Establishing and using threshold of surrogate endpoint in relation to clinical endpoint^#

Clinical trials are an essential aspect of the drug development process. Clinical endpoints and surrogate endpoints are two terms used in clinical trials to measure the effectiveness of a treatment. While clinical endpoints typically require higher costs and longer durations of observation to show direct clinical benefits, surrogate endpoints have been introduced as a cheaper and faster method that may be used to predict clinical effects. When there is a linear relationship between the surrogate and the clinical endpoint, the surrogate may still need to rule out a threshold that corresponds to no clinical benefit. The determination of such a threshold uses the knowledge of numerous parameters in the bivariate statistical distribution of the clinical response and the surrogate. In our work, we present a concept of “working” threshold to incorporate statistical uncertainties in determination of such a threshold.