The key to unlocking effective preclinical forecasting is individual patient data from clinical trials, both retrospectively and prospectively.
Manjari is addressing the more than 90% failure rate of new therapies that enter clinical trial evaluation. What if instead, every 1 in 2 drugs that entered clinical trials succeeded? Early drug development requires effective proxy signals, such as animal models and biomarkers, to evaluate whether a drug modifies the right biological mechanism, cures disease in the relevant tissue, or displays the right safety profile. Yet, the ability of these proxy signals to emulate unbiased human response to drugs is rarely even quantified, let alone systematically improved. ATHENA aims to increase clinical trial success rates by quantitatively evaluating these proxies, combining them into improved predictive models using causal AI.
