Abstract PS5-10-08: Advance-bc: a prospective single-center registry to capture real-world outcomes in advanced breast cancer

Abstract Background: Advanced Breast Cancer (ABC) remains a leading cause of cancer mortality among United States women, yet real world outcomes and quality of life data remain underreported. Population-level studies are often limited by incomplete data capture and lack of quality-of-life variables. We present an updated, comprehensive analysis of real-world clinical outcomes that integrates recent therapeutic advances and key social determinants of health. Methods: Data from 361 MD Anderson patients was summarized as counts (%) for categorical variables. Group differences were assessed using χ2/Fisher’s exact tests and t-tests/ANOVA or Wilcoxon/Kruskal-Wallis, as appropriate. Comparisons between subgroups utilized chi-square tests. Overall and progression-free survival (OS and PFS) was estimated with Kaplan-Meier curves and compared by log-rank tests. Cox proportional-hazards models—adjusting for age, subtype, stage, ECOG status, therapy line, and any variable with univariate P < 0.10—were used to identify independent predictors; proportional-hazards assumptions were tested. Missing data <10 % was handled by complete cases; others with multiple imputation. All two-sided analyses (α = 0.05) were conducted in R 4.4 and independently validated. Results: Of 361 patients, 99% (n = 360) were female, 1% male (n=1). Median age was 62 (range: <49 = 19%, 50-69 = 51%, ≥70 = 27%). Most were married (73%), and 54% lived within 150 miles of the center. The cohort was 81% White, 9% Black, 6% Asian, and 5% Hispanic. Subtype classifications were 70% HR+/HER-, 19% TNBC, and 11% HER2+. At diagnosis, 36% presented with de novo stage IV disease. Metastatic sites included bone (47%), liver (19%), lung (19%), and lymph nodes (15%). At last assessment, 75% had an ECOG of 0 or 1. Genomic data was available for 13%. Advance-care-planning notes were present for 32%, and 40% participated in an interventional clinical trial since their ABC diagnosis. In multivariate Cox analysis, younger age predicted worse survival (HR 0.97 per year, p = 0.003). Genomic testing trended toward improved outcomes (HR 0.50, p = 0.35). ER, PR, and stage were not independently prognostic. Conclusions: This analysis highlights key clinical and demographic features and survival outcomes in a real-world ABC cohort. However, structured quality-of-life, toxicity, and patient-reported outcomes—essential to modern ABC care—remain largely absent from standard EHR fields. Many relevant data elements reside in unstructured text, limiting extraction. Future work will leverage large language models (LLMs) to extract and analyze these data at scale. Additionally, the single-center design and demographic homogeneity limit generalizability. To improve representation and impact, we plan to expand the registry through partnerships with community hospitals. Citation Format: A. J. Kaler, A. Singareeka, G. Kirklin, A. Anderson, J. Harris, T. Jacobsen, B. Lim. Advance-bc: a prospective single-center registry to capture real-world outcomes in advanced breast cancer [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 PS5-10-08