AI companies run virtual drug trials, aim to improve success of human studies
XLK•AI companies are working with drugmakers on virtual trials intended to improve the odds that experimental medicines succeed in human studies. The industry spends about $140 billion annually on human clinical testing, while around 12% of drug candidates are approved.
1. Simulating drug trials
AI firms say simulations can help drugmakers assess experimental medicines and acquisition targets before committing capital to costly human trials. Some simulations take a month or less, compared with years for traditional trials. BioinvestGPT said its predictions of six high-profile trial results were correct in five cases, including the failed late-stage trial of Novartis’ del-desiran.
2. Limits and adoption
AI predictions are not always right: BioinvestGPT forecast positive results for Novartis’ pelacarsen, which failed to reduce the risk of major heart attacks or strokes in a late-stage trial. The company later said it had misjudged the drug’s mechanism. Pharma companies are beginning to use AI trial modeling, while experts say human trials will remain necessary.
3. Predictions ahead
BioinvestGPT predicts failure for two Phase 3 trials of Biogen’s litifilimab in lupus and Phase 2 trials of Takeda’s zasocitinib in Crohn’s disease and ulcerative colitis. Biogen said it uses available tools, including AI, and noted that historical data could lead algorithms to predict negative results. Takeda research chief Andy Plump said he did not believe AI could yet make definitive predictions.
