In the 1990s, the breast cancer drug trastuzumab (Herceptin; Genentech/Roche)--an antibody specific for human epidermal growth factor receptor 2 (HER2; also known as ERBB2)--was approved based on trials in which HER2 expression levels were used to select patients in clinical trials. This provided support for analogous efforts for drugs that target the epidermal growth factor receptor (EGFR). However, the development of these drugs, such as cetuximab (Erbitux; Bristol-Myers Squibb/Lilly) and gefitinib (Iressa; AstraZeneca), has revealed that EGFR expression is an insufficient and unreliable biomarker to select patients for EGFR-targeted therapies in both lung and colon cancer. Indeed, evidence on patient populations that are likely to respond to such therapies, on the basis of specific mutations in proteins of the targeted pathway, has only recently been clinically validated and incorporated into some of the drug labels. This article highlights lessons learned from the development of the first drugs targeting the EGFR family and discusses strategies to decrease the risk of failure in clinical development by more effectively integrating molecular diagnostics into anticancer drug discovery.