Background: Immune checkpoint inhibitor (ICI) based regimens are standard of care for patients with TNBC and clinical trials are investigating ICI in HR+/HER2- breast cancer, with encouraging results in a subset of patients. The tumor immune microenvironment (TIME) plays a critical role in ICI response and can differ significantly by metastatic site. Given the expanding integration of ICIs in HR+/HER2- breast cancer trials, we examined the TIME across sites of disease and subtypes to illustrate how these anatomical sites are molecularly different and may impact ICI benefit.

Methods: We used Tempus Lens (Tempus AI, Inc., Chicago, IL) to retrospectively analyze de-identified next-generation sequencing data from patients with HR+/HER2- breast cancer (n=6,818) and TNBC (n=2,715) in the Tempus Database. Metaplastic histology was excluded. Among HR+/HER2- breast cancer and TNBC, tumors from primary breast (PB, n=4451), liver (n=1117), bone (n=845), lymph node (LN, n=425), lung (n=388), and brain/CNS (n=234) metastatic sites were sequenced with Tempus xT DNA (648-gene panel) and xR RNA assays. TMB, MSI, PD-L1 (CPS, clone 22c3), and proportions of B, T (CD4+, CD8+), NK cells, and macrophages of immune cells via quanTIseq deconvolution were compared across metastatic sites in HR+/HER2- breast cancer and compared to TNBC. Chi-squared/Fisher’s exact or Kruskal-Wallis tests were used to assess statistical significance.

Results: In this study, the median age at diagnosis in the HR+/HER2- and TNBC cohort was 58 and 57, respectively and both cohorts were diverse: White (50%/44%), African American (8%/14%), Hispanic (6.7%/7%), Asian (3.1%/2.3%), other race (5.5%/5.5%) and unknown (33%/34%). Overall, 67% of patients were stage IV. Among patients with HR+/HER2- breast cancer, brain/CNS (14%) was most commonly TMB high (≥10 mut/mB) followed by bone (9.4%) and liver (8.9%) (p<0.001). PD-L1 >=10% (positive) proportions were highest in PB (4.1%), lung (3.6%) and LN (4.9%) (p<0.001). LN had the highest proportions of B cells and T cells (CD4+ and CD8+) and the lowest proportion of macrophages (M1 and M2) and NK cells (p<0.001). Besides LN, CD8+ T cells were in higher proportions in lung and PB compared to liver, bone, and brain/CNS metastases (p<0.001). In TNBC, PB, lung, and LN metastases had the highest proportions of PD-L1 positive status as well as CD8+ T cell proportions compared to liver, bone, and brain/CNS (both, p<0.001). Compared to TNBC, HR+/HER2- breast cancer had different proportions for all immune cell populations in PB and all non-breast sites (exception NK cells, all else p<0.05). Notably, HR+/HER2- breast cancer had lower proportions of CD8+ T cells and M1 macrophages and higher proportions of M2 macrophages across PB and all non-breast sites compared to TNBC (all, p<0.001). Liver metastases had the highest proportions of M1 macrophages, NK cells, and neutrophils compared to all other sites in HR+/HER2- breast cancer, with a similar pattern seen in TNBC (all, p<0.001).

Conclusions: In HR+/HER2- and TNBC, the TIME significantly differed according to site of disease. Although more commonly TMB high, liver, bone, and brain/CNS metastases were composed of lower proportions of CD8+ T cells in HR+/HER2- breast cancer. HR+/HER2- breast cancer was less immunogenic across all disease sites compared to TNBC. However, patients with breast, lung, and LN metastases had the highest proportions of CD8+ T cells in both subtypes, indicating a subset of patients with HR+/HER2- breast cancer that could be enriched for ICI response. These analyses may support TIME biomarkers for patient selection in ICI focused clinical trials for patients with breast cancer.

VIEW THE PUBLICATION

VIEW THE POSTER