Construction and Validation of a Robust Cancer Stem Cell-Associated Gene Set-Based Signature to Predict Early...

Area of Science:

• Oncology
• Genomics
• Prostate Cancer Research

Background:

• Early biochemical recurrence (BCR) is a key indicator of prostate cancer (PCa) progression and metastasis.
• Identifying patients at high risk for early BCR is crucial for effective treatment strategies.

Purpose of the Study:

• To develop a predictive signature for early BCR in PCa patients.
• To identify a subgroup of PCa patients at high risk of early BCR using a cancer stem cell (CSC)-associated gene set.

Main Methods:

• Utilized The Cancer Genome Atlas (TCGA) PCa dataset, split into discovery and validation sets.
• Employed propensity score matching and differential gene expression analysis to identify candidate CSC-associated genes.
• Applied the LASSO Cox regression model to select prognostic CSC-associated genes for early BCR prediction.

Main Results:

• A thirteen-CSC-associated gene signature was developed to predict early BCR.
• High-risk patients identified by the signature showed significantly poorer BCR-free survival in both discovery and validation sets.
• The CSC gene signature demonstrated superior predictive value over clinicopathological features, with an AUC of 0.800 at 1 year.

Conclusions:

• A novel CSC-associated gene set-based signature accurately predicts early BCR in PCa.
• This signature offers significant clinical utility for identifying high-risk PCa patients.