive or close margins, occurred in 15% of BCL patients and 23% of WL patients. The seroma rate was higher with BCL (9.2% vs 4.7%).

Mean operative time was 98.5 minutes with BCL and 83.5 minutes with WL (TD, 13 minutes; 95% CI, 5.0-21.0). Among the 2 surgeons who performed at least 20 BCL cases, the mean operative time was 75 minutes with BCL vs 70 minutes with WL, a difference the investigators described as clinically insignificant.

The authors also noted that BCL guidance removes the need for a separate preoperative wire or seed placement procedure and does not require coordination with radiology on the day of surgery. They concluded that the probability of benefit is high, “particularly in patients with tumor types (DCIS and invasive lobular cancer) associated with higher PMRs.”

Reference

Gass JS, Pesek S, Czerniecki B, et al. An international randomized prospective trial comparing a supine MRI-based guidance system, the Breast Cancer Locator, to wire localization for breast-conserving surgery for non-palpable breast cancer. Ann Surg Oncol. Published online October 22, 2026. doi:10.1245/s10434-026-20657-6