Despite advancements, breast cancer outcomes remain stagnant, highlighting the need for precise biomarkers in precision medicine. Traditional TNM staging is insufficient for identifying patients who will respond well to treatment.
Our study involved over 6,900 breast cancer patients from 14 datasets, including in-house clinical data and single-cell data from 8 patients (37,451 cells). We integrated 10 machine learning algorithms in 55 combinations and analyzed 100 existing breast cancer signatures. IHC assays were conducted for validation, and potential immunotherapies and chemotherapies were explored.
We pinpointed six stable Panoptosis-related genes from multi-center cohorts, leading to a robust Panoptosis-model. This model outperformed existing clinical and molecular features in predicting recurrence and mortality risks, with high-risk patients showing worse outcomes. IHC validation from 30 patients confirmed our findings, indicating the model’s broader applicability. Additionally, the model suggested that low-risk patients benefit more from immunotherapy, while high-risk patients are sensitive to specific chemotherapies like BI-2536 and ispinesib.
The Panoptosis-model represents a major advancement in breast cancer prognosis and treatment personalization, offering significant insights for effectively managing a wide range of breast cancer patients.