Background: TALA is a poly(ADP-ribose) polymerase inhibitor approved as monotherapy for adult pts with gBRCA1/2-mutated HER2-negative locally advanced/metastatic BC. We report biomarker analyses from a phase 2, nonrandomized, single-arm, open-label study (NEOTALA; NCT03499353) evaluating the efficacy and safety of TALA in the neoadjuvant setting for pts with early gBRCA1/2-mutated HER2− BC. Efficacy and safety results are presented separately. Methods: The biomarker analysis population was all pts treated with TALA for whom biomarker results are available. To support molecular eligibility, blood was tested using BRCAnalysis CDx (Myriad Genetics). Baseline tumor tissue was retrospectively tested using FoundationOne CDx, with BRCA1/2 zygosity assessed using somatic-germline-zygosity (SGZ; Sun et al. JCO PO, 2018). Germline mutational status of 14 non-BRCA DNA damage response (DDR) genes was retrospectively assessed in baseline saliva samples using Ambry CustomNext-Cancer. Mutations were defined as known/likely pathogenic/deleterious variants, including copy number alterations (CNAs). Association between mutational status of MYC or RAD21 and primary endpoint pathological complete response (pCR) as per Independent Central Review was investigated with logistic regression. Results: Of 52 evaluable tumor samples from 61 treated pts, 39 (75%) and 13 (25%) pts exhibited BRCA1 and BRCA2 mutations, respectively; 1 (2%) pt exhibited mutations in both genes, and 1 (2%) pt had mutations in neither. BRCA loss of heterozygosity (LOH) was seen in 42/43 (98%) evaluable BRCA-mutant tumors. Of 45 pts evaluable centrally for both germline and tumor, 44/45 (98%) pts exhibited the same BRCA mutation in tumor as originally detected in germline, with the remaining pt exhibiting a gBRCA1 mutation, but lacking a tumor BRCA mutation. None of 49 saliva-evaluable pts exhibited non-BRCA germline DDR mutations. TP53 (51 [98%] pts) was the most frequently mutated gene in tumors. MYC and RAD21 (each 14 [27%] pts) were the most frequent CNAs. No evidence of association between mutational status of MYC or RAD21 and pCR was found (odds ratio=0.39, 95% CI 0.12-2.30). Based on a cutoff of ≥16%, genomic LOH was elevated in 24/27 (89%) tumors evaluable for both gLOH and pCR, precluding assessment of the potential association of gLOH high/low status with pCR. Conclusions: Tumor BRCA mutations were evident in nearly all pts in the biomarker analysis population, with BRCA LOH evident in all but 1 BRCA-mutated tumor. No pts had non-BRCA germline DDR gene mutations; tumor TP53 mutations were near-universal. MYC and RAD21 each exhibited CNAs in 27% of tumors, with no association with pCR. These results support the central role of BRCA mutations in tumor pathobiology in this indication. Clinical trial information: NCT03499353