AI-Developed Lung-Disease Drug Enters Human Testing

AI-Developed Lung-Disease Drug Enters Human Testing

Healthcare delivery and drug discovery and development have used computers for decades.

From combinatorial chemistry to rational drug design and now artificial intelligence (AI), innovative technical breakthroughs have helped drive an ongoing flywheel of advances in healthcare and medicine.

Sector leaders aren’t sweating the regulatory uncertainty around AI as much as other industries might be. The medical community, after all, has always been cognizant of the high stakes implicit in practice and delivery.

“When it comes to healthcare, there are a lot of existing regulatory frameworks that have been put in place over a long period of time that are adequate to deal with these [AI] tools and the incorporation of these tools,” Dr. Scott Gottlieb, former commissioner of the U.S. Food and Drug Administration (FDA), said in a Wednesday (June 28) CNBC interview.

“Some need to be modernized, but you don’t need a new set of rules,” he added.

The “existing rules” point of view is similar to the approach taken by the federal regulatory bodies tasked with overseeing another technical innovation: cryptocurrency.

But applying AI capabilities to medicine is likely to have far greater societal utility than the digital asset sector has so far been able to generate.

For the first time, a pharmaceutical drug discovered and designed by an AI system is set to undergo human testing in the second phase of clinical trials.

See also: Healthcare Industry Could Be Generative AI’s Biggest Proving Ground

Pharma Companies Invest in Partnerships With AI Companies

Drug development and discovery is a notoriously difficult business but one that offers outsize rewards and life-changing patient impacts.

Still, roughly 9 out of every 10 potential drugs that reach clinical trials ultimately fail to make it into commercial production, according to the National Library of Medicine.

That’s why deploying AI models that can quickly and efficiently churn through massive datasets to come up with new molecules that can eventually be turned into a drug represents such an attractive innovation.

“With some of these large language models (LLMs), you can create multi-modal correlations [across imagery and text] that are just too hard to do across traditional discovery models,” Gottlieb said.

That’s how biotech firm Insilico Medicine discovered and developed its NS018_055 drug, which is meant to treat idiopathic pulmonary fibrosis, a chronic lung disease, according to a Tuesday (June 27) press release.

Scientists are now testing whether the drug works on humans.

“This first drug candidate that’s going to Phase 2 is a true highlight of our end-to-end approach to bridge biology and chemistry with deep learning,” Insilico CEO Alex Zhavoronko said in a statement. “This is a significant milestone not only for us but for everyone in the field of AI-accelerated drug discovery.”

Per the company’s statement, developing the NS018_055 drug would have cost more than $400 million and required up to six years of development.

By using AI, the firm was able to design and develop the chronic lung treatment for “one-tenth of the cost and one-third of the time,” successfully reaching the first phase of clinical trials two and a half years after starting the project.

Its success is one reason why pharma companies are trying to invest in partnerships with AI companies.

“These [AI] tools can be very helpful, as long as they are used appropriately and the models are reliable and well-validated,” Gottlieb explained during his CNBC interview. “AI offers unprecedented opportunity… We need proper oversight, but existing models can adapt to ensure safeguards.”

Insilico’s AI platform has separately discovered 12 pre-clinical drug candidates, three of which have advanced to early clinical trials so far, according to the Tuesday press release.

The sector is also littered with examples of AI drug development failure. London-based AI-biotech Benevolent AI laid off said last month it would lay off 180 employees after its lead drug candidate flopped.

Data-Driven Development

Other sectors could look to the healthcare industry’s example of ensuring both patient data privacy protection and leveraging the most relevant data sets when tapping AI solutions to improve the quality of the product being provided as they chart their own paths forward.

In healthcare, the importance of data quality and integrity is much higher than in consumer-facing AI applications.

“You have to find ways to ensure the quality of the data and the integrity of the data that is going in, and you also want transparency around it and consistent terminology across medical records,” Gottlieb told CNBC.

“Quality matters when lives are on the line,” Erik Duhaime, co-founder and CEO of AI-for-healthcare data provider Centaur Labs, told PYMNTS last month.