A novel prostate cancer subtyping classifier based on luminal and basal phenotypes

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Prostate cancer (PCa) is a heterogeneous disease with varying clinical behaviors.
• A need exists for expression-based subtyping models reflecting prostate-specific biological processes.

Purpose of the Study:

• To develop a novel RNA signature for classifying prostate cancer subtypes.
• To identify distinct biological and clinical features associated with each subtype.

Main Methods:

• Unsupervised machine learning applied to gene expression profiles from over 100,000 primary prostate tumors.
• Development of a prostate subtyping classifier (PSC) based on basal/luminal cell expression and PCa-relevant gene signatures.

Main Results:

• Four subtypes identified: luminal differentiated (LD), luminal proliferating (LP), basal immune (BI), and basal neuroendocrine (BN).
• Subtypes exhibit distinct molecular pathways, clinical characteristics, and treatment responses.
• LD tumors show less aggressive behavior; BI tumors benefit from radiotherapy; LP tumors benefit from docetaxel.

Conclusions:

• A PSC based on gene expression profiles can effectively differentiate PCa into four biologically and clinically distinct subtypes.
• This classification aids in predicting tumor aggressiveness and guiding personalized treatment strategies for prostate cancer.