Study highlights possible inaccuracies in thyroid cancer detection tests

Protein biomarkers are used to test for cancer before and after surgeries to remove tumors. To test thyroid malignancy, many biomarkers are tested separately to confirm cancer. However, new research from Michigan Technological University shows that the detection tests that measure two major biomarkers to diagnose thyroid cancer may be inaccurate. The study recently came out in the journal Thyroid, published by the American Thyroid Association.

One biomarker, thyroglobulin (Tg), is a glycoprotein that has sugar in its molecular structure. It's "candy-coated" says Tarun Dam, an assistant professor of chemistry at Michigan Tech who led the research. He points out that the second biomarker, galectin-3 (Gal-3) is not a glycoprotein, but "it has a sweet tooth."

And like a kindergartner on Halloween, Gal-3 can't resist Tg. The two proteins tend to clump, an interaction not accounted for in the current thyroid cancer detection assays.

The clumping cycle of Tg and Gal-3 is fairly straightforward from a biomechanics perspective. A thyroid cancer cell secretes Tg and Gal-3, and the sweet-tooth nabs the glycoprotein. As more Tg is secreted, the influx partially breaks up the larger clumps. The body can potentially remove some globs quickly, taking away some of the biomarkers before the samples are collected from the patients.

Currently, detection assays only look at Tg and Gal-3 separately. They have no way to account for the biomarkers that may be tied up or removed in clumps. Plus, there is no way to know what stage of the clumping cycle the proteins are in; they could be in a big clump, already metabolized, partially dissolved or mostly free.

Dam and his team tested the physical and biochemical properties of the biomarkers to ensure this clumping cycling was driven by the proteins' interactions.

Based on his team's findings, Dam suggests adding a step to break up the clumps before running the assay tests, which should help make the tests more accurate.