Many promising therapies fail because of poor patient selection; Harris warns that only about one third of tumors are truly hypoxic, underscoring the need for precise biomarkers in trials.