|Risk index: Fifty genes that matter in lung cancer|
By Kate Dalke
August 2, 2002
ow a patient will do in the years following surgery for lung cancer can be predicted by studying fifty human genes, a new study reports. The genes constitute what the researchers call a 'risk index': These genes behave differently in different types of tumor cells, and their activity correlates with the risk for relapse.
Many patients in the early stages of non-small cell lung cancer respond well to surgery, but the disease recurs in a third of these patients. If doctors know which patients are likely to relapse, they could consider using more aggressive or alternative therapies.
The research, led by David G. Beer at the University of Michigan in Ann Arbor, is one of many projects currently testing the use of DNA microarrays, or gene chips, to guide decisions about treating cancer and other diseases. Doctors hope to design the most appropriate treatment for patients by recognizing patterns of gene expression in cells related to the progression of disease.
"The risk index predicts outcome and survival, which suggests there is a profile of gene expression that correlates with high-risk patients," says Beer. He adds that the index may not necessarily include the top 50 genes for predicting survival, but larger studies will help find such genes.
In this study, which included 86 patients, the researchers monitored the activity of 7,000 genes in tumor cells using microarrays. They analyzed a subset of nearly 5,000 genes expressed specifically in the lung. These expression data were used to create an index of 50 genes that can identify high- and low-risk lung cancer patients in the early stages of disease, the scientists report in Nature Medicine.
Beer's team confirmed published data on genes previously linked to non-small cell lung cancer, but many genes in the index had never before been associated with patient outcome.
Similar microarray studies on non-small cell lung cancer are generating different lists of genes relevant to patient survival, says Beer. Differences may be due to the different designs of the studies, the tumors chosen, or the methods of analysis.
"There needs to be some consensus about how the studies are performed so the results can be related," Beer adds. "In addition, larger studies will help in designing better arrays that contain genes most relevant to cancer."
See related GNN article
. . .