Artificial intelligence has been used by scientists to develop a new treatment regimen for children suffering from a fatal form of brain cancer. The survival rate for these children has not improved in over half a century.
Experts say the breakthrough, which was published in Cancer Discovery, will usher in a new era in AI that can be used to develop and invent new treatments for all types.
The promise of AI in drug discovery is transformative, according to Prof Kristian Helin (CEO of The Institute of Cancer Research, London), where scientists, doctors, and data analysts made the discovery.
This study has shown that AI can help to identify a combination of drugs that could be used as a treatment for children with incurable brain tumors. It is exciting to imagine that this could be the first AI-based treatment to benefit patients.
Computer scientists and specialists in cancer at the ICR, Royal Marsden NHS Foundation Trust used AI for the discovery that everolimus and vandetanib could be combined to treat diffuse intrinsic pontine growth (DIPG) a rare and rapidly-growing type of brain tumor in children.
DIPG and similar tumours in children are extremely difficult to remove. They are diffused, meaning they don't have specific borders that are suitable for operation.
After analyzing data from existing drugs, the team discovered that everolimus could increase vandetanib's ability to pass the blood-brain barrier to treat the cancer.
This combination was successful in mice, and it has been successfully tested in children. It is now being tested in clinical trials on larger numbers of children by experts.
Researchers found that mice with the combination of both drugs had a 14% increase in survival rate compared to those who received a standard treatment.
The drugs used in the research, which was funded primarily by Brain Research UK, Brain Collaborative, Children with Cancer UK, and the Royal Marsden Cancer Charity, were already approved for use in treating other types of cancer.
DIPG is an aggressive and rare form of childhood brain cancer. Survival rates are low so we need to find better treatments. Chris Jones, professor of paediatric brain tumor biology at the ICR, stated that DIPG is extremely rare.
Our study shows how AI can be used to discover new treatments for cancers such as DIPG.
To treat children with DIPG, the AI system recommended that two drugs be combined to treat the condition. One drug would target the ACVR1 mutation and the other would sneak past the blood brain barrier. We tested the treatment in mice to extend survival, and have begun testing it in small numbers of children.
Although we still require a clinical trial on a large scale to determine if the treatment is effective for children, we have made significant progress without AI.
Dr Fernando Carceller is a consultant in paediatric, adolescent and neuro-oncology at Royal Marsden NHS Foundation Trust and the leader of the ICR's paediatric, adolescent and neuro-oncology and drug-development team. He said that the breakthrough was encouraging.
BenevolentAI, a company that built an AI drug discovery platform, was the first to suggest the research. Researchers from the ICR collaborated with BenevolentAI in order to find drugs that could treat DIPG.
Professor Peter Richardson, vice president for pharmacology, BenevolentAI, stated that the early results were encouraging.
He said that AI-enhanced methods are already showing their value in expanding researchers' capabilities to discover new treatments, or repurposing old ones. This is not only for DIPG but also for other diseases.
