Subtype-Stratified Consensus Gene Signatures: Bridging Tumor Cell Biology, Immune Microenvironment, and Clinical...

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Breast cancer exhibits significant molecular heterogeneity, hindering universal prognostic tool effectiveness.
• Existing prognostic tools lack precision due to this heterogeneity.

Purpose of the Study:

• To develop robust, subtype-specific prognostic gene signatures for breast cancer.
• To improve overall survival prediction by integrating tumor cell and immune microenvironment data.

Main Methods:

• Systematic transcriptomic profiling across three independent breast cancer cohorts (TCGA, METABRIC, SCAN-B).
• Unsupervised clustering to identify consensus prognostic gene signatures.
• Single-sample Gene Set Enrichment Analysis (ssGSEA) for prognostic score derivation.
• Validation using C-index, time-dependent AUC, NRI, IDI, and independent cohort analysis.

Main Results:

• Identified pan-cancer and PAM50 subtype-specific prognostic gene signatures.
• ssGSEA-derived scores significantly stratified overall survival across cohorts, outperforming established features.
• Discovered subtype-specific mechanisms: immune pathways linked to good prognosis (HER2-enriched, Basal-like), oncogenic pathways to poor prognosis.
• Good prognostic scores correlated with anti-tumor immune cells and inversely with pro-tumor cells.
• Luminal signatures validated for stratifying distant relapse-free survival.

Conclusions:

• Subtype-specific consensus signatures offer superior prognostic accuracy in breast cancer.
• These signatures integrate tumor biology and immune microenvironment insights.
• Potential for clinical translation as improved prognostic tools for breast cancer patients.