The identification of prognostic or predictive markers are one of the major tasks in clinical cancer research. In a recent, Phase II clinical trial for Pertuzumab™ (Omnitarg, Genentech, Inc. South San Francisco, CA), we used a hypothesis-driven candidate gene approach to develop clinically feasible diagnostic assays for markers known to be related to the target pathway of the drug. We conducted exploratory analyses of the data to identify cutoffs for the diagnostic markers by correlating them to progression free survival (PFS). We used the minimum p-value method to dichotomize the continuous marker values. However, appropriate corrections need to be applied to correct for bias in p-values and the log-hazard ratios from multiple testing. We discuss various methods and challenges arising in the context of co-development of drug-diagnostic in our clinical trial settings.