eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Scientists working on cancer treatment are looking into using artificial intelligence (AI) as a way of cutting drug development time. The research could save countless lives by helping predict and treat the deady disease.
At least two research and development teams — one at the London Institute of Cancer Research (ICR) and another at the University of Virginia (UVA) — are exploring independent AI solutions for medical drug discovery. While one is focusing on developing new drugs, the other is exploring ways to use old drugs for new diseases.
Pioneering AI “fingerprint” technology
A team of ICR scientists recently unveiled AI “fingerprint” technology that can successfully predict how cancer cells will respond to new drugs. The team first had to figure out how to train an AI model of this type — past AI tools were trained using 2D images, which doesn’t fully capture the shape of a cell. Their solution combines next-gen microscopy and geometric deep learning, which they used to analyze nearly 100,000 3D images depicting cells with melanoma skin cancer.
According to their tests, their new AI tech showed a 99.3% success rate when predicting the drug that was applied to cancer cells. It achieves this by looking at the cell’s 3D shape and monitoring the changes that occur after the drug has been administered. Not only does their research prove that AI can identify the biochemical changes associated with melanoma and cancer drugs, but their tool also identifies proteins that could be used in the development of new cancer drugs. Their AI tech could even be used to treat other diseases.
Introducing LogiRX
Meanwhile, a team at UVA’s School of Medicine developed LogiRX, a computational AI tool designed to reduce the time it takes to develop medical treatments for newly discovered diseases. Much like the tech pioneered by ICR, it does so by predicting the effects of various drugs on different biological processes.
Early tests have already detected a potential candidate for heart failure prevention in an antidepressant known as escitalopram. Led by UVA biomedical engineering professor Jeffrey Saucerman, the team analyzed 62 drugs in total. LogiRX predicted off-target effects for seven, two of which were ultimately confirmed.
“AI is accelerating many aspects of drug development, but it has made less progress in providing the required understanding of how these drugs work in the body,” Saucerman told Medical XPress. “LogiRx is a step towards combining AI with existing knowledge of how cells work to find new uses for old drugs.”
The two types of AI models being developed by ICR and UVA hold a lot of promise for the future of medical and healthcare AI systems around the globe.