Abstract P2-10-09: Can an AI-based Clinical Decision Support System (CDSS) in the treatment decision-making process for breast cancer contribute to healthcare cost reduction ?

Le résumé fourni par la source

Abstract Background: Reducing breast cancer treatment costs is vital for global patient care, especially in low-income countries, and alleviates the financial burden on healthcare systems, fostering better resource allocation for new therapies. Material and methods : The BreaCS consortium developed an AI-based CDSS for early breast cancer patients, targeting three questions at diagnosis : definitive tumor size (pT), nodal involvement and response to primary systemic therapy (PST). A multimodal fusion of the EUSOMA clinical data, biopsy pathology data and the preoperative MRI trained the AI based-model. The economic impact was assessed using the real-world EUSOMA database from the consortium, comprising 5,000 early breast cancer patients. The impact on healthcare costs was calculated based on the treatment of 1,200 new early breast cancer patients annually by the consortium. Only the costs of the unnecessary surgical procedures were calculated according to the Belgian reimbursement system excluding non-surgical costs, hospital admission and other related expenses. Results: Accurate prediction of the tumor size will refine decisions about breast-conserving surgery (BCS) versus mastectomy (ME), reducing the need for reconstructive surgeries. In our database, 42,5 % of patients had a superior pathological T stage compared to clinical staging. This led to 5% of patients requiring second breast surgery, with half undergoing a re-excision and the other half a ME. The annual cost for a second surgery for the consortium is estimated at 9000 euro. Conversely, 40,5 % of patients had a smaller pathological T stage compared to the clinical estimation. 25 % of these patients had a pT1 while presenting with a cT2-3. Considering that the cT3 patients (approximately 3%) received unnecessary ME’s, the cost for unnecessary reconstructions is estimated on a yearly basis to be 53,536 euro for the entire consortium. Prediction of lymph node involvement will eliminate the need for any axillary surgery in pN0 patients. Of the entire population 71,7% underwent a sentinel lymph node biopsy (SLNB), of which 40,4 % had pN0. Approximately 350 SLNB could have been avoided, related to a cost of 70,000 euros on a yearly basis. 18% of patients receiving PST had a pCR. In these cases, predicting response might allow omitting all surgical procedures. For patients with a pCR, 66.4% received SLNB, 42,7 % axillary LN dissection, 38,6 % a ME and 56,8 % a BCS. All these surgeries have a combined total cost of 13,800 euro. Conclusions: An AI-based CDSS that can predict tumor size, lymph node involvement and response to PST could reduce the surgical costs for early breast cancer treatment with approximately 150,000 euro on a yearly basis for 1200 patients. This excludes non-surgical costs, treatment of sequelae and impact on quality of life of the patients. Citation Format: Barbara Bussels, Isabelle Kindts, Sandra Steyaert, Philip Poortmans, Adelheid Roelstraete, Caroline de Beukelaer, Mona Bové, Sander Goossens, Peter De Jaeger. Can an AI-based Clinical Decision Support System (CDSS) in the treatment decision-making process for breast cancer contribute to healthcare cost reduction ? [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P2-10-09.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.