This study aims to identify precise biomarkers for breast cancer to improve patient outcomes, addressing the limitations of traditional staging in predicting treatment responses.
Our analysis encompassed data from over 7,000 breast cancer patients across 14 datasets, which included in-house clinical data and single-cell data from 8 patients (totaling 43,766 cells). We utilized an integrative approach, applying 10 machine learning algorithms in 54 unique combinations to analyze 100 existing breast cancer signatures. Immunohistochemistry assays were performed for empirical validation. The study also investigated potential immunotherapies and chemotherapies.
Our research identified five consistent glutamine metabolic reprogramming (GMR)-related genes from multi-center cohorts, forming the foundation of a novel GMR-model. This model demonstrated superior accuracy in predicting recurrence and mortality risks compared to existing clinical and molecular features. Patients classified as high-risk by the model exhibited poorer outcomes. IHC validation in 30 patients reinforced these findings, suggesting the model’s broad applicability. Intriguingly, the model indicates a differential therapeutic response: low-risk patients may benefit more from immunotherapy, whereas high-risk patients showed sensitivity to specific chemotherapies like BI-2536 and ispinesib.
The GMR-model marks a significant leap forward in breast cancer prognosis and the personalization of treatment strategies, offering vital insights for the effective management of diverse breast cancer patient populations.