Two of the University of Florida’s Environmental and Global Health faculty, Associate Professor Zhoumeng Lin, PhD, and Research Assistant Professor Wei-Chun Chou, PhD., are hard at work developing computer modeling technologies using machine learning and artificial intelligence (AI). They are using these approaches to address research questions about the safe and efficient delivery of nanomedicines to tumors. Much of their work is anchored in physiologically based pharmacokinetic (PBPK) modeling, a mathematical modeling technique that predicts how synthetic or natural chemicals move throughout the bodies of humans and other animal species. This information is critical to understanding chemical and drug safety and highly informs regulation and policy in the pharmaceutical drug development and toxicology sectors.
In a new research study entitled Predicting Nanoparticle Delivery to Tumors Using Machine Learning and Artificial Intelligence Approaches published in International Journal of Nanomedicine, Drs. Lin and Chou use PBPK modeling to better understand how nanoscale materials – very small materials that are typically < 100 nm and much smaller than your average bacterial cell – are targeted to tumors after they are injected. The article reports several machine learning and AI models that help in the design of nanomedicines with higher delivery efficiency to tumor sites. With information about the chemical and physical properties of the nanomaterials (e.g., size, shape, charge, core material) from their previously published Nano-Tumor Database, along with tumor characteristics, the team looked at four variations of tumor delivery efficiency after intravenous injection in mice in conjunction with nine machine learning and AI modeling algorithms to create 36 models. The models were trained to predict the optimal delivery of the nanomaterials to the tumors. “In addition, this overall framework provides a basis to support creation of a publicly available AI-based web interface that allows researchers to predict organ and tumor delivery efficiency of nanomaterials, thereby reducing or eliminating the need for animal testing and to accelerate the clinical and translational impact to the healthcare community,” said Dr. Lin. The team’s long-term research goal is to develop AI-based computer models to support decision making in nanomedicine, human health, animal health, and environmental health broadly.
Dr. Chou discussed how he hopes this project will impact cancer research. “The developed AI model provides accurate prediction and mechanistic understanding of bio-distribution for nanoparticle-carried cancer drugs. I hope this model can offer an approach to optimize the design of nanomedicine delivery and improve the success rate during drug development.”
Dr. Lin also shared that one of the most rewarding aspects of this research was the opportunity to collaborate with his students. “I enjoy working with my lab members as I like to share my knowledge, skills, and experiences with them, provide guidance on the research process, and to provide professional development support to help move them along on their career path,” said Lin. “In turn, I learn a lot from these interactions. I am always very happy whenever I see my lab members complete a project and publish it in a quality journal, receive an award, and find a good position.”
The project is a multi-year collaborative research study that is still ongoing. The group is currently collecting additional nanomaterial tumor delivery efficiency data in tumor-bearing animals and cancer patients. They are collaborating with Santosh Aryal, Ph.D., an associate professor in the department of pharmaceutical sciences and health outcomes at the University of Texas at Tyler. Dr. Aryal will conduct pharmacokinetic lab experiments using nanoparticles and data from these experiments will be used to validate and/or optimize the new AI-PBPK model developed by Dr. Lin’s lab. The additional data will help them optimize their machine learning and AI-PBPK models which will then be converted to an open source web-based interface that will be shared broadly by the scientific community.