Drug companies are struggling with a fundamental problem: promising treatments that work in the lab often fail when tested in real patients. Now, Stanford researchers and their peers are using a combination of patient-derived stem cells, artificial intelligence, and genetic analysis to predict which drug candidates will succeed before they ever reach human trials. This approach, called "clinical trial in a dish," could dramatically reduce the billions of dollars wasted on failed therapies each year.

Why Do So Many Promising Drugs Fail in Clinical Trials?

The disconnect between laboratory success and real-world failure stems from a fundamental mismatch. Traditional drug testing relies on animal models or generic cell cultures that cannot capture the full complexity of human disease. Animal models often cannot reproduce the genetic diversity or disease mechanisms found in actual patients, while conventional two-dimensional cell cultures lack the structural and functional complexity of human tissues. Human disease is influenced by genetics, environmental factors, and biological variability that these simplified systems simply cannot replicate.

This gap has created a persistent problem in drug development: therapies that perform well in preclinical models often fail to demonstrate efficacy or safety once they reach patients. The result is wasted time, money, and hope for patients waiting for new treatments.

How Are Scientists Building Better Preclinical Models?

Researchers are moving toward what experts call New Approach Methodologies (NAMs), which generate data directly from human-derived systems rather than relying solely on animal testing. The foundation of this shift involves human induced pluripotent stem cells (iPSCs), which can be generated from adult patient samples and reprogrammed into many different cell types while retaining the patient's genetic background. This means researchers can study disease using cells that reflect the biology of afflicted individual patients rather than relying on generic laboratory models.

These stem cell-derived systems can also be used to generate organoids, engineered tissues, and microphysiological systems that better reproduce key aspects of human organs. The additional complexity provided by these structures improves the evaluation of drug efficacy and toxicity during preclinical development, generating data that are more relevant to later clinical outcomes.

"One of the greatest strengths of iPSC-derived models is that they retain the genetic background of individual patients who are being treated. This enables researchers to investigate disease mechanisms directly in a patient-specific context and to study how genetic diversity influences therapeutic responses," explained Professor Joseph C. Wu, Director of the Stanford Cardiovascular Institute at Stanford University School of Medicine.

Professor Joseph C. Wu, Director of the Stanford Cardiovascular Institute, Stanford University School of Medicine

Steps to Integrate Multiple Technologies for Predictive Drug Discovery

• Combine Patient-Derived Models with Multi-Omics Data: Rather than using stem cell-derived models in isolation, researchers integrate them with single-cell sequencing, transcriptomics, proteomics, epigenomics, and advanced imaging to generate vast amounts of biologically relevant data from human tissues.
• Apply CRISPR-Based Functional Genomics: Researchers use CRISPR technology to identify genes that play causal roles in disease rather than simply being associated with it, moving beyond correlation to understand actual disease mechanisms.
• Leverage AI to Interpret Complex Datasets: Artificial intelligence is now essential to integrate and interpret these multi-omics datasets, helping researchers prioritize the most promising targets and experiments before drug candidates progress through the discovery pipeline.
• Validate Targets Through Large-Scale Perturbation Experiments: Before compounds progress to clinical testing, researchers conduct large-scale experiments to validate potential therapeutic targets, improving confidence in the candidates moving forward.

Professor Wu emphasized that the real power lies in integration rather than isolation. "The greatest value comes from integrating these technologies rather than using them in isolation. Human-derived experimental models generate biologically relevant data, whereas genomics and AI provide the analytical framework to interpret that information and make predictions," he stated.

Professor Wu

What Makes "Clinical Trial in a Dish" Different?

One longstanding limitation of drug development is that therapies are often evaluated using models that represent an "average" patient. In reality, genetic differences among patients can lead to marked variation in treatment response, making it difficult to predict which individuals will benefit and which may experience adverse effects. The "clinical trial in a dish" approach addresses this by assessing drug efficacy, toxicity, and biological responses across collections of patient-derived cells and tissues representing different genetic backgrounds.

This method enables researchers to identify potential responders and non-responders, investigate population-specific safety concerns, and better understand the biological factors influencing treatment outcomes before clinical trials begin. As the resulting experimental datasets grow, they could support predictive computational frameworks such as digital twins, which show promise of forecasting drug responses at both individual and population levels.

Rather than replacing clinical trials, these systems serve as powerful tools that strengthen confidence in therapeutic candidates before they enter the clinic. They support better target validation, earlier patient stratification, and more informed decision-making throughout drug discovery.

How Is AI Accelerating Every Stage of Drug Discovery?

Artificial intelligence has become one of the most widely discussed technologies in biomedical research, with one of its key strengths being the ability to help researchers interpret the exponentially expanding volume of biological data generated throughout drug discovery. Rather than replacing laboratory research, AI is used to analyze experimental data to help researchers prioritize the most promising targets and experiments, improving confidence before drug candidates progress to clinical testing.

"AI has the potential to accelerate nearly every stage of the drug discovery process, from target identification and drug design to efficacy prediction and safety assessment," said Professor Joseph C. Wu.

Professor Joseph C. Wu, Director of the Stanford Cardiovascular Institute, Stanford University School of Medicine

The integration of patient-derived experimental models with genomic and computational analyses represents a fundamental shift in how researchers approach drug development. By combining human-relevant preclinical models with AI-powered data interpretation, researchers are moving closer to a future where drug candidates are evaluated not just for their average effectiveness, but for their potential to help specific patient populations. This precision approach could reduce the attrition rate of promising therapies and ultimately get more effective treatments to patients faster.