Integrated Single-Cell RNA-Seq and Machine Learning to Construct an EMT Infiltration Scoring Model for Prostate...

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Prostate cancer (PCa) poses a significant global health challenge, with aggressive forms often developing despite current treatments.
• Epithelial-mesenchymal transition (EMT) is crucial for tumor invasion and metastasis but its heterogeneity in PCa is not fully understood.

Purpose of the Study:

• To characterize EMT dynamics in malignant prostate cancer cells at the single-cell level.
• To develop a prognostic model for predicting PCa recurrence and guiding therapeutic strategies.

Main Methods:

• Single-cell transcriptomic analysis to identify malignant cells and quantify EMT activity.
• Gene expression profiling, gene set enrichment analysis, and LASSO regression to identify key EMT genes.
• Construction and validation of an EMT infiltration scoring model using bulk transcriptomic data.

Main Results:

• Identified key EMT-associated genes and characterized EMT heterogeneity in PCa.
• Developed a robust EMT infiltration model that accurately predicts early biochemical recurrence in validation cohorts.
• Demonstrated associations between EMT scores and clinicopathological features, immune infiltration, and genomic instability.

Conclusions:

• The study systematically delineates EMT heterogeneity in PCa at the single-cell level.
• The developed EMT infiltration model offers a valuable tool for prognostic prediction and personalized treatment strategies in PCa.
• Findings provide insights for precision risk stratification and individualized therapy in prostate cancer.