How accurate are CertisAI platform’s predictions?

Results from internal hold-out cross-validation studies indicate CertisAI is 90%+ accurate for monotherapy predictions and 80%+ accurate for combination therapies. External validation studies specific to various indications and predictive biomarkers, in which we use clinically relevant PDX models and PDX-derived 3D cell cultures, are ongoing. Externally validated data is used as training datasets to continually improve predictive accuracy.

Here, Orthotopic PDX study validated CertisAI’s rank-order efficacy predictions for an individual with triple-negative breast cancer.