Noetik's core thesis is that the 95% failure rate in cancer trials isn't due to bad drug design, but an inability to identify the correct patient sub-population. Their models aim to solve this patient selection problem from the outset, rescuing potentially effective drugs.

Instead of the traditional lab-to-clinic pipeline, a "reverse translation" approach uses AI to analyze data from patients who fail standard-of-care treatments. This identifies the specific unmet need and biological target first, guiding subsequent lab research for higher success rates.

In preclinical drug development, choosing the right biological model is the most critical initial decision. Selecting an inappropriate model, such as the wrong PDX or organoid line, guarantees the research program will fail as it will be designed to answer the wrong question from the outset.

Only 5% of investigational cancer drugs reach the market due to the gap between lab models and human biology. Dr. Saav Solanki highlights organoids, which use real patient tissue, as a key translational model to improve the predictive accuracy of preclinical research and increase the low success rate.

An individual tumor can have hundreds of unique mutations, making it impossible to predict treatment response from a single genetic marker. This molecular chaos necessitates functional tests that measure a drug's actual effect on the patient's cells to determine the best therapy.

Drawing an analogy from neuroscience, Noetik argues for a top-down modeling approach. Instead of building a perfect simulation of a single cell and scaling up, they model the functional interactions at the tissue level first. This abstraction is more likely to predict patient-level outcomes, which is the ultimate goal.

To bridge the gap between animal models and human trials, Noetik trains models on its human data and then runs inference on mouse histology (H&E) images. This allows them to predict human-relevant biology and gene expression directly from the mouse model, overcoming a key translational hurdle in drug development.

The progress of AI in predicting cancer treatment is stalled not by algorithms, but by the data used to train them. Relying solely on static genetic data is insufficient. The critical missing piece is functional, contextual data showing how patient cells actually respond to drugs.

Unlike using genetically identical mice, Gordian tests therapies in large, genetically varied animals. This variation mimics human patient diversity, helping identify drugs that are effective across different biological profiles and addressing patient heterogeneity, a primary cause of clinical trial failure.