Design and analysis of drug combination experiments

Area of Science:

• Pharmacology and Toxicology
• Biostatistics
• Computational Biology

Background:

• Assessing drug synergy is crucial for developing effective combination therapies.
• Existing methods for synergy assessment can be complex and prone to bias.
• Plate-location effects can introduce variability in high-throughput screening data.

Purpose of the Study:

• To present a novel, simple, and easily implementable parametric modeling approach for synergy assessment.
• To introduce an extended three-parameter log-logistic model for analyzing interaction data.
• To correct for bias introduced by plate-location effects in synergy studies.

Main Methods:

• Utilized an extended three-parameter log-logistic model for data analysis.
• Employed PROC NLMIXED in SAS for statistical analysis.
• Provided SAS code for the implementation of the proposed method.
• Illustrated the approach with an oncology study involving a fixed-ratio design of two compounds in 96-well plates.

Main Results:

• The proposed model effectively analyzes synergy and calculates confidence intervals for interaction indices.
• The method successfully corrects for bias arising from plate-location effects.
• Demonstrated the practical application and ease of implementation through an oncology study.

Conclusions:

• The novel parametric modeling approach offers a simple and robust method for synergy assessment.
• The log-logistic model with plate-location correction enhances the reliability of interaction index calculations.
• This method provides a valuable tool for researchers studying drug combinations, particularly in oncology.