PS5-09-13: Evaluation of the diagnostic accuracy of ultra-fast ConfocAl imaging for rEtroareolar Limit in nIpple-sparing mAstectomy: Camelia project

Abstract Background Nipple-sparing mastectomy (NSM) followed by breast reconstruction is now widely recognized as a safe technique, including in patients with a history of radiotherapy or chemotherapy, regardless of tumor size or axillary status, except in cases of inflammatory breast cancer or direct involvement of the nipple-areola complex on imaging or clinical examination. The main oncological concern is the potential persistence of tumor remnants in the residual subareolar breast tissue. At least 10% of patients undergo a second surgery to excise this tissue. Such reinterventions may increase patient anxiety, negatively impact body image, raise healthcare costs, and delay adjuvant treatments. A new generation of ultra-fast confocal microscopes (UFCM) enables high-resolution imaging of fresh tissue in a shorter time than frozen section analysis, potentially guiding intraoperative assessment and avoiding reoperation [1-4]. Methods The CAMELIA project is a prospective, non-interventional, ex vivo clinical study analyzing retroareolar margins after NSM, imaged using UFCM and interpreted by pathologists and surgeons. The examined tissue is preserved for subsequent conventional histology (HES). At this stage, UFCM results do not influence surgical management. The primary objective is to use UFCM to provide an accurate intraoperative diagnosis of retroareolar tissue. Accuracy, sensitivity, and specificity are calculated based on the proportion of images correctly interpreted compared to definitive histology on corresponding HES slides, the reference standard. Results Since October 2024, 40 out of 194 patients who underwent Nipple-Sparing Mastectomy (NSM) have been included in this study. Of these, 8 had prophylactic mastectomies and 32 had therapeutic interventions. There were no bilateral procedures, and 16 patients underwent surgery after neoadjuvant chemotherapy. Importantly, four of these patients had positive margins. Three pathologists participated in the image analysis: two seniors (PT1 and PT2) and one junior (PT3). They analyzed 20 out of 40 images. Both PT1 and PT3 have experience with Ultra-Focalized Cancer Mammography (UFCM). Additionally, an experienced UFCM surgeon interpreted all 40 UFCM images. Images were categorized into four groups: deferred diagnosis, cancer, no cancer, or suspicious. For the first 20 cases, the pathologists' concordance rate ranged from 59% to 75%. Specifically: PT1 classified 2 out of 3 positive margins as suspicious, PT2 deferred diagnosis for 3 cases and classified 1 out of 3 as suspicious. PT3 classified 1 out of 3 positive margins as suspicious. The surgeon's concordance rate for the first 40 cases was 72.5%, with one positive margin identified out of four existing positive margins. On average, the intraoperative image acquisition time was 16 minutes. Readers rated the quality of these images at 3.4/5. The average image interpretation time across all specialists was 1 minute and 30 seconds. Conclusions Expected outcomes include high accuracy, sensitivity, and specificity. These results will provide proof of concept for cellular imaging of the retroareolar margin during NSM, with the aim of completing tumor resection if necessary and avoiding reoperation. References: 1. Conversano A. et al. BJS Open. 2023 May 5;7(3): zrad046. 2. Mathieu MC et al. Virchows Arch. 2025 Feb;486(2):299-311. 3. Mathieu MC et al. Life (Basel). 2024 Oct 28;14(11):1384. 4. Spoor J et al. Eur J Surg Oncol. 2024 Jun;50(6):108320. Citation Format: A. Conversano, N. Ben-Romdhane, N. Joyon, A. Puchar, A. Ilenko, A. Turpin, A. Alfaro, M. Mathieu, M. Abbaci. Evaluation of the diagnostic accuracy of ultra-fast ConfocAl imaging for rEtroareolar Limit in nIpple-sparing mAstectomy: Camelia project [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS5-09-13.