Companion diagnostic tests are essential for the safe and effective use of a therapeutic product. From a regulatory perspective, when a drug relies on a companion diagnostic test to classify patients for drug eligibility, regulatory acceptance of both the drug and the test are needed for joint licensure. Like other diagnostic tests, companion diagnostic tests are evaluated for analytical performance (i.e., quality of measurement) and clinical performance (association of test result with clinical condition of interest). Clinical performance evaluation can be complicated by less than ideal pivotal study designs. In this talk, I will discuss statistical methodology and regulatory developments in the analytical and clinical evaluation of companion diagnostic tests.