Unmasking Neuroendocrine Prostate Cancer with a Machine Learning-Driven 7-Gene Stemness Signature that Predicts Progression

Agustina Sabater ^1,2,3, Pablo Sanchis ^1,2,3, Rocio Seniuk ^1,2

⁰Departamento de Química Biológica, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Buenos Aires C1428EGA, Argentina.

Medrxiv : the Preprint Server for Health Sciences +

|

October 14, 2024

View abstract on PubMed

Summary

A new 7-gene signature predicts aggressive prostate cancer (PCa) and neuroendocrine prostate cancer (NEPC) progression. This tool aids personalized medicine by identifying high-risk patients for tailored treatment strategies.

Area Of Science

Oncology
Genomics
Biomarkers

Background

Prostate cancer (PCa) is a major global health concern.
Progression to aggressive neuroendocrine prostate cancer (NEPC) presents significant challenges.

Purpose Of The Study

Develop and validate a stemness-associated gene signature for PCa.
Identify patients with poor prognosis and NEPC subtypes.

Main Methods

Utilized Random Forest and Lasso regression on large-scale transcriptomic data.
Validated a 7-gene signature (KMT5C, MEN1, TYMS, IRF5, DNMT3B, CDC25B, DPP4) in independent cohorts and xenograft models.

Main Results

The 7-gene signature showed strong prognostic value for multiple survival endpoints.
Successfully identified NEPC subtypes and predicted poor outcomes in non-NEPC PCa with the signature.
Demonstrated dual prognostic and classifier capabilities.

Conclusions

The validated 7-gene signature is a robust tool for personalized PCa management.
Enables prediction of disease progression and guides treatment strategies.
Offers valuable insights for identifying high-risk PCa patients.

Keywords:

Gene signature Large Cell Neuroendocrine Carcinoma Machine Learning Neuroendocrine Transdifferentiation Prognosis Prostate Cancer Stemness