Active Learning-Based Prediction of Drug Combination Efficacy.

Journal: ACS nano
PMID:

Abstract

Combination therapy, which involves the use of multiple drugs, has emerged as a promising approach to cancer treatment. However, traditional combination therapy development is constrained by the vast experimental design space, requiring exhaustive testing of drug ratios, concentrations, and encapsulation strategies. In this study, we present a computational intelligence method combining active learning and fine-grid optimization to predict the efficacy of drug combinations, focusing on dual-drug-loaded polymeric nanoparticles for cancer therapy. Our approach harnesses Gaussian Process Regression to predict both drug efficacy and associated uncertainty, enabling rapid identification of optimal conditions with only 25% of the experimental effort. This method was successfully applied to optimize dual-drug systems, including doxorubicin and docetaxel, demonstrating significant reductions in experimental workload without compromising precision. Our study has demonstrated the potential of AI-driven methodologies in overcoming the challenges posed by traditional experimental designs in the drug delivery field.

Authors

  • Song Jin
    Key Laboratory for Cerebral Artery and Neural Degeneration of Tianjin, Department of Medical Imaging, Tianjin Huanhu Hospital, Tianjin 300350, China.
  • Xinyu Li
    School of Pharmacy, Binzhou Medical University, Yantai, China.
  • Guangze Yang
    School of Chemical Engineering, Faculty of Sciences, Engineering and Technology, The University of Adelaide, Adelaide, SA 5005, Australia.
  • Zhen Zhang
    School of Pharmacy, Jining Medical University, Rizhao, Shandong, China.
  • Javen Qinfeng Shi
  • Yun Liu
    Google Health, Palo Alto, CA USA.
  • Chun-Xia Zhao
    School of Chemical Engineering, Faculty of Sciences, Engineering and Technology, The University of Adelaide, Adelaide, SA 5005, Australia.