Abstract Background: In France, access to PARP-inhibitors for metastatic breast cancer (mBC) is restricted to BRCA1, BRCA2 and PALB2 germline variant carriers. Genomic tumor sequencing or circulating tumor DNA assays (ctDNA or liquid biopsy (LB)) can incidentally reveal potential germline alterations. Currently, LB assays are nor designed nor validated for germline variant detection or interpretation, and it is not known whether their performances are good enough to replace a regular germline test. Our aim is to quantify the concordance between germline and LB in BC, and to explore the potential of LB to serve as a proxy for germline testing. Methods: Among patients (pts) with mBC enrolled between February 2021 and March 2025 in the STING molecular screening program (NCT04932525), pts with a pathogenic/likely pathogenic variant (P/LPVs) in BRCA1, BRCA2, and PALB2 genes in ctDNA were retrieved through the study database. LB in STING was carried out by using the FoundationOne Liquid CDx assay covering the analysis of 324 genes. We checked all the selected pts medical records to look for germline test results. We first performed descriptive analyses of clinical and pathological variables of the selected pts. We then assessed the positive predictive value (PPV) and false discovery rate (FDR) of ctDNA. Variant allele frequency (VAF) values were explored to evaluate their potential as a surrogate marker of germline origin, with a threshold of 40% VAF prospectively examined based on prior biological rationale. Results: Among 1,325 pts enrolled in STING program, 119 (9.0%) LB showed P/LPVs in BRCA1 (n=33; 27.7%), BRCA2 (n=71; 59.7%), or PALB2 (n=15; 12.6%). Germline genetic testing was performed in 93 of these pts (79.8%: 27 BRCA1, 56 BRCA2, and 12 PALB2).Among the entire cohort of 93 pts with BRCA1, BRCA2, or PALB2 mutations detected by LB and who underwent germline testing, 67 pts had confirmed germline P/LPV, resulting in an overall PPV of 72 %. Conversely, 26 pts were germline negative. In this subcohort, median age at diagnosis was 43.5 years (SD ± 12.1) for germline P/LPV carriers versus 45.6 years (SD ± 13.1) for somatic pts.Among the 27 LB BRCA1 pts, 21 (77.78%) pts showed concordance, all with a VAF 40%. LB FDR for BRCA1 variants was 22.2%. Among the 6 discordant pts, 5 pts harboured mutations with VAF 40% (median 4.9%), consistent with somatic origin, while one case involved a rearrangement intron 4 with a VAF of 42%.Among 54 LB BRCA2 pts, 39 (57.4%) showed concordance. Of these, 37 (94.9%) had VAF 40%, while 2 pts exhibited VAF between 26-38%. LB FDR for BRCA2 variants was 27.7%. Among the 15 discordant pts, 14 exhibited low VAF consistent with somatic origin; for one pt with exon 2 loss alteration, VAF was unavailable.Among 12 LB PALB2, 7 (58.3%) pts showed concordance, all with VAF 40%. LB FDR for PALB2 variants was 41.7%. Among the 5 discordant pts, all presented mutations with VAF 40% (median 0.28%) on LB, suggesting somatic variants. Interestingly, in our population, the PALB2 mutation rate was higher than previously reported. Conclusion: Our findings highlight a substantial concordance (PPV 98.5%) between germline and LB results for BRCA1, BRCA2, and PALB2 P/LPVs in mBC pts enrolled in the STING program, when applying a 40% VAF threshold. Although the overall FDR appears elevated, 92% of discordant cases involved variants with VAFs below 40%, which would not typically raise suspicion for germline origin in clinical practice. These results suggest that 40% VAF threshold could be a practical and reliable surrogate marker for germline status provided it is confirmed in larger independent cohorts. Data on LB false negative rates will be presented at the conference, to measure the likelihood of a positive germline test when the LB shows no PV and thereby evaluate the safety of omitting germline testing in LB-negative cases. Citation Format: A. Spata, V. Goldbarg, M. Brichard, M. Ibanez-Alda, I. Valle, F. Giugliano, T. Grinda, A. Viansone, C. Poisson, C. Bousrih, S. Akla, C. Roussel-Simonin, E. Rassy, A. Sabau, L. Antoun, C. Nicotra, M. NgoCamus, B. Verret, A. Bayle, E. Rouleau, B. Pistilli, A. Italiano, O. Caron, J. M. Ribeiro. Evaluating Liquid Biopsy as a Surrogate for Germline Testing in Metastatic Breast Cancer: Concordance and Discordance Analysis in the STING Program 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 PS3-01-06.
Spata et al. (Tue,) studied this question.