Personalized medicine is gaining more attention in medical research and practice. A market-ready companion diagnostic assay (CDx) is used in personalized medicine for choosing the best treatment for an individual patient. In the ideal situation, the CDx is used for patient enrollment in device-drug pivotal clinical trial(s) so that Food and Drug Administration can ensure that appropriate clinical and analytical validation studies are planned and carried out for CDx. Unfortunately, development of the CDx may lag behind the development of the drug, and consequently, it is unavailable during the drug pivotal clinical trial. Instead, a clinical trial assay (CTA) may be used to enroll patients in the trial. Thus when CDx is available, to estimate the drug efficacy in the CDx intended use population, a bridging study will be required to assess the agreement between CDx and CTA in order to bridge the clinical data from CTA to CDx. The main challenge we face in a bridging study is covariate imbalance between treatment arms for the subpopulation with both positive CDx and CTA. In this paper, we introduce an estimation method for estimating the effect of a drug in a bridging study with missing data under a causal inference framework.