Skip to Content


Test Distinguishes Gastrointestinal Cancers

CancerWise - April 2007

A powerful but simple two-gene test distinguishes between a pair of nearly identical gastrointestinal cancers that require radically different courses of treatment, researchers say.

Significance of study

Researchers at M. D. Anderson and the Institute for Systems Biology in Seattle say that in addition to differentiating between the two cancers, the technique also will have wider application in other types of cancer as well, according to their February report in the Proceedings of the National Academy of Sciences.

The cancers that the researchers studied are:

  • Gastrointestinal stromal tumor (GIST), which is:
    • Treatable with the targeted medication Gleevec®
    • Unresponsive to chemotherapy
  • Leiomyosarcoma (LMS), which is:
    • Not treatable with Gleevec
    • Responsive to chemotherapy

Both cancers begin in the gastrointestinal tract.


An existing test distinguishes between the two cancers with about 87% accuracy, but intensive and time-consuming additional analyses are required for uncertain cases, says Wei Zhang, Ph.D., a professor in the Department of Pathology at
M. D. Anderson and senior author on the study.

Genomic approaches to diagnosing, selecting treatment and predicting patient response to therapy are beginning to work their way into the clinic, the researchers note. These approaches can rely on dozens of genes as biomarkers.

Research methods

First, the researchers measured genes present in 68 GIST or LMS tumors.

Then, they applied an analytical twist. Rather than identifying multiple genes that might distinguish each cancer type, they looked at every possible pair of genes.

Primary results

This resulted in a cancer classification based on two genes:

OBSCN – If this gene possesses more of its RNA than the gene C9orf65, then the tumor is GIST.

C9orf65 – If C9orf65 is more abundant in RNA, the tumor is LMS.

The test differentiated with near perfect accuracy between the two types of cancer, accurately identifying 67 of 68 tumors. One tumor, which had nearly a 50-50 split between the two genes, could not be diagnosed.

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% accuracy.

Secondary results

The analysis uses fewer genes to distinguish between:

  • Similar cancers
  • Patients who respond to treatment differently

In regard to treatment, the analysis may be used to determine who will benefit from different types of chemotherapy and who is at risk of relapse.

Zhang says the research group is using this analytical strategy to identify gene pairs that can predict which GIST patients respond to Gleevec and how other types of cancer respond to this treatment as well.

What’s next?

“This simple and accurate test has the potential to be quickly implemented into the clinic and used to guide appropriate treatment,” Zhang says.

– From staff reports


© 2015 The University of Texas MD Anderson Cancer Center