When scientists at Bristol-Myers Squibb Research Institute in Princeton, NJ, recently studied whether a new drug could fight colorectal cancer, they started with an unusual assumption: It wouldn't help most patients. They bet that the drug, cetuximab, would benefit just a fraction of sufferers.

That's the new thinking in much cancer research: Drugs often work best when targeted to patients with specific molecular tumor profiles, says David J. Mauro, director of medical strategy at the institute. Those patients might be just a small percentage of those with a particular cancer.

Big pharmaceutical companies are starting to invest in this approach, though universities have done so to some extent for several years, says Mauro. "Pharmaceutical companies are becoming more convinced that personalized medicine, particularly in oncology, is not only an option but will be a requirement." It's part of what some scientists have hailed as an upcoming revolution in medicine: treatments tailored to an individual's genetic makeup, often called personalized medicine.

To accept that idea explains Mauro, businesses had to overcome an obvious worry: treating fewer patients would mean fewer profits. "It's a reflex reaction across many companies," he says, but they're realizing that usually what benefits patients, boosts the bottom line. A well-targeted medicine can see use in a niche market for decades, but one that helps only a few, yet is given to many - under the old, less-targeted approach - would come off looking only marginally effective on average and would fade from use, or never win regulatory approval. "All these companies have spent millions and millions of dollars and enrolled thousands and thousands of patients in failed clinical trials," Mauro notes.

As reported in the Aug. 1, 2007, issue of the Journal of Clinical Oncology, Bristol-Myers Squibb and academic researchers reported on 110 cancer patients who were taking cetuximab. The drug was developed by Branchburg, NJ-based Imclone Systems and is marketed by NY-based Bristol-Myers Squibb as Erbitux. They found a good response to the drug in about 25 to 30 patients whose tumors had high expression of two ligands, epiregulin and amphiregulin.

"They all had the same type of tumor - metastatic colorectal cancer - that on a slide looks the same," Mauro notes, adding that the findings made sense. Cetuximab, a monoclonal antibody, blocks the epidermal growth factor receptor (EGFR), part of a pathway linked to tumor growth. Some tumors seem more "addicted" to this pathway than others, Mauro says. The involvement of epiregulin and amphiregulin is logical because they activate the receptor, though their connection to better treatment response became evident only after the study, he adds.

Breast cancer is the only solid-tumor cancer currently treatable through a personalized-medicine approach, Mauro says, and the new work reflects an effort to extend this to other cancers. All the half-dozen oncology drugs in Bristol-Myers Squibb's pipeline have an associated "biomarker" research program in this vein, he adds, but the fraction such programs consume of the company's total R&D costs (reported as $3 billion in 2006) is unavailable.

The drawback of these trials is that they don't help most of the participants much. "They're doing it solely to advance science and to help future patients," says Mauro. Future research, he adds, will involve validating the cetuximab findings with prospective trials, and collaborating with San Francisco-based Genomic Health to develop an assay for the ligands useable in regular clinical practice.