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