How AI-Driven Integration of Data Improves Preclinical Studies, Reduces Drug Trial Failures: Jo Varshney, PhD, DVM

Despite decades of advances in drug discovery, nearly 90% of drug candidates still fail in clinical trials, often due to shortcomings in preclinical research, says Jo Varshney, PhD, DVM, CEO and founder of VeriSIM Life. By harnessing artificial intelligence (AI) to combine human and animal data, VeriSIM Life is pioneering a new methodology that strengthens the reliability of preclinical studies, optimizes animal model choice, lowers development costs, and accelerates the path to regulatory approval—all while reducing unnecessary animal testing.

Furthermore, these findings highlight how AI can be used to improve data integration, speed up regulatory timelines, and reduce trial error.

This transcript was lightly edited; captions were auto-generated.

Transcript

How does VeriSIM Life improve the reliability of preclinical studies and reduce trial failures?

VeriSIM was founded on only 1 core and only 1 focused mission, which is reducing the translational gap. What that really means is, most of the drugs that enter into clinical trials fail; about 90% of them fail, and it's due to the things that happen in a nonclinical setting, either in a lab setting or in an animal study, that do not translate into human trials, and there are several reasons for that. What our approach is, is that it combines the mechanistic understanding of different diseases, artificial intelligence, and multispecies data, which is from rats to primates to humans and several other species, to really learn what are the differences and what are the similarities between these trials before the drug candidate enters into clinical trial. And that helps, not only for the researchers to understand how to design an animal trial better, but also how to find the best animal model.

For instance, if there is a program where a rat is a better model to reflect what is happening in the human disease, then just utilize that instead of using a mouse model or a guinea pig or rabbit, which are not really going to be useful to capture the complexity of the disease in the human trials. Additionally, it's pretty timely, given there's a significant effort by the FDA and NIH to really hone in on the nonanimal and more human-centric methods, and we've been doing this for about 8 years now. These simulations that we've created are to really not only capture those differences between animals and humans but also make the whole drug development process efficient and reduce the trial-and-error method.

What happens when we are reducing the animals, because in a humane way, this is an amazing use of technology to be able to not have to do unnecessary animal testing. But if you're looking at it from the business economics and also for patients, it's a huge win. Why? Because now we are derisking the challenges one may see in human trials very early on. You're now reducing the costs that it takes to really identify those challenges in human trials, because often human trials are from tens to billions of dollars within the range of the clinical trials a company pursues.

Then the other thing is, you are reducing the time to get to the regulatory acceptance and also getting into clinical trials, which means now the patients who are waiting for new therapies have a higher chance of receiving novel therapies or participating in research trials much more quickly than ever before, and that's what we are doing. We are very excited and very proud to be one of those companies doing that.