AI-Selected Drug Candidate for Rare Brain Cancer Enters Clinical Trial
A drug that could be used to treat a rare form of brain cancer has moved into a phase I/II clinical trial. What makes this drug special is the artificial intelligence that guided its development.
Drug development is a tortuous process. In the US it can take 12 years on average for an experimental drug to pass through the various testing stages and pass regulatory approval. For patients whose prognoses can often be measured in months, this process isn’t just tortuous; it can be deadly.
Patients with glioblastoma multiforme (GBM), a rare form of brain cancer, are one such group. With a five year relative survival rate at only 5.1%, there is an enormous unmet need for effective treatments. Moreover, the majority of patients experience recurrence of the disease, even with existing treatment options.
So the news that a new candidate drug has passed into phase I/II clinical trial is good news. At this stage the trial is focused on evaluating how safe the drug – BPM 31510-IV – in patients experiencing recurrence of GBM.
What makes BPM 31510-IV special is how it was developed. The compound was guided through early development by artificial intelligence, specifically by the AI-based BERG Interrogative Biology® platform.
In a normal drug development model, explains BERG Chief Scientific Officer Ranga Sarangarajan PhD, the process is to look at efficacy signals such as influence on cell viability in cancer cells when they are exposed to candidate drugs, examining one pathway after another in a sequential manner to figure out actual mechanism that is responsible for efficacy signal.
However, using a combination of patient biology and artificial intelligence-based analytics, BERG’s platform can build a “comprehensive picture of the cancer biology”.
“Basically, once a comprehensive disease model is built, we can ask very specific questions,” says Ranga. “For example, what is the effect of hypoxia on the outcome? We can generate molecular networks in silico that provides insights into the disease biology that can be rapidly validated in appropriate models removing years of investigation typical of conventional hypothesis driven basic science initiatives.”
Not only does this approach have the potential to remove years from the process of drug development, but building models based entirely on human data has the potential to remove another element of uncertainty from the process: animal models.
All drugs will pass through a phase of animal testing, both to look at how a compound performs in a whole living system, and as part of legally required regulatory testing. While Ranga believes that this is unlikely to change in the immediate future, AI could very well be the technology that changes drug development into an entirely human in vivo process.
“All our drug discovery is done from models established using human biospecimens,” explains Ranga. “This offers the opportunity pick and choose for a large spectrum of genetically defined well characterized animal models that is consistent and ideal for pathway validation and its relevance to human disease. So ultimately this allows us to validate our drug candidates in appropriate animal models, but in a much more informed manner.”
These advances bring one crucial benefit to the process of drug development: reducing the number of failures. Drugs that fail during pre-clinical testing, or in clinical trials, are one of the critical reasons for the sheer length of time it can take to bring a new drug to market. As Ranga explains, applying artificial intelligence coupled with biological data to the process can quickly flag why a drug doesn’t work in a population, “allowing us to remove it from our testing cycle.”
The advance of BPM 31510-IV is an exciting new development for drug development. But even more than that it is a critical step forward in realising a future for precision medicine, one in which patients may only wait months, rather than years, for the drugs that will help them the most.