Many of the genetic alterations identified using tumor-only sequencing are not actually associated with the cancer, but instead reflect inherited germline mutations already present in the normal cells of the person. These findings were published in Science Translational Medicine (2015;7:283ra53).
Scientists at Personal Genome Diagnostics Inc (PGDx), a provider of advanced cancer genome analysis and testing services, conducted the study working in collaboration with company co-founders Victor Velculescu, MD, PhD, and Luis Diaz, Jr, MD, and their colleagues at Johns Hopkins University in Baltimore, Maryland.
In the study, researchers from PGDx and the Johns Hopkins Kimmel Cancer Center compared DNA from tumors and from normal cells of 815 patients with a wide variety of cancers. They found that almost half of patients analyzed using tumor-only approaches had genetic alterations detected in their tumors that were also present in their normal cells, indicating that the alterations were ‘false positive’ changes not specific to the tumor.
The growing use of personalized medicine is predicated on tailoring treatments to the genetic makeup of the patient’s tumor, so the high rate of false positives uncovered in the study has implications for the accuracy of the approach when it relies on tumor-only sequencing.
“We knew from our pioneering whole exome analyses of cancer patients that a significant number of the genetic alterations that were thought to be associated with tumors were also present in the inherited germline DNA. By comparing tumor DNA to DNA from normal tissue, we were able to separate out those genetic alterations that are truly tumor-specific,” said coauthor Siân Jones, PhD, Vice President of Genome Sciences at PGDx.
“Accurately identifying tumor-specific alterations is essential to realizing the potential of personalized medicine to achieve better treatment outcomes,” added Jones. “As a result of this work, we decided to include the option to analyze both normal and tumor tissue DNA when we launched our CancerSelect targeted gene panel, which is designed to detect those genetic alterations in the [patient’s] cancer that are most relevant to optimizing treatment. The study published today with our colleagues at Johns Hopkins University is a powerful validation of that decision.”
The researchers identified 382 genetic alterations in the study patients that were potentially tumor-specific by first detecting all of the genetic changes in their tumors and then eliminating those that were well-known germline alterations. However, when the remaining genetic alterations were compared to the genomic profiles of the patients’ germline DNA, an average of 249, or 65%, turned out to be false-positive changes that were already present in the normal cells.
The researchers also looked at the alterations in actionable genes, which were defined as genes that have been identified as potential targets for cancer drugs or investigational cancer therapies. In the study, 48%, or almost half of the tumor samples, had at least one false-positive mutation in an actionable gene. These findings are noteworthy because use of false-positive findings to guide personalized treatment decisions could result in a substantial number of patients receiving therapies that are not optimized for their cancer.