Impact of pathogenic germline BRCA1/2 and PALB2 mutations and tumor aneuploidy in patients with HR+/HER2- metastatic breast cancer treated with CDK4/6 inhibitors.

Abstract

1075 Background: CDK4/6 inhibitors lack established predictors of efficacy in HR+/HER2- mBC. Pathogenic germline BRCA2 variants, somatic homologous recombination deficiency (HRD) and APOBEC signatures conferred worse survival in small series. Broader tumor DNA panels enable the detection of these and other parameters such as aneuploidy and tumor mutational burden. Methods: Clinical and molecular data from a single institution cohort of CDK4/6i-treated HR+/HER2- mBC pts were analyzed. ECOG PS was 0-1. Clinical variables were collected retrospectively. Primary endpoints were progression-free survival (PFS) and overall survival (OS). Tumor-only sequencing was performed on clinical samples with 161-gene OCAv3, 523-gene TSO500 or 555-gene lab-developed NGS panels. HRD was assessed with the Signature Multivariate Analysis (SigMA) tool using a likelihood-based measure to estimate single base substitution signature 3. Aneuploidy scores were defined as the fraction of covered arms harboring arm-level somatic copy number alterations as determined by ASCETS v.1.1. Results: 153 pts were available for analysis. Median age was 54 y (29–88). 74% were postmenopausal, 28% de novo metastatic and 67% had visceral disease. Palbociclib was used in 80%, ribociclib in 17% and abemaciclib in 3%, mostly as early lines (1L 65%, 2L 22%). Median follow up was 46 mo (7–84) and median time from sample collection to CDK4/6i was 0.6 y (-18.5–+2.6). 76% of samples have been sequenced. The median number of mutations (muts) was 3 (1–32), most commonly in PIK3CA (37%), GATA3 (26%) and TP53 (20%). ESR1, KRAS and ARID1A, but not RB1 muts were numerically enriched in post–CDK4/6i mets (n=8). TP53 (HR 1.85 CI 1.02–3.33) and GATA3 (HR 0.48, CI 0.28–0.83) muts were associated with PFS. Germline ( g) testing was available in 68 (44%) pts, with 21 pts carrying pathogenic variants including 11 BRCA2, 2 BRCA1 and 2 PALB2. Cox regression showed a strong negative impact of gBRCA1/2-PALB2 on PFS (HR 4.0 CI 2.0–7.9) and OS (HR 3.5 CI 1.6–7.7) adjusted for age, visceral disease and prior chemo. In 1L, mPFS/OS was 9.9/28.1 mo for gBRCA1/2-PALB2 carriers, 26.8/NR mo for non-carriers and 23.7/49.2 mo for untested (Cox HR PFS 5.1 CI 2.1-14.0; HR OS 4.0 CI 1.4-11.2). Pts with gATM/ CHEK2/ BRIP1 had a 1L PFS/OS>15/45 mo. Among 36 pts assessed for HRD, 22 were HRDlow and 14 HRDhigh with no significant survival association. 3/4 known gBRCA1/2 carriers were HRDhigh and 7/8 gWT HRDlow. In these pts, higher aneuploidy was associated with shorter PFS and OS (Cox HR PFS 2.0 CI 0.9-4.3, p=0.07; HR OS 2.6 CI 1.0-7.1; p=0.05). Conclusions: Pathogenic gBRCA1/2- PALB2 and higher aneuploidy were associated with shorter survival. Profiling of aneuploidy, HRD, TMB and APOBEC in all pts is ongoing including 55 baseline ctDNA samples and 361 samples from the GENIE BPC cohort. Results will account for FDR, left truncation and optimal cutoffs.