Integrating genomic signatures for treatment selection with Bayesian predictive failure time models

Area of Science:

• Genomics
• Biostatistics
• Computational Biology

Background:

• Developing genomic signatures for targeted cancer therapies is resource-intensive.
• Single-gene studies and prognostic signatures limit personalized treatment selection.
• Current statistical methods for treatment selection are complex and require large patient cohorts.

Purpose of the Study:

• To describe a Bayesian predictive failure time model for treatment selection that integrates multiple-gene signatures.
• To enhance the matching of patients to personalized therapies based on genomic data.

Main Methods:

• A Bayesian predictive failure time model integrating multiple-gene signatures.
• Utilizing a heuristic similarity measure to assess historical patient data contribution.
• Leveraging existing clustering methods for the similarity measure.

Main Results:

• The proposed method demonstrates robustness with small sample sizes.
• Simulation studies and a lung squamous cell carcinoma case study validate the approach.
• The model effectively uses genomic signatures for personalized treatment selection.

Conclusions:

• The Bayesian predictive failure time model successfully leverages genomic signatures for improved cancer treatment selection.
• This approach enhances patient survival by matching them with the most beneficial therapies.
• The method offers a feasible solution for personalized medicine, especially with limited data.