While most of the world is focused on how artificial intelligence is illustrating near-masterpieces and writing thrilling novels in mere minutes, one company is leveraging its power to de-risk drug development, a move that could substantially decrease time to market for novel treatments, and ultimately improve or even save lives.
VeriSIM Life, a San Francisco-based biotechnology firm focused on the application of machine learning techniques to the drug development and discovery process, is using AI to help drug developers more easily navigate the Food & Drug Administration’s approval process and avoid the pitfalls that keep life-saving novel drugs from making it to patients. At the top of the list is the historic bedrock of drug development: animal testing.
“What has historically been a stepwise process based on error-prone experimental models (animals), is now being parallelized by the use of AI systems that better represent both human and animal biology,” said Dr. Jo Varshney, CEO and Founder of VeriSIM Life. “This radically accelerates the process of getting drugs into clinical trials.”
The FDA drug approval process can take anywhere from 10 to 12 years or even longer, meaning that novel drugs entering the initial stages of development today likely won’t make it to market in time to help the patients who are reading about their development today. Importantly, VeriSIM is turning the tides for drug development by improving accuracy and dramatically reducing the industry fail rate.On average, just one in 10 drugs that enter preclinical trials ever become available to doctors and their patients.
According to Varshney, VeriSIM’s methods are nearly flipping that number by using millions of data points to predict the way that both patients and diseases will react to potential treatments before ever actually involving a person. That gives drug developers the opportunity to correct the course of development before investing limited research time and dollars into the clinical trials that could determine whether their drug makes it to market.
“AI is extremely useful at predicting the interactions between a drug and patients with a disease, though it’s not 100 percent accurate. At VeriSIM Life, we have literally run millions of validation scenarios against our AI to confirm its accuracy,” said Varshney. “And today, I’m extremely pleased to say that on average, we are at 86 percent accuracy. That is a huge advantage for drug companies, where the industry average failure rate is over 90 percent.”
That accuracy has additional benefits that could usher in what Varshney calls an era of AI-driven precision medicine, or medicine that is tailored to the unique biological needs of an individual patient. “AI can decrypt human biology so well today it is identifying our unique differences which can impact a drug’s effectiveness, and/or side effects,” she said. “With this insight, a drug can be modulated, or prevented from being prescribed to someone whose genetics or medical history may make them a poor candidate, enabling what some people call precision medicine.”
And VeriSIM isn’t alone. Researchers all over the world are making huge strides in the detection and diagnosis of diseases, and they’re doing it with extreme precision. A team of microbiologists at Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts recently developed an AI-powered microscope that can accurately diagnose blood infections. Also known as sepsis, roughly 1.7 million Americans suffer from a blood infection each year, and according to the Mayo Clinic, between 30 and 40 percent die of septic shock. Interestingly, this isn’t the first time that AI has been used to detect blood infections. In fact, AI was deployed in the medical field for this very purpose in 1970, and according to Varshney, it’s only going to keep getting better, and that means a bright future for patients with all types of conditions.
“I am very optimistic for the future of therapies for cancer, Alzheimer’s, heart disease, rare diseases, and more. Today, AI is helping unlock our understanding of disease at extremely granular levels of detail. And it will allow us to move away from our current one-drug-fits-all approach to healthcare," said Varshney. “Imagine a future where instead of multiple prescriptions, you receive a single cocktail designed just for you. In this paradigm, drugs and biology would work together instead of fighting each other. That’s the future I’m trying to create, along with so many other mission-driven innovators in the AI space.”