Toxicology Expert Sees AI, Brain Organoids Accelerating Drug Development
By Rebecca Stauffer
Artificial intelligence (AI) will serve as a co-pilot for toxicology testing but will not replace it. In addition, AI will reduce the need for animal models and help reduce development timeframes.
2023 PharmSci 360 opening plenary speaker Thomas Hartung, M.D., Ph.D., of Johns Hopkins University opened his presentation with these predictions. His talk offered a glimpse into a future of testing that looks much different than today.
Hartung is the Doerenkamp-Zbinden-Chair for Evidence-based Toxicology in the Department of Environmental Health and Engineering at the Johns Hopkins Bloomberg School of Public Health and Whiting School of Engineering, Baltimore.
Hartung offers this example: a toxicology study involving 506 chemicals found that AI findings were accurate 80% of the time when predicting human skin allergies. In contrast, animal tests were accurate just 74% of the time. The 80% figure is also expected to rise as AI improves.
He then provided a short primer on this disruptive technology.
“AI is, in essence, making big sense of big data, i.e., large, multimodal information,” Hartung said. The amount of data in the world in general is doubling every 18 months, and simultaneously, computers are following Moore’s law by doubling every two years. On top of this, AI has been doubling its capacity every three months since 2010. Together, there has been more than a billion-fold increase in capability since AI was coined in 1956.
As far as the promise of AI in the industry, he sees the following occurring in our lifetimes, possibly even in our careers:
- AI will replace animal models with more accurate results that translate better to human conditions
- AI will accelerate drug development due to improved assessment
- AI will democratize knowledge as AI becomes a tool for mining scientific literature
Animal testing is expensive, resource-heavy, and can take four to five years to identify possible carcinogens, while only 57% of the findings are reproducible even between rats and mice, Hartung explained, adding that AI’s impact on efficiency and effectiveness will make resources go much farther. It will also be tremendously disruptive to supply chains and companies that have made today’s animal testing system possible.
But scientists do not need to fear being replaced by AI. “We need humans in the loop,” he said. “AI will only replace those toxicologists, who don’t use it.”
Brains in a Petri Dish
Hartung ended his talk with a look at brain organoids. Much of his research at Johns Hopkins University has focused on this area, drawing coverage in international media outlets, including the Wall Street Journal, The Financial Times, Forbes and CNN. Connecting brain organoids to AI opens up opportunities for studying learning and memory. He characterized this work as “living science fiction” that will disrupt human cell and tissue research.
In his view, brain organoids offer additional opportunities for improved testing due to “organoid intelligence” as the human brain is constantly learning.