Background – MpBC is a rare, aggressive subtype with a dismal prognosis. Standard of care treatments are limited and novel therapies are needed. Recent clinical studies have demonstrated a response to immune checkpoint inhibitors. However, characterization of the immune and genomic environments in MpBC is paramount to inform new biomarkerselected strategies. Here, we investigated the TIME and somatic landscape of MpBC.

Methods – We retrospectively analyzed de-identified next-generation sequencing data from unique patients with a breast cancer diagnosis (n=13,512) in the Tempus database. We selected patients with MpBC (n=171, 1.3%) and non-MpBC (n=13,341,98.7%) histologies based on clinical documentation reported within 180 days of sample collection. For patients with multiple samples sequenced, the sample closest to the date of reported diagnosis was utilized. Tumors were sequenced with the Tempus xT DNA (648-gene panel) and/or xR RNA assays. Somatic alterations, immune cell infiltration predicted from gene expression patterns, PD-L1 from IHC, TMB (tumor mutational burden), and MSI (microsatellite instability) were evaluated. Wilcoxon rank-sum and Pearsons Chi-squared/Fisher’s exact tests assessed statistical significance (p<0.05, q<0.05 for false discovery rate correction for multiple testing).

Results – The MpBC and non-MpBC cohorts comprised a diverse population (White, 72% vs 73%, Black/African American, 12% vs 15%, Asian, 7.9% vs 4.4%, Other, 7.9% vs 7.7%) (p=0.5). The median age at diagnosis in the MpBC and non-MpBC cohorts were 61 and 57 years, respectively (p<0.001). Breast cancer receptor subtype distribution for MpBC vs non-MpBC was HR+/HER2- (20% vs 58%), TNBC (72% vs 24%), and HER2+ (1.9% vs 9.2%) (p<0.001). PD-L1 expression (CPS ? 10) was higher in the MpBC (35%) vs nonMbBC group (14%) (p<0.001). In MpBC, the proportion of M1 macrophages and neutrophils was higher vs non-MpBC (p<0.001 for both), and the proportion of B and NK cells was lower (p<0.001 for both). There was no clinically significant difference in the percentage of patients with TMB-H or MSI-H status (p>0.05 for both). The top five somatic alterations that were more frequent (q<0.001) in MpBC vs non-MpBC were: TERT (23% vs 1.3%), CDKN2A (23% vs 5.1%), CDKN2B (21% vs 5.1%), MTAP (16% vs 3.3%), and PIK3R1 (13% vs 3%). Additional potentially therapeutic-relevant alterations associated with PI3k pathway were prevalent in the MpBC and non-MpBC groups: PTEN (25% vs 10%, q<0.001), PIK3CA (26% vs 32%, q=0.2) and AKT1 (1.8% vs 3.7%, q=0.3). Within the MpBC group, there was a similar incidence of PD-L1 expression (CPS ? 10) in both TNBC (38%) and HR+/HER2- subtypes (33%). No significant difference was observed between the immune cell subsets within MpBC subtypes (p>0.05). Somatic alterations between TNBC and HR+/HER2- subtypes in the MpBC group were analyzed and there was a trend towards differences in PIK3CA (23% vs 39%) and PTEN (22% vs 33%) (q=0.6 for both).

Conclusion – In this large, real-world analysis, patients with MpBC displayed a distinct molecular phenotype compared to non-MpBC patients. In patients with MpBC, TNBC was more common, whereas the HER2+ subtype was rare. Patients with MpBC had higher PD-L1 expression and therapeutically relevant alterations, including those within the PI3k pathway, were frequently encountered in MpBC. Although limited by sample size, this is one of the first studies to compare the molecular phenotypes between subtypes within MpBC. These findings are hypothesis-generating and provide further rationale for developing novel combinatorial therapeutic clinical trial strategies for MpBC.

VIEW THE PUBLICATION

VIEW THE POSTER