Drug trials should aim to identify subgroups who respond.

Three prominent cancer organizations are on a quest to redesign clinical trials, saying that additional testing before and during trials will highlight individual differences in drug response and detect successful targeted therapies faster. Their call comes as the US Food and Drug Administration (FDA) in March released its guidelines for personalized medicine.

Traditional clinical trials test for safety first and efficacy later, but that approach fails to take advantage of continuing advances in pharmacogenomics, experts say. “Clinical trials need to be designed so that you know if a drug is working within the first few patients,” says William Hait, director of The Cancer Institute of New Jersey.

Clinical trials need to be designed so that you know if a drug is working within the first few patients. , . William Hait, The Cancer Institute of New Jersey

According to a new model proposed by the American Society of Clinical Oncology, the American Association for Cancer Research and the Association of American Cancer Institutes, early clinical trials should include ongoing analysis of patients' tissue and blood samples. If a drug fails, scientists can determine whether it doesn't work because the target is inappropriate, or because genetic differences stop the drug from hitting the target in some individuals.

Such a design may have prevented the failure of the lung cancer drug Iressa: those trials were well underway by the time researchers identified a mutation in the EGF receptor that underlies the positive response in a small subset of patients (Nat. Med. 11, 107; 2005)

Before entering the clinic, scientists should understand the biology of the drug and its target, says Hait. “Then you can begin to predict how a mutation could alter that interaction,” he says. Some drug designers are trying to create 'irresistible inhibitors,' compounds that will bind to a target regardless of mutations.

The guidelines also recommend unbiased testing, such as blood-based proteomics assays, to see whether those who respond to the drug show a particular profile. Once researchers identify likely 'responders,' they can design a second clinical trial using only this population. Such a trial is currently underway for Iressa at the Massachusetts General Hospital in Boston.

The organizations plan to test the new design with a small-scale trial of EGF receptor inhibitors for non–small cell lung cancer. James Doroshow and colleagues at the US National Cancer Institute are also trying to designate money to include blood and tissue analysis in clinical trials. Their proposal is expected to be reviewed in June.

Experts hope drug companies will develop diagnostic tools along with new drugs, but at a workshop in April, Janet Woodcock, the FDA's acting deputy commissioner for operations, said that the business model and regulatory path for such markers is not clear. “And I'm not sure it's clear to the FDA either,” she said.

The FDA's new guidelines encourage—but don't require—companies to submit data on the impact of genetic variation on drug responses. In March, the FDA also issued dosing recommendations for Asians taking the cholesterol drug Crestor, after research showed this group metabolizes the drug differently. Although such differences are becoming increasingly apparent, the agency says it does not plan to ask for specific studies in ethnic subgroups unless there is a reason to suspect significant clinical differences.