Abstract 145P
Background
Triple-negative breast cancer (TNBC) that do not achieve pathological complete response (pCR) have unfavorable prognosis. The RAD51 score is a functional assay able to identify Homologous Recombination Repair (HRR)-deficient tumors. In this setting, it may add prognostic value and guide post-neoadjuvant treatments.
Methods
We quantified RAD51 and BRCA1 foci by immunofluorescence, content of tumor-infiltrating lymphocytes (TIL) and expression of immune markers on diagnostic biopsies of 26 high-risk breast cancer (BC) patients (pts), namely TNBC or early onset BC (≤ 35 yo) or gBRCA1/2-mutated BC admitted at the University Hospital of Parma from 01/2011 to 03/2020. All pts received neoadjuvant chemotherapy (neoCT) with epirubicin, taxanes and cyclophosphamide. Functional HRR deficiency (HRD) was predefined as a RAD51 score ≤10% (RAD51-low).
Results
RAD51 was successfully scored in 26/29 (90%) samples. 16/26 (62%) tumors were RAD51-low (HRD). 14 pts presented HRR alterations: 4 gBRCA1, 2 gBRCA2 and 2 gPALB2 mutations and 6 BRCA1-low foci, surrogate of lack of BRCA1 function likely due to promoter hypermethylation. Median RAD51 score was 3.4 in HRR-mutated tumors and 19.2 in HRR-WT tumors (p=.01). Disease-Free Survival (DFS) at 4 years was 100% for pts who achieved pCR vs 67.5% for non-pCR tumors (p=.12). The addition of RAD51 status to pCR information improved the model capacity to predict DFS (ANOVA test, p=.05). Pts with HRD tumors by RAD51 showed a trend towards better DFS (HR=0.28, 95% CI 0.05–1.54, p=.14). 4/5 TIL-high tumors in this cohort were RAD51-low, suggesting a crosstalk between HRD and an active antitumor immune response. In support, RAD51-low/TIL-high tumors had higher CD20+ TIL (p=.01), lower CD3+ TIL (p=.02), higher PD-L1 Combined Positive Score (p=.03), and a trend towards higher PD1+ TIL (p=.05); no differences were found in FOXP3+ TIL. Moreover, only 1/4 RAD51-low/TIL-high tumors achieved pCR but none of them relapsed.
Conclusions
The RAD51 test is able to identify HRR-altered tumors, beyond gBRCA1/2 mutations, and to select a cohort of RAD51-low pts with better prognosis in a platinum-free neoCT setting. Biomarker analyses on treated paired tumors and on a larger cohort of pts are ongoing. Results will be available for the congress.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
B.P. was supported by ESMO with a Clinical Translational Fellowship aid supported by Roche. This work was supported by the project ERAPERMED2019-215.
Disclosure
B. Pellegrino: Non-Financial Interests, Institutional, Sponsor/Funding: Novartis; Financial Interests, Institutional, Writing Engagements: BMS; Non-Financial Interests, Institutional, Sponsor/Funding: Lilly. V. Serra: Non-Financial Interests, Institutional, Research Grant: AstraZeneca; Non-Financial Interests, Institutional, Research Grant: Tesaro; Financial Interests, Institutional, Sponsor/Funding: Abbie. A. Musolino: Non-Financial Interests, Institutional, Research Grant: EISAI; Non-Financial Interests, Institutional, Research Grant: Ltd.; Financial Interests, Personal, Sponsor/Funding: Roche; Financial Interests, Personal, Sponsor/Funding: MacroGenics; Financial Interests, Personal, Sponsor/Funding: Merck; Financial Interests, Personal, Sponsor/Funding: Lilly; Non-Financial Interests, Institutional, Research Grant: Pfizer. All other authors have declared no conflicts of interest.