Abstract
Antitumor immunotherapy can enable promising and durable responses following their clinical application. However, heterogeneity in the tumor immune microenvironment leads to differences in the individual response rates. In this study, we identified novel immune-related molecular subclasses of breast cancer using a non-negative matrix factorization analysis. We enrolled 4184 patients with breast cancer, including 1104 patients from The Cancer Genome Atlas as a training cohort and 3080 patients from another four independent datasets as validation cohorts. In the training cohort, 36.9% of patients who exhibited significantly higher immunocyte infiltration and enrichment of immune response-associated signatures were categorized into an immune class, which was confirmed by probing the expression of immunocyte markers (CD3, CD19, and CD163). Within the immune class, 53.3% of patients belonged to an immune-suppressed subclass, characterized by the activation of stroma-related signatures and immune-suppressive cells. The remaining patients in the immune class were allocated to an immune-activated subclass. The interferon-γ and granzyme B levels were higher in the immune-activated subclass, whereas the transforming growth factor-β1 and programmed cell death-1 (PD-1) levels were higher in the immune-suppressed subclass. The established molecular classification system was recapitulated in validation cohorts. The immune-activated subclass was predicted to have a better response to anti-PD-1 immunotherapy. The immune-related subclasses were associated with differences in copy number alterations, tumor mutation burden, neoantigens, tumor-infiltrating lymphocyte enrichment, PD-1/programmed death-ligand 1 expression, mutation landscape, and various infiltration immunocytes. Overall, we established a novel immune-related molecular classification of breast cancer, which may be used to select candidate patients for immunotherapy.
Highlights
Breast cancer is the most common malignancy in women worldwide, with a continuous increase in its incidence [1]
A total of 4184 patients with breast cancer were enrolled from public databases, including 1104 patients from The Cancer Genome Atlas (TCGA; as the training cohort) and 3080 patients from four external cohorts
The maximum negative matrix factorization (NMF) decomposition weight among the remaining nine patterns was selected as the representative of these nine patterns, and the genes were sorted according to the difference between weight of pattern 8 and maximum weight of the other patterns, and the top 150 genes were selected as "exemplar genes,” which are presented in Supplementary Table 1
Summary
Breast cancer is the most common malignancy in women worldwide, with a continuous increase in its incidence [1]. Advancements in determining the molecular mechanisms of breast cancer have led to the identification of canonical markers for breast cancer subtypes, including immunohistochemical, proliferative, genomic, and immune markers [2]. Endocrine therapy, www.aging-us.com targeted therapy, chemotherapy, and immunotherapy have been preferably used for individual breast cancer subtypes [3, 4]. Neoadjuvant combination therapy, which adopts targeted agents and chemotherapy, is recommended for human epidermal growth factor receptor 2 (HER2)-positive and triplenegative breast cancers [5]. For advanced breast cancer with metastasis, the use of small molecule inhibitors and immunotherapy are fundamental strategies based on the tumor subtype and molecular characteristics [6]. Traditional classification systems and prognostic prediction markers do not accurately reflect the biological heterogeneity and clinical complexity of breast cancer. It is important to identify novel molecular subclasses contributing to tumor heterogeneity to guide optimal clinical management
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
