Cancer research is increasingly turning to genetics to expose the inner workings of tumors and to guide treatment. But tumor-only analyses offer up many “false positive” mutations that appear to contribute to cancer, but which actually show up elsewhere in an individual’s healthy tissue, a new study finds. Sampling both tumor and healthy tissues might provide a way to sort out truly cancerous mutations, the scientists report.
A team of researchers in Baltimore tested tumor tissue and healthy tissue from 815 patients who had various cancers. Using only the tumor analysis, the tests spotted an average of 382 mutations per case that appeared associated with cancer. But nearly two-thirds of these variations, on average, also showed up in healthy tissues, suggesting that they weren’t driving the cancer, the authors report in the April 15 Science Translational Medicine.
The findings offer a new spin on personalized medicine, which targets mutations in tumors to refine therapy and limit side effects. The mutations that were simultaneously found in tumor and healthy tissues turned out to be germline, or hereditary, variations that are present at conception. Study coauthor Victor Velculescu, an oncologist at Johns Hopkins University School of Medicine, says that the vast majority of such inherited gene variations are simply rare variants that make people different from one another. Very few predispose people to cancer. He calls them “passenger mutations in the germ line,” contrasting them with the “driver mutations” that propel the cancer process.
These inherited variants are common and typically have minor effects. But they might occur in particular gene regions and look similar to variations that are related to cancer. “They are masking themselves as cancer mutations,” Velculescu says. “We get tricked by these alterations, which are really in all the cells in our body.”
When the researchers concentrated strictly on what appeared to be well-known cancer-related mutations in tumors, they found that patients averaged two or three such mutations. But in 48 percent of these patients, at least one of these mutations was a false positive —it also appeared in healthy tissue and wasn’t unique to the tumor.
Analyzing healthy tissue also revealed some hidden pitfalls not apparent in the tumor-only analysis. In 3 percent of the group, healthy-tissue analysis found a potentially detrimental germline mutation predisposing them to cancer. All but one of these 27 patients were unaware that they carried such a risk. They had no family history of it, Velculescu says.
“In principle, I completely agree with the notion of studying paired tumor and normal specimens,” says Elaine Mardis, a biochemist and geneticist at Washington University in St. Louis. Tumor-only analysis might have value in very focused testing of known hot spot mutations, she says. But more broadly applied, it risks providing misleading information.
Analyses that include normal tissues would cost more than current testing but might improve cancer treatment, Velculescu says. “You cannot have precision medicine without precision diagnostics,” he says. “Inaccurate genetic information from tumor-only tests can have substantial consequences,” such as side effects from inappropriate treatments. Velculescu and study coauthor Luis Diaz Jr. are cofounders of a gene analysis and testing service that has a patent application pending on some of this work.