PALM BEACH, Florida-Integrating molecular biology with patient care will require testing methods that are feasible in routine care, reproducible, and manufactured with good quality assurance, according to Al B. Benson III, MD. Early work in colorectal cancer suggests that such tests can be developed, that they will improve the efficiency of clinical trials, and that they will help clinicians select regimens most likely to benefit the individual patient, Dr. Benson said at the First International Colorectal Cancer Congress. He is director of clinical investigations, Northwestern University Robert H. Lurie Comprehensive Cancer Center, Chicago. "At present, eligibility criteria for clinical trials represent a generic population that includes a high percentage of patients who will receive no benefit from the study treatment," Dr. Benson said. "Targeting markers, such as those for epidermal growth factor receptor (EGFR) or vascular endothelial growth factor (VEGF), may help avoid giving patients a test drug from which they can expect only toxicity." For a marker to become clinically useful, it must have significant and independent value, be validated by clinical testing, be feasible for routine use, be reproducible, and be widely available with rigorous quality controls, Dr. Benson noted. "Furthermore, we should demonstrate that a given result will offer benefit to the individual patient," he said. The list of candidates for clinically useful markers in colorectal cancer is long (see Tables 1 and 2). Some of these will be markers for unfavorable prognosis, suggesting that the patient is a candidate for particularly aggressive treatment. Others are predictors of response to a specific drug such as fluorouracil (5-FU). "These include loss of the 18q allele in microsatellite stable cancer, mutations of the gene for type II receptor for TGF-β1 (TGF- β1-RII) in cancers with microsatellite instability (MSI), and high expres sion of thymidylate synthase," Dr. Benson said. One challenge for clinical investigators is how to help patients understand why taking tissue samples for marker analysis is important. "Many trials require additional patient approval for tissue acquisition, and recent privacy laws in the United States with their accompanying regulatory processes may make it much more difficult to obtain tissue and to establish tissue banks," he said. Tissue acquisition increases the cost of conducting clinical trials, and new funding sources will be required to establish tissue banks and develop usable tests for tumor markers. He said that there is a tension between industry's push for rapid drug development and researchers' need for biologically driven trials. Multidisciplinary collaboration required "This will require multidisciplinary collaboration. In the past, most trials did not include funding for the pathologists and surgeons, who will be critical for obtaining and evaluating tissue for our biological analyses," Dr. Benson said. Small uncontrolled or retrospective correlational trials may not provide definitive answers about the predictive power of a molecular marker, he said. "We must resist the temptation to use the growing number of commercial laboratories that market tests for molecular and other markers," Dr. Benson stressed. Trial Data Data from ECOG patients in intergroup trials of stage III colon cancer patients showed that retention of the 18q allele, TGF-β1-RII, and MSI with a TGF-β1-RII mutation were associated with favorable outcome after 5- FU. Five-year survival was 74% in patients who had the 18q allele vs 50% in those without the allele, and 74% in those who had MSI but also had the TGF-β mutation vs 46% in those with MSI but no such mutation. Dr. Benson said that clinical trial E5202, which is now on the drawing board for patients with stage II colon cancer, would be the first to prospectively evaluate presence of the 18q allele and MSI status in tissue samples obtained immediately after surgery. High-risk and low-risk profiles will be determined according to whether the patient has microsatellite stability but has lost the 18q allele (high risk) or has MSI without loss of the 18q allele (low risk). Low-risk patients will be assigned to observation, and Dr. Benson estimated that survival of low-risk patients should approach 90%. Highrisk patients will be randomized to 5-FU/ leucovorin or to an experimental arm that will include a targeted therapy yet to be determined.