A powerful two-gene test shows near perfect accuracy in distinguishing gastrointestinal stromal tumor from leiomyosarcoma, according to an article published in the online early edition of the Proceedings of the National Academy of Sciences (USA).

"This simple and accurate test has the potential to be relatively quickly implemented in the clinic to benefit patients by guiding appropriate treatment," said senior author Wei Zhang, PhD, professor in the Department of Pathology at The University of Texas M. D. Anderson Cancer Center.

The analytical technique employed to distinguish gastrointestinal stromal tumor (GIST) from leiomyosarcoma (LMS) with near perfect accuracy will have wider application in more individualized diagnosis and treatment of other types of cancer, according to the authors.

Although gastrointestinal stromal tumor was once thought to be a type of leiomyosarcoma because both originate in the smooth muscle cells of the gastrointestinal tract, stromal tumors are treatable with imatinib but relatively unresponsive to chemotherapy. The opposite is true of leiomyosarcoma.

An existing test distinguishes among the two cancers with about 87 percent accuracy, but intensive and time-consuming additional analyses are required for uncertain cases, Zhang said.

The researchers used common whole genome microarrays to measure gene expression in 68 tumors, but then applied an analytical twist. Rather than identifying multiple genes that might distinguish each type of cancer, the researchers instead analyzed every possible pair of genes, said first author Nathan Price, PhD, research scientist at the Institute for Systems Biology, a process called Top Scoring Pair analysis.

The result was a cancer classifier based on the expression ratio of two genes. If the gene OBSCN expressed more RNA than the gene C9orf65, the tumor was gastrointestinal stromal tumor. If C9orf65 was more abundant, the diagnosis was leiomyosarcoma.

The test accurately identified 67 of 68 microarrayed tumors, with the exception being one tumor with nearly a 50-50 split between the two expressed genes. An additional test using a more accurate measurement procedure on the two genes identified 20 of the original samples (including the sample with near equal gene expression) and 19 independent samples with 100 percent accuracy.

Top scoring pair analysis allows the use of fewer genes to distinguish between similar cancers or between groups of patients who have one type of cancer yet respond differently based on genetic indicators, the authors said. For example, paired gene analysis may be used to determine which patients benefit from different types of chemotherapy and which patients are at risk of relapse.

Zhang said the research group is using this analytical strategy to identify gene pairs that can predict which stromal tumors patients will respond to imatinib and how other types of cancer respond to treatment as well.