High-throughput gene expression and protein technologies are enabling the discovery of cellular profiles that will aid in the prognosis of various malignant states and, most importantly, the prediction of which therapeutic intervention will most likely benefit a given patient. 1 The discovery, refinement, and use of these molecular diagnostic tools will ultimately revolutionize the practice of oncology. As these promising technologies are being developed, the refinement of more traditional diagnostics is progressing, particularly for patients with colorectal cancer. Although colorectal cancer is the second leading cause of mortality due to cancer in the United States, more than 60% of patients present with localized disease. For such individuals, adjuvant treatment is considered the standard of care. Yet 65% of patients are cured with surgery alone and only one third of the 35% of individuals who would have relapsed after surgical intervention derive benefit from currently available adjuvant therapy. Thus, only a minority of patients with stage II and III colorectal cancer either require or derive benefit from the use of adjuvant therapy. Stage has been found to be the single most reliable prognosticator of outcome. However, individuals within a given stage, particularly stages II and III, clearly may have widely divergent outcomes. Thus, there is an urgent need to discover, verify, and apply in clinical practice molecular diagnostics capable of more precisely prognosticating clinical outcome and predicting which patients are most likely to derive benefit from therapeutic interventions.

In this issue of the Journal of Clinical Oncology, Edler et al2 examined the prognostic and predictive value of thymidylate synthase (TS) protein expression in 862 patients with stage II and III colorectal cancer. Approximately half of these patients received adjuvant treatment with fluorouracil (5-FU) and leucovorin with or without levamisole. These