ERBB2 as a Prognostic Biomarker in Prostate Cancer: Integration of Single-Cell Transcriptomics, Deep Learning, and...

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Biochemical recurrence (BCR) significantly impacts prostate cancer (PCa) prognosis.
• T cell exhaustion and metastatic PCa (mPC)-related genes are crucial in tumor progression.

Purpose of the Study:

• Identify key genes associated with BCR in PCa.
• Validate the clinical significance of identified genes.
• Develop a deep learning model for predicting BCR risk.

Main Methods:

• Integrated single-cell RNA sequencing (scRNA-seq) and deep learning.
• Intersected highly expressed genes in CD8+ T cell exhaustion clusters with mPC-related genes.
• Utilized Cox regression and a neural network model for prediction and validation via immunohistochemistry (IHC).

Main Results:

• ERBB2 was identified as the most predictive gene for BCR.
• High ERBB2 expression correlated with shorter biochemical recurrence-free survival (bRFS).
• ERBB2 expression was associated with higher PSA, NM stage, and ISUP grade.

Conclusions:

• This study reveals ERBB2's critical role in PCa BCR using integrated omics and deep learning.
• High ERBB2 expression serves as a potential biomarker for poor prognosis in PCa.
• ERBB2 may represent a novel therapeutic target for personalized PCa treatment.