Abstract PS3-06-06: A comparison of prognostic predictors in node-positive hr+/her2- breast cancer

Abstract Background: RxPONDER showed that Oncotype DX (ODX) can assist with adjuvant chemotherapy selection in postmenopausal women, but in premenopausal patients, treatment benefit prediction is unresolved. RSClinN+, an online calculator that adds clinical factors to ODX, was introduced to close this gap. However, RSClin+ lacks independent validation, and requires prior ODX testing, which risks tissue exhaustion, financial burdens, and significant turnaround times. Ataraxis Breast RISK (ATX) is a multimodal artificial intelligence prognostic test that natively combines H&E-derived morphology with clinicopathologic data and predicts risk of recurrence (RoR). We compared its performance with ATX in a real-world cohort. Methods: We retrospectively analyzed 112 HR+/HER2- patients with 1-3 positive lymph nodes from Basel University Hospital who previously underwent ODX testing. 12 patients (11%) experienced breast cancer recurrence. RSClinN+ scores were computed with the online tool. ATX scores were generated using a locked model using baseline diagnostic H&E slide and clinical data. Prognostic accuracy of ODX, RSClin+ and ATX was assessed using Harrell’s C-index (concordance index). Results: ATX was significantly associated with overall recurrence (p<0.01) with a hazard ratio of 1.67 (95% CI: 1.44-2.44). In a head-to-head comparison, ATX continuous risk score was more accurate with a C-index of 0.74 (0.61-0.86), compared to 0.67 (0.50-0.84) for RSClinN+ and 0.53 (0.35-0.71) for ODX. In multivariable Cox analysis including both ATX and RSClin, ATX added independent prognostic information (LR χ2 = 4.81, p=0.028). Using an exploratory threshold of 10% risk of 5-year recurrence, the high-risk group (N=64) had a recurrence rate of 18.8%, while there were no recurrences in the low-risk group (N=48). Conclusions: ATX and RSClinN+, two multi-modal prognostic tools, were more accurate than ODX alone. ATX had the greatest prognostic accuracy, outperforming both RSClin+ and ODX in a real-world cohort of N+ patients. Citation Format: E. Chiru, S. Muenst, C. Kurzeder, S. Ebner, F. Schwab, B. Seifer, M. Vetter. A comparison of prognostic predictors in node-positive hr+/her2- 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 PS3-06-06.