Abstract
BACKGROUD: There were limitations existing in programmed cell-death ligand 1 (PD-L1) as predictive biomarkers for breast cancer (BC), hence exploring the correlation between PD-L1 levels and other biomarkers in BC may become a very useful therapeutic clinical tool. METHODS: A total of 301 Chinese patients with different BC subtypes including 47 HR+/HER2+, 185 HR+/HER2-, 38 HR-/HER2+, and 31 triple-negative breast cancer (TNBC) were enrolled in our study. Next-generation sequencing based Yuansu450 gene panel was used for genomic alteration identification and PD-L1 expression was tested using immunohistochemistry. RESULTS: The most prevalent BC-related mutations were TP53 mutations, followed by mutations in PIK3CA, ERBB2, CDK12, and GATA3 in our Chinese cohort. We found that mutations DDR2 and MYCL were only mutated in HR-/HER2+ subtype, whereas H3-3A and NRAS mutations were only occurred in HR-/HER2- subtype. The percentage of patients with PD-L1-positive expression was higher in patients with HR-/HER2- mainly due to the percentage of PD-L1-high level. Mutational frequencies of TP53, MYC, FAT4, PBRM1, PREX2 were observed to have significant differences among patients with different BC subtypes based on PD-L1 levels. Moreover, a positive correlation was observed between TMB and PD-L1 level in HR+/HER2- subtype, and showed that the proportion of patients with high PD-L1 expression was higher than that of patients with low PD-L1 expression in the HR+/HER2- and HR+/HER2+ cohorts with high Ki67 expression. CONCLUSIONS: The genomic alterations based on PD-L1 and other biomarkers of different cohorts may provide more possibilities for the treatment of BC with different subtypes.