Application of Kriging Models for a Drug Combination Experiment on Lung Cancer

TL;DR

This paper introduces Kriging models for efficient analysis of combinatorial drug experiments on lung cancer, reducing experimental runs while improving prediction accuracy.

Contribution

It applies Kriging models to drug response surface modeling, incorporating measurement error and optimizing experimental design for fewer runs.

Findings

Only 27 runs needed to predict 512 original experiments

Kriging models provide better precision than existing methods

Efficient experimental design reduces resource use

Abstract

Combinatorial drugs have been widely applied in disease treatment, especially chemotherapy for cancer, due to its improved efficacy and reduced toxicity compared with individual drugs. The study of combinatorial drugs requires efficient experimental designs and proper follow-up statistical modelling techniques. Linear and non-linear models are often used in the response surface modelling for such experiments. We propose the use of Kriging models to better depict the response surfaces of combinatorial drugs and take into account the measurement error. We further study how proper experimental designs can reduce the required number of runs. We illustrate our method via a combinatorial drug experiment on lung cancer. We demonstrate that only 27 runs are needed to predict all 512 runs in the original experiment and achieve better precision than existing analysis.