Determining breast cancer histological grade from RNA-sequencing data

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Histologic grade (HG) is a prognostic factor in breast cancer, with grade 3 being most aggressive.
• Grade 2 (intermediate) breast cancer poses challenges for clinical decision-making due to risks of under- and over-treatment.
• Accurate prognostic tools are crucial for tailoring breast cancer treatment strategies.

Purpose of the Study:

• To develop and validate a transcriptomic grade (TG) model for breast cancer using RNA-sequencing data.
• To reclassify intermediate-grade (HG2) breast tumors into high and low TG categories.
• To assess the potential of TG to improve risk stratification and clinical decision-making in breast cancer patients.

Main Methods:

• RNA-sequencing was performed on 275 invasive breast cancer samples.
• Multivariate prediction models were built using gene and isoform expression data to classify tumors into high and low TG.
• Recurrence-free survival analysis was conducted using the Cox proportional hazards model to evaluate the TG model's predictive power.
• The model was validated using The Cancer Genome Atlas (TCGA) dataset (N=487).

Main Results:

• The TG model achieved high accuracy in differentiating grade 1 and grade 3 tumors (AUC=0.97).
• A significant association was confirmed between HG and recurrence-free survival (HR=2.62 for HG3 vs HG1).
• Reclassification of HG2 tumors into high TG and low TG groups revealed differential recurrence risks (HR=2.43 for high TG HG2).
• 8200 genes and 13,809 isoforms were identified as differentially expressed between HG1 and HG3 tumors.

Conclusions:

• RNA-sequencing data can accurately differentiate between high and low histologic grades of breast cancer.
• The developed TG model effectively reclassifies intermediate-grade (HG2) breast tumors, identifying distinct risk groups.
• Implementing the TG model could potentially reduce overtreatment and undertreatment in breast cancer management.