Some may think that prediction of metastasis is meaningless since metastatic breast cancer is currently incurable. We argue that effective identification of developing metastasis will enable us to design and conduct clinical trials specifically targeting those patients at high risk. The current study sought to generate the KAM score by 4 genes (BRSK2, EYA1, SIGLEC15, and AGTR1) overexpressed in primary breast cancer that developed metastasis to bone compared with matched controls without metastasis longer than 10 years. A high KAM score was prognostic of poor overall (OS), disease free survival (DFS), and disease specific survival (DSS) in the METABRIC, and OS in the GSE96058 cohorts. A high KAM score was significantly associated with clinical aggressiveness, such as high American Joint Committee Cancer (AJCC) stage, lymph node metastasis, Nottingham pathological grade, and triple negative breast cancer (TNBC). Subgroup analysis revealed that a high KAM score was associated with worse OS in ER-positive/HER2-negative breast cancer in both cohorts. A high KAM breast cancer enriched all 5 cell proliferation-related gene sets of the Hallmark collection and interferon (IFN)-γ response gene sets. Furthermore, a high KAM breast cancer was significantly infiltrated with a high fraction of not only anti-cancer but also pro-cancer immune cells and associated with high level of cytolytic activity. Finally, a high KAM breast cancer was significantly associated with lung metastasis. In conclusion, we developed KAM score using 4 gene expressions that predict lung metastasis and patient survival in breast cancer.
Keywords: Breast cancer; KAM score; gene sets; metastasis; survival; tumor microenvironment.
AJCR Copyright © 2021.