We are currently experiencing a fundamental shift in the way in which we approach the characterization of cancer. Never before has the make up of cancer tissues and individual cells been so exhaustively researched and char- terized. We are now capable of producing molecular “fingerprints” that ch- acterize the expression of all known and unknown genes within tumors and their surrounding tissues. More than 30,000 different genes may be measured in each patient’s tumor in a single experiment. Simultaneously, novel therapies that exploit the molecular roadmap have been developed and are now being offered to patients. These novel agents, such as Glivec, Herceptin, Iressa, and others, specifically target individual genes within tumors and can produce d- matic responses in some patients. These drugs are only the forerunners of a coming tidal wave of novel therapeutics that individually target specific m- ecules within cancer cells—more than 300 such agents are currently in phase I or II clinical trials. This is an exciting time for cancer specialists and patients alike. However, if we have learned anything from the past 50 or more years of research into cancer, it is that Lord Beaverbrook, in founding the British national health service in the 1950s, was frighteningly prescient when he defined the primary goal of health care to be “Diagnosis, Diagnosis, Diag- sis. ” Now, more than ever, it is essential that appropriate diagnostic methods and approaches are applied to the selection of patients for treatment.