The gold standard for regulatory approval in oncology is overall survival (OS). Because OS data are initially limited, early drug development decisions are often based on early efficacy endpoints, such as objective response rate and progression-free survival. Tumor size (TS)-OS models provide a framework to support decision-making on potential late-stage success based on early readouts, through leveraging TS data with limited follow-up and treatment-agnostic TS-OS link functions, to predict longer-term OS. Conditional simulations (also known as Bayesian forecasting) with TS-OS models can be used to simulate long-term OS outcomes for an ongoing study, conditional on the available TS and OS data at interim data cuts of the same study. This tutorial provides a comprehensive overview of the steps involved in using such conditional simulations to support better informed drug development decisions in oncology. The tutorial covers the selection of the TS-OS framework model
applying the TS-OS model to the interim data
performing conditional simulations
generating relevant output
as well as correct interpretation and communication of the output for decision making.