Preclinical drug testing is often a slow, costly process that relies heavily on animal models. Despite modern advances, 90 percent of compounds deemed promising in these models fail in clinical trials. The consequences for the pharmaceutical and biotechnology industries are staggering, with billions of dollars in sunk costs, years spent chasing nonviable candidates, and ethical concerns. Companies are now seeking alternative methods driven by a desire for more predictive, human-relevant data. One emerging solution, known as the Bio-AI clinical prediction platform, was developed by Quris-AI to address these challenges.
Quris-AI tests compounds on miniature, human-relevant organoids instead of relying solely on animals. CEO Isaac Bentwich, MD, explains, “At Quris-AI, we have developed a Bio-AI clinical prediction platform that integrates machine learning and organ-on-a-chip technology to deliver better safety and later efficacy predictions in the drug development life cycle.”
In recent news, pharma and life sciences giant Merck KGaA, in Darmstadt, Germany, began integrating the clinical prediction platform into its drug development program after a successful two-year validation study.
With Merck among the early adopters of Quris-AI’s Bio-AI platform, this new approach to preclinical testing has gained significant traction. Integrating advanced AI analytics with miniaturized organ-on-a-chip systems may potentially offer a more accurate and cost-effective route to identifying promising drug candidates—something Merck will no doubt monitor as it moves forward with Bio-AI.
Lab managers may want to keep an eye on these innovations, as they eventually could translate into:
- Fewer dead ends;
- Streamlined budgets; and
- Faster delivery of safe, effectual treatments to patients.
The limitations of traditional testing
“Out of all the drug programs having gone through in vitro and in vivo testing, 90 percent fail in clinical trials. That tells us animal testing is inaccurate in predicting human outcomes.”
Pharmaceutical research has long depended on in vivo animal studies to identify drug safety risks before advancing to human clinical trials. “Currently, drug development relies on animal testing, but it’s ineffective,” explains Bentwich. “Out of all the drug programs having gone through in vitro and in vivo testing, 90 percent fail in clinical trials. That tells us animal testing is inaccurate in predicting human outcomes.” For lab managers, these inefficiencies translate into lost time, depleted budgets, and higher odds of investing in drug candidates or therapeutic compounds that fail clinical trials.
Relying on animal testing raises additional concerns, including the well-being of test subjects and the biological differences between humans and other species that can lead to less reliable results. As the pressure mounts for more robust, human-relevant screening tools, the industry is searching for alternatives that reduce animal usage without compromising safety or efficacy data.
Combining patient-on-a-chip technology with AI
According to Bentwich, referring to this approach as a patient-on-a-chip is somewhat misleading. “When we speak about a patient-on-a-chip, it’s actually a misnomer,” he says, clarifying that a more accurate description would be “multiple intermittently interconnected, miniaturized human organs in a microfluidic chip.” Unlike single-cell lines, which Bentwich calls “convenient but very much not lifelike,” these three-dimensional organoids typically measure under a millimeter and can encompass multiple cell types. “A typical such organoid or spheroid may be less than a millimeter in size, about 1,000 cells,” he explains, “but it can include hepatocytes, endothelial cells, and immune cells, giving a more realistic snapshot of human physiology.”
By placing these mini-organs in microfluidic chips, Quris-AI can test how a compound is metabolized by a tiny liver, passes the blood-brain barrier, and interacts with brain cells. “This is a far cry from a true patient-on-a-chip,” Bentwich notes, “but it’s the best science has to offer right now, and it’s much more lifelike than any biology until now.” He adds that the company overcomes variability, which has long been considered a setback in organ-on-a-chip systems, by “combining it with AI […] to generate massive data that is highly predictive,” ultimately enabling a more robust and industry-ready alternative to traditional in vitro or animal-based testing. For lab managers, this synergy of organ-on-a-chip and AI means faster, more confident decision-making and fewer late-stage surprises.
Navigating regulatory changes with Bio-AI
As organ-on-a-chip and AI tools prove their value in preclinical testing, regulatory agencies are beginning to adapt. In recent years, legislative efforts such as the FDA Modernization Act 2.0, enacted in late 2022, and the proposed Modernization Act 3.0 have signaled a shift away from exclusive reliance on animal testing. They also encourage the adoption of AI, organ-on-a-chip, and stem cell technologies to improve predictive power.
“Both these laws press the FDA to gradually move away from animal testing and incorporate more accurate methods,” says Bentwich. “Animal testing is not disappearing overnight, but we should minimize it and use it more responsibly.”
By integrating these new methodologies, regulatory agencies aim to increase the predictive power of preclinical testing, reduce ethical concerns surrounding animal welfare, and expedite the drug development process. According to Bentwich, Quris-AI does not claim to eliminate animal studies but offers a way, as he puts it, “to know before animal testing which drug candidates are likely to be safe in humans and animals.”
“These changes,” Bentwich continues, “are opening the door, and [the FDA] is encouraging us all to move in this direction. Scientists have been saying it for over a decade now.”
From a lab management perspective, integrating Bio-AI solutions alongside traditional preclinical approaches can reduce wasted resources by flagging potential failures earlier. As regulators continue to acknowledge the limits of current models, platforms like Bio-AI represent a pivotal step toward safer, more efficient, and more ethically sound drug development pipelines.
Practical implementations for lab managers
Quris-AI’s platform-as-a-service model relieves labs of the need to purchase specialized equipment or hire dedicated robotics and AI teams. “Essentially, labs can send us their compounds, and we handle the testing in-house,” says Bentwich. “We provide real-time visibility, so they can oversee our process without building a huge organization.”
Although Quris-AI is still in the process of launching its platform and does not yet have finalized return-on-investment data, Bentwich predicts “dramatic results.” He notes that every day of delay in drug discovery can cost a large pharmaceutical company around $1 million, while smaller labs must still contend with the high price of animal studies and extended timelines. “If you can save even one out of five animal runs for a single compound,” he says, “that’s a significant savings in both money and time.”
Future outlook with patient-specific organ-on-a-chip
Looking ahead, induced pluripotent stem cell (iPSC) technology promises to replicate patient-specific biology on chips, driving a new era of personalized medicine. “Patient specificity, encompassing patient safety and diversity, is extremely important,” says Bentwich. “You can think of this technology unfolding in generations.”
Quris-AI’s recent acquisition of Nortis, renowned for its NASA-tested, kidney-on-a-chip technology, expands its organ portfolio. “By bringing Nortis on board,” Bentwich explains, “we’re streamlining the process so lab managers can efficiently oversee multiple organ-on-a-chip systems under a unified platform.”
