AI-Driven “Patient Avatars” Aim to Replace Animal Testing and Predict Drug Toxicity

Study in a Sentence: Cedars-Sinai researchers are developing KronosRx, an artificial intelligence-powered platform that uses human-derived organoids and deep-learning models to forecast adverse drug reactions before clinical trials begin—potentially reducing and replacing animals used in testing and improving drug safety.

Healthy for Humans: A major challenge in drug development is predicting how a new medicine will affect real human patients. Currently, more than 30% of clinical trials fail due to harmful reactions that were not detected in preclinical testing, which often relies on animals and cannot predict how medicines behave in people over time. To address this, the Cedars-Sinai team is building “patient avatars”—sophisticated, experimentally engineered models of human organs, such as heart and brain tissues grown from human stem cells. These tiny organoids and organ-on-a-chip systems mimic how human tissues respond to drugs on a cellular level. What makes KronosRx distinctive is the AI component: Responses from these human tissue models are combined with millions of anonymized electronic health record data points to train deep-learning algorithms. This lets the system simulate how different patients—accounting for age, existing health conditions, and other medications—might respond over time.

Redefining Research: With KronosRx, Cedars-Sinai hopes to move beyond decades-old animal testing standards by creating a dynamic, human-centric predictive platform. The system is not just checking if a drug is toxic or safe—it’s modeling how and why adverse reactions might develop in diverse patients before they ever enter human trials. If successful, this approach could reduce clinical trial failures by flagging potential drug safety issues early on, accelerate development for safer treatments, and improve patient safety by using human biology and real-world clinical data rather than animals. By integrating stem-cell-based human models with AI trained on real patient histories, KronosRx represents a step toward more reliable, predictive drug research—with the promise of fewer animal experiments and more effective therapies reaching patients sooner.