All Times EDT
Keywords: Artificial Intelligence, Pre-clinical Discovery, Decision Analysis, Real World Evidence
Artificial Intelligence (AI) has been widely used in the pharmaceutical industry. In this discussion, we will focus on AI applications on three different stages of drug development. 1) AI application in pre-clinical discovery: Identifications of biomarker targets are the essential process for the new drug discovery treating the disease. However, some traditional machine learning methods either ignore the biological basis in modeling or have issue for handling the high dimensional genomic dataset. In this roundtable, we will discuss how to tackle these problems, the benefit and risk of novel AI algorithms and how to collaborate with scientists to interpret the biological evidence from the model. 2) AI application in decision analysis: After the phase 2 read-out, sponsors are required to evaluate the probability of success for the future phase 3 trial and then make the go/no-go decision. So it takes up the question of how AI can help to predict the clinical outcomes for a longer period to mimic the ph3 scenario given the data from the short period phase 2 study and some historical trials. In this discussion, multiple deep learning algorithms will be covered and compared. 3) AI application in real world evidence: In clinical practice, healthcare providers usually need to generate the hypothesis for patients on their later performance from their first or first few visits and then a quick decision can be made on whether and when to switch the therapy. For payers, they also need the method to identify the potential drug responders among all patients and then help on their commercialization strategies. Therefore, we will discuss some traditional method vs. AI algorithms for predicting patients’ future performance given their baseline characteristics and the data from first few visits and how the AI can guide the real-world clinical practice.