A new method of identifying certain cancer-causing mutations in tumors could make it possible for physicians to routinely test for these changes. If the experimental method can be adapted for clinical use, it might be a cost-effective way to identify some of the underlying genetic changes in cancer.

The method is new, but the basic technology is not. Researchers at the Dana-Farber Cancer Institute and their colleagues adapted mass spectrometry genotyping, which has long been used to analyze normal variation in DNA, for the purpose of large-scale tumor profiling.

A pilot study to test the method on 1,000 tumors found it to be a reliable and relatively low-cost way to detect known mutations in cancer-causing genes, or oncogenes. The findings were reported online in Nature Genetics on February 11.

"Profiling oncogene mutations with mass spectrometry was not only faster and cheaper but also more sensitive, accurate, and specific than the traditional Sanger method," says former Dana-Farber researcher Dr. Roman Thomas, the first author of the study.

The Sanger method of sequencing DNA is used in laboratories around the world.

Featured Meetings and Events A calendar of scientific meetings and events sponsored by the National Institutes of Health (NIH) is available at http://calendar.nih.gov.

The new method could potentially help guide decisions about treatments in real-time, notes Dr. Thomas, who is now a principal investigator with the Max Planck Institute for Neurological Research in Cologne, Germany.

The study was conducted in part because the researchers anticipated a tremendous need to profile large numbers of clinical samples as current and future cancer genome projects neared completion. The profiling would have to be done rapidly and at reasonable cost.

An ideal method would capture information on oncogene mutations across the genome.

The catalog of genetic alterations in cancer keeps growing. Last year, researchers at Johns Hopkins published an analysis of 13,000 genes in colon and breast tumors. They identified 189 mutated genes, most of which had not been linked to cancer previously.

The Cancer Genome Atlas Pilot Project, a large-scale NCI effort, will soon begin to catalog genomic changes in lung, brain, and ovarian cancers.

"We need to know the essential alterations across cancer, but we also need ways to glean the relevant clinical information from every patient who walks through the door," says Dr. Levi Garraway of Dana-Farber, who led the new study.

His team tested the method by screening the 1,000 tumors for the presence of 238 known mutations in 17 oncogenes.

They focused on three types of mutations - those that are common (such as mutations in the RAS gene family), those that have clinical implications (such as mutations in the genes KIT and EGFR), and those that may interact with targeted therapies (such as those in the genes EGFR, KIT, or BRAF).

The tumors that were screened were "high-quality" samples. A potential problem in adapting the method for clinical use is that tumor specimens often degrade when they are preserved in paraffin.

Another limitation of the method is that it detects known mutations. "You have to know what you're looking for," says Dr. Garraway. But the researchers did see several instances of oncogene mutations in tumors where they had never been reported before.

"This told us that if such diagnostics did exist, you could get very useful information on every patient," says Dr. Garraway. Such information could potentially lead to a more focused and effective use of existing and emerging therapies, he adds.

The researchers also identified an unexpectedly high number of co-occurring mutations in some tumors.

An example of how tumor profiling might benefit patients was the discovery of two KIT gene mutations in a gastrointestinal tumor. The patient's tumor had relapsed after treatment with imatinib (Gleevec), and the profiling revealed that one of the mutations was D816H, which is associated with resistance to imatinib.

Physicians who know whether a tumor contains mutations associated with drug resistance may be able to select appropriate therapies for patients.

"This study focused on how best to apply the available information on the biology of the disease to the care of patients," says Dr. Frederic Kaye of NCI's Center for Cancer Research and a co-author of the study.

By Edward R. Winstead