Abstract PS4-08-07: Predicting response and survival after neoadjuvant systemic treatment with on-treatment biopsies

Abstract Background: Early prediction of response to neoadjuvant systemic treatment (NST) in breast cancer (BC) may guide treatment tailoring even before NST completion. Patients (pts) could avoid ineffective treatments and be assigned to non-cross-resistant therapies. Here, we analysed core needle on-treatment biopsies (OTB) taken after 2-4 cycles of NST for residual invasive cancer, Ki67, stromal tumor-infiltrating lymphocytes (sTILs) and their ability to predict treatment outcome. Study Design: Pts with available and evaluable OTBs (n=449) of neoadjuvant clinical trials GeparSixto (n=47), GeparSepto (n=194), GeparNuevo (n=99), GeparX (n=47), and GeparOLA (n=62) were included in the analysis. OTBs were retrospectively evaluated for the presence of tumorbeds, residual cancer defined as cellularity of viable invasive cancer cells >0 (OTB+) vs 0 (OTB-), Ki67, and sTILs, which was thereafter correlated with post-treatment response (pCR defined as ypT0/is ypN0) and time-to-event outcomes. Results: Most tumors were cT2 (55%), of no special type (88%), G3 (66.1%), and 69.6% of pts were clinically nodal negative. Most prevalent subtype at baseline was TNBC (n=257, 57.2%), followed by HR+/HER2- (n=108, 24.1%), and HRany/HER2+ BC (n=84, 18.7%). In 79.1% (355/449) a tumorbed in OTBs was identified. The remaining OTBs had unspecific (10%) or uncertain changes (6.2%) or no signs of tumorbed (4.7%). Pts with OTB+ were identified in 60% (213/355) and had a pCR rate of only 20.2% (43/213) after completion of NST, whereas pts with OTB- (40%, 142/355) had a pCR rate of 73.9% (105/142). In univariate logistic regression analysis, pts with OTB+ had a lower pCR rate (OR 0.091 CI 0.055-0.152, p<0.001). In multivariate analysis adjusting for other established prognostic factors like nodal stage, intrinsic subtype and grading, the presence of OTB+ still showed a statistically significant decrease in the odds for pCR (OR 0.111 CI 0.063-0.196, p<0.001). When analysed according to subtype, the pCR rate at surgery for pts with OTB+ was 48% for HER2+ (12/25), 21.6% for TNBC (24/111), and 9.1% for HR+/HER2- (7/77). Significantly lower chances for pCR in pts with OTB+ were consistent in all intrinsic subtypes (HRany/HER2+ OR 0.24 CI 0.07-0.80 p=0.02; HR+/HER2- OR 0.13 CI 0.03-0.52 p=0.004, TNBC OR 0.09 CI 0.05-0.18 p<0.001, interaction test p=0.468). OTB+ was linked to worse DFS and OS (DFS: HR 2.59 CI 1.72-3.91; OS: HR 3.99 CI 2.09-7.59, log-rank p<0.001 respectively), remaining significant in multivariate analysis (DFS: HR 2.37 CI 1.52-3.68; OS: HR 3.76 CI 1.91-7.40, p<0.001 respectively). In subgroup analysis, the overall results for DFS were confirmed for all intrinsic subtypes (HRany/HER2+ HR 2.77 CI 0.919-8.7 log-rank p=0.067; HR+/HER2- HR 3.82 CI 0.922-15.9, log-rank p=0.065; TNBC HR 2.19 CI 1.33-3.60 log-rank p=0.003, interaction test p=0.774). In multivariate analysis an impact on OS could only be demonstrated for TNBC (HR 4.17 CI 1.92-9.07, log-rank p<0.001), probably limited in other subtypes by paucity of OS events. Baseline Ki67 was lower in OTB+ (median 39%) vs OTB- (median 48%, p<0.001). Overall, Ki67 values decreased from baseline to OTB by a median absolute value of 30%. The highest decrease appeared in pts with high-grade TNBC (median absolute value of 37,5%, stratified signed rank test p=0.03). Pts with OTB+ had lower sTILs at baseline (median 10%) as compared to OTB- (median 20%, p<0.001). Pts with pCR had a median absolute increase of 5% in sTILs from baseline to OTB, whereas pts with no pCR showed no changes in sTILs (p-value 0.122). Conclusion: Our findings suggest that OTB are suitable to monitor neoadjuvant therapy response. Residual cancer cells have a prognostic impact, even if detected early during treatment, therefore OTB should be evaluated prospectively as biomarker to guide response adapted treatment in neoadjuvant clinical trials. Citation Format: C. Denkert, A. Schneeweiss, V. Schaser, M. Untch, M. Frank, M. van Mackelenbergh, J. Blohmer, F. Holms, V. Müller, D. Zahm, T. Karn, P. Fasching, E. Simon, F. Marmé, M. Darsow, E. Stickeler, S. Jud, T. Fehm, B. Heinrich, C. Schem, B. Felder, J. Holtschmidt, S. Loibl, A. Noske. Predicting response and survival after neoadjuvant systemic treatment with on-treatment biopsies [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 PS4-08-07.