Clinicogenomic Insights for Progression-Free Survival in Prostate Cancer

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Prostate cancer (PrCa) is a leading global cancer in men, necessitating advanced prognostic tools.
• Current PrCa assessment relies on limited clinical parameters, highlighting the need for integrated clinicogenomic approaches.
• Early detection and accurate risk stratification are crucial for effective PrCa treatment and improved patient outcomes.

Purpose of the Study:

• To evaluate the prognostic value of clinicogenomic profiles in predicting progression-free survival (PFS) in prostate cancer patients.
• To compare the performance of three distinct survival models in assessing PrCa progression risk.
• To identify key clinical and genomic predictors of PrCa progression.

Main Methods:

• Utilized a cohort of 494 PrCa patients from The Cancer Genome Atlas (TCGA) via cBioPortal.
• Employed penalized Cox model, random survival forest, and deep learning survival neural network for survival analysis.
• Integrated clinical features and single-nucleotide variant data for prognostic modeling.

Main Results:

• Survival models achieved strong discriminatory performance (Harrell's C-index 0.80-0.87) on test data.
• Consistently identified neoadjuvant treatment history, cancer status, tumor recurrence, and the gene MYH6 as significant predictors of PrCa PFS.
• All models demonstrated robust ability to rank patients by progression risk.

Conclusions:

• Clinicogenomic data integration enhances the prediction of prostate cancer progression-free survival.
• The study underscores the importance of incorporating genomic information alongside clinical data for personalized oncology.
• Findings support the development of advanced models for improved PrCa risk stratification and therapeutic decision-making.