Biomarker development is considered as the most important area for improving drug development. Successful use of biomarker requires a close collaboration of early researchers and clinicians where early researchers need to look forward to understand how their research would impact clinical development, and clinicians need to look backward to conisder all the preclinical experiments performed into clinical trials. We will present a few recent oncology cases of utilizing biomarker in early phase adaptive clinical trials aiming for identifying patients who can benefit the greatest from the experimental drug which lead to better phase III study design and increased chance of success. Statistically, we will use appropriate procedure for the inference of treatment effect that would adjust for potential random high. We will share our excitement, challenges and issues with recommended solutions.