Background: As the available information about breast cancer is growing every day, the decision-making process for the therapy is getting more complex. ChatGPT as a transformer-based language model possesses the ability to write scientific articles and pass medical exams. But is it able to support the multidisciplinary tumor board (MDT) in the planning of the therapy of patients with breast cancer? Material and Methods: We performed a pilot study on 10 consecutive cases of breast cancer patients discussed in MDT at our department in January 2023. Included were patients with a primary diagnosis of early breast cancer. The recommendation of MDT was compared with the recommendation of the ChatGPT for particular patients and the clinical score of the agreement was calculated. Results: Results showed that ChatGPT provided mostly general answers regarding chemotherapy, breast surgery, radiation therapy, chemotherapy, and antibody therapy. It was able to identify risk factors for hereditary breast cancer and point out the elderly patient indicated for chemotherapy to evaluate the cost/benefit effect. ChatGPT wrongly identified the patient with Her2 1 + and 2 + (FISH negative) as in need of therapy with an antibody and called endocrine therapy “hormonal treatment”. Conclusions: Support of artificial intelligence by finding individualized and personalized therapy for our patients in the time of rapidly expanding amount of information is looking for the ways in the clinical routine. ChatGPT has the potential to find its spot in clinical medicine, but the current version is not able to provide specific recommendations for the therapy of patients with primary breast cancer.

Purpose: Large language models (LLM) show potential for decision support in breast cancer care. Their use in clinical care is currently prohibited by lack of control over sources used for decision-making, explainability of the decision-making process and health data security issues. Recent development of Small Language Models (SLM) is discussed to address these challenges. This preclinical proof-of-concept study tailors an open-source SLM to the German breast cancer guideline (BC-SLM) to evaluate initial clinical accuracy and technical functionality in a preclinical simulation. Methods: A multidisciplinary tumor board (MTB) is used as the gold-standard to assess the initial clinical accuracy in terms of concordance of the BC-SLM with MTB and comparing it to two publicly available LLM, ChatGPT3.5 and 4. The study includes 20 fictional patient profiles and recommendations for 5 treatment modalities, resulting in 100 binary treatment recommendations (recommended or not recommended). Statistical evaluation includes concordance with MTB in % including Cohen’s Kappa statistic (κ). Technical functionality is assessed qualitatively in terms of local hosting, adherence to the guideline and information retrieval. Results: The overall concordance amounts to 86% for BC-SLM (κ = 0.721, p < 0.001), 90% for ChatGPT4 (κ = 0.820, p < 0.001) and 83% for ChatGPT3.5 (κ = 0.661, p < 0.001). Specific concordance for each treatment modality ranges from 65 to 100% for BC-SLM, 85–100% for ChatGPT4, and 55–95% for ChatGPT3.5. The BC-SLM is locally functional, adheres to the standards of the German breast cancer guideline and provides referenced sections for its decision-making. Conclusion: The tailored BC-SLM shows initial clinical accuracy and technical functionality, with concordance to the MTB that is comparable to publicly-available LLMs like ChatGPT4 and 3.5. This serves as a proof-of-concept for adapting a SLM to an oncological disease and its guideline to address prevailing issues with LLM by ensuring decision transparency, explainability, source control, and data security, which represents a necessary step towards clinical validation and safe use of language models in clinical oncology.

Background: Invasive lobular breast carcinomas (ILC) have different histological features compared to non-special type carcinomas (NST), but the effect of histological subtypes on survival is controversial. In this study, we compared clinicopathological characteristics and outcomes between ILC and NST based on a large pooled data set from three adjuvant breast cancer trials (SUCCESS A, B, and C) and investigated a potential differential effect of recurrence risk related to nodal stage on survival. Methods: From 2005 to 2017, the large randomized controlled SUCCESS A, B, and C trials enrolled 8190 patients with primary, intermediate-to-high-risk breast carcinoma. All patients received adjuvant chemotherapy, and endocrine and/or HER2-targeted treatment was given where appropriate. Survival outcomes in terms of disease-free survival (DFS), overall survival (OS), breast cancer-specific survival (BCSS), and distant disease-free survival (DDFS) were estimated using the Kaplan–Meier method and analyzed using log-rank tests as well as univariable and adjusted multivariable Cox regression models. Results: In the SUCCESS trials, 6284 patients had NST and 952 had ILC. The median follow-up time was 64 months. ILC patients were older, more likely to receive mastectomy, and more likely to have larger tumor sizes, lymph node infiltration, hormone receptor-positive, HER2neu-negative, and luminal A-like tumors than NST patients. In the overall cohort, no significant differences between ILC and NST were detectable regarding the four survival endpoints, with hazard ratios obtained in adjusted multivariable cox regressions of 0.96 (95% CI 0.77–1.21, p = 0.743) for DFS, 1.13 (95% CI 0.85–1.50, p = 0.414) for OS, 1.21 (95% CI 0.89–1.66, p = 0.229) for BCSS, and 0.95 (95% CI 0.73–1.24, p = 0.689) for DDFS. However, a differential effect of nodal stage on survival was observed, with better survival for ILC patients with pN0/pN1 tumors and worse survival for ILC patients with pN2/pN3 tumors compared to NST patients. Conclusions: Our results revealed that ILC was associated with worse survival compared to NST for patients at high risk of recurrence due to advanced lymph node infiltration. These findings should be taken into account for treatment decisions and monitoring.

The current standard for the preoperative marking of non-palpable breast lesions is wire guided localization (WGL) which is associated with logistical efforts and patient discomfort. Non-radioactive seeds (NRS) recently challenged the use of WGL; but do they provide a better alternative from a logistical and environmental perspective? WGL standard was compared with NRS available in Germany: Magseed®, Pintuition®, SAVI SCOUT ® and LOCalizer™ on a logistical and carbon-footprinting basis. In the logistical analysis the number of patient contacts with the healthcare system for lesion localization/removal and the number of breast punctures were evaluated in two different clinical scenarios (primary surgery and secondary surgery after neoadjuvant treatment). The carbon footprints of WGL and NRS (with exception of LOCalizer) were assessed based on their material compositions and operating energy in a streamlined approach. Application of NRS reduces the number of contacts by 33.3 % (2 vs. 3) in primary, by 50 % (2 vs. 4) in secondary surgery, and the number of breast punctures by 33.3 % (2 vs.3). Annual Germany-wide material- and energy-based carbon footprints of NRS (1.6-3.2 tons CO₂eq) are significantly lower in comparison to WGL (10.3 tons CO₂eq). The implementation of NRS would lead to a CO₂eq reduction by around 79 % compared to WGL. The use of NRS for the localization of non palpable breast lesions is more favorable from the environmental and logistical perspective, when compared to WGL with possible benefits for patients, healthcare providers and the environment.

With the increasing amount of information related to breast cancer (BC) management, artificial intelligence (AI) has emerged as a tool with the potential to enhance the quality of treatment through the efficient integration of large datasets; however, the specific areas for which AI may be ready for clinical implementation remain unclear. In this narrative review, we recapitulate the available data on AI utilization in BC treatment by focusing on surgical therapy, radiation therapy, systemic and supportive treatment, but including the diagnostics, too. While AI has been implemented successfully in mammography screening, preoperative consultation, and radiation oncology, its use intraoperatively, post-operatively, and in systemic and supportive treatment is still in development. AI has potential to improve care, but since the accuracy of AI varies, careful consideration of its benefits and limitations is necessary.

Background/Objectives: Young women frequently present with more aggressive breast cancer tumors. This retrospective study analyzed the oncological outcomes of patients under the age of 40 according to the tumor biology. Methods: Group comparisons were performed via the log-rank test. Recurrence and survival rates are presented according to the Kaplan–Meier method. Results: In total, 88 women (mean age 36) were included, but two presented with bilateral cancer, resulting in 90 tumors. Triple-negative carcinoma was most common, with 26.7% (n = 24); 11.1% (n = 10) were luminal A; 23.3% (n = 21) were luminal B HER2-negative; 15.6% (n = 14) were luminal B HER2-positive; and 6.7% (n = 6) were HER2-positive (non-luminal). Moreover, 26.1% (n = 23) of patients experienced recurrence (mean 40 months), with the highest recurrence rate in the HER2-positive (50%) and triple-negative (30.4%) groups. The 3- and 5-year recurrence-free survival rates were 84.9% and 77.3%, and the overall survival rates were 93.1% and 90.3%, respectively. No statistically significant differences in oncological outcomes were observed (p = 0.164). Conclusions: The results show that young women tend to have triple-negative and fast-growing breast carcinomas, with worse overall survival in the triple-negative group. More research is needed on the pathomechanisms of breast cancer development in young women, especially those leading to disease progression and resistance to therapy.