With a platform that can predict which drug candidates will work in humans—and eliminate the costly and inefficient practice of testing on animals—Quris hopes to revolutionize drug development.
Quris is among a stampede of startups coming to biopharma with artificial intelligence (AI) and machine learning (ML) tools that are changing the industry. The shift is happening quickly. By this time next year, the drug discovery landscape will look considerably different, according to industry watchers.
2023 is set (PDF) to be a year of “revolution and value” for AI and ML as they are finally “fit for purpose,” according to management consulting firm Syneos Health.
Backing up Syneos’ findings is a survey of 131 C-suite executives, conducted by the Deloitte Center for Health Solutions, which identified the development and acquisition of innovative products as their top objectives in 2023.
Precedence Research estimates that the AI market in healthcare will exceed $20.6 billion in 2023, which is nearly double the figure from 2021 ($11.1 billion).
According to Arda Ural, EY Americas Industry Markets Leader, Health Sciences and Wellness, next year will be an “inflection point” in the industry.
“What got you here won't get you there. That’s the 2023 mantra,” Ural said in an interview. “The practices we used to be successful in the pharma industry will have to be reconsidered. And we need to evolve to overcome that innovation deficit.”
Sharing this “inflection point” view is Harietta Eleftherochorinou, the vice president of the innovation hub at IQVIA. She points to stats that show emerging biopharmas account for 65% of the molecules in the global R&D pipeline and are increasingly holding onto these assets.
To hear more from Harietta Eleftherochorinou, listen to our podcast The Top Line
With market growth expected to slow over the next five years, Eleftherochorinou believes that employing AI will be the key X-factor for pharma companies seeking to stay ahead of the curve.
“Innovation should be embraced more broadly and reflected across applications on innovative mechanisms of action, innovative trial design, innovative drug delivery,” Eleftherochorinou wrote.
Forcing the revolution is the exorbitant cost of R&D given the typical time needed to develop a therapy (12 to 18 years) and the failure rate (90%). AI tools can significantly reduce both.
Enter companies like Quris, which has discovered hundreds of novel microRNA genes through analysis of the human genome. The Boston and Tel Aviv-based firm uses patients-on-a-chip technology to generate a proprietary data set that is automated, predictive and uses classification algorithms to identify which drug candidates will work safely in humans.
As for increased efficiency, Quris’ platform can reduce the failure rate to 50%, CEO Isaac Bentwich said in an interview with Fierce Pharma.
The timing is right as well, as the U.S. is on the verge of passing the FDA Modernization Act. The measure will halt animal study mandates, which have been in effect for more than eight decades. Animal testing will still be permitted to show a drug is nontoxic, but not required.
“It’s three revolutions that are culminating now, with Hollywood style timing,” Bentwich said. “It’s a perfect storm—organs on a chip coming of age, AI becoming powerful and focused on this problem and the regulator saying animal studies suck.”
Other companies, such as U.K.-based CN Bio, which just named Paul Brooks its new CEO, also are pursuing organ-on-a-chip solutions.
Another advantage of the technology that makes it attractive in facilitating drug development is its ability to “compress time,” Bentwich said, pointing to the investigation of treatments for nonalcoholic steatohepatitis (NASH).
“It takes 10 to 15 years for cirrhosis and fatty liver to develop this pathology within the human liver,” he said. “On the chip, you achieve that state within 15 days.”
Another company making noise in AI is New York City-based Envisagenics, which inked a research collaboration with Bristol Myers Squibb last month.
SpliceCore compares sequenced RNA data from patients to Envisagenics’ data base of 5 million-plus splicing events, using proprietary algorithms to identify the incorrectly spliced RNA and to develop therapeutics to fix them.
SpliceCore is more than a decade in the making, as Envisagenics was spun out of Cold Spring Harbor Laboratory. Over that time, CEO Maria Luisa Pineda has seen increased acceptance in the potential of the platform.
“We had to educate a lot of people because we were disrupting something that was always done manually,” Pineda said in an interview. “Imagine telling folks, hey, we understand you like doing chemistry one compound at a time. But now we have this machine—software that can test all of your compounds, thousands of millions of compounds at once in 2,000 patients.”
Milestones are falling fast. In 2020, a drug discovered in less than 12 months through AI entered a phase 1 trial for the first time. The obsessive-compulsive disorder treatment DSP-1181 was developed by Sumitomo of Japan using a platform created by Exscientia of the U.K.
Since that pivotal moment, more than a dozen products that were identified through AI have entered clinical trials.
“AI has been through numerous cycles of promise, hype, despair and rebirth,” Syneos’ Leigh Householder said in a podcast summarizing the consulting firm’s findings. “In 2023, programs that are well-designed, well-trained and right-sized are said to have a positive impact on healthcare. In other words, emerging out of that hype and disappointment cycle, into delivering real value.”