NEW YORK (GenomeWeb) — Genetic testing company InVitae has evaluated the clinical performance of a next-generation sequencing-based test for determining hereditary cancer risk and found that it is as good as Myriad Genetics' BRACAnalysis test at identifying variants in the BRCA genes and can also find additional variants that increase the risk of hereditary cancer.
InVitae collaborated with researchers at Stanford University and the results were published this month in the Journal of Clinical Oncology.
The researchers invited women that had been referred to Stanford for clinical BRCA1/2 testing between 2002 and 2012 to participate in the study, which assessed their germline DNA for variants in 42 genes with cancer risk associations — including the BRCA1/2 genes — and returned potentially actionable results.
In total, 198 women consented to participate in the study. Of those, 174 had breast cancer and 57 had BRCA1/2 mutations.
Steve Lincoln, senior vice president at InVitae, told Clinical Sequencing News that the study had three main goals. First, was to see whether a next-generation sequencing-based gene panel could find additional risk variants other than in BRCA1/2 and whether those risk variants could help the doctor make treatment decisions or advise the patient on preventative measures. A second goal was to see how the NGS panel compared to Myriad Genetics' commercial Sanger-based test for identifying variants in the BRCA1/2 genes. And finally, the third goal was to assess the number and potential impacts of uncertain findings from the multi-gene panel.
Currently, InVitae offers a hereditary cancer diagnostic panel that consists of 29 genes for $1,500. The test actually sequences around 200 genes, but the firm only analyzes and interprets results from 29 genes, Lincoln said. Additionally, the test can be customized depending on the physician's preference. For instance, a doctor can just order the genes involved in Lynch syndrome, Lincoln said. No matter what gene combination is ordered, the price remains the same, and the physician can then request interpretation of additional genes for no extra cost.
The tests are run in InVitae's CLIA-certified and CAP-accredited laboratory on Illumina's HiSeq system. In the JCO study, however, the firm ran a 42-gene panel on the Illumina MiSeq to an average of 400x coverage.
Of the 57 women that had tested positive for BRCA1/2 mutations via Myriad's clinical test, the InVitae assay detected 58 of the 59 variants. The variant the assay missed was a large insertion, which would have likely been picked up with deletion/duplication testing, which InVitae performs on its clinical samples.
Interestingly, one of the 59 variants detected by Myriad's assay and called as pathogenic was classified as benign according to InVitae's assay. InVitae called variants as pathogenic based on recommendations set by the American College of Medical Genetics.
Lincoln said that InVitae is currently in the midst of a study of over 1,000 participants comparing the ability of the firm's NGS panel to analyze BRCA1/2 variants to Myriad's test. Looking at the first 600 individuals, he said that the "results so far are substantially concordant" and he expects the full study to be completed within in the next few months.
Looking at genes other than BRCA1/2, the researchers found 16 pathogenic variants in the 141 women negative for BRCA1/2 mutations. Eleven had previously been reported and five were novel. Of those 16 variants, 15 were determined to be potentially actionable and the researchers then attempted to contact the 14 women carrying them. The genetic counselor was able to reach 11 of the women by phone. Ten accepted follow-up counseling appointments and one woman had passed away but her children accepted the appointment.
The participants each met with a genetic counselor and oncologist, who explained their sequencing results and gave recommendations based on those results. Six participants carried mutations to genes that conferred a doubling of their breast cancer risk and were advised to receive annual breast MRIs. An additional six participants harbored mutations that put them at an increased risk for gastrointestinal cancer and were advised to consider a colonoscopy and/or endoscopic gastroduodenoscopy every one to two years. And one woman was found to have a frameshift mutation in a gene know to cause Lynch syndrome, an inherited condition that increases the risk of getting colon and other types of cancer. This participant, who had been diagnosed with triple-negative breast cancer at age 35, had already undergone a hysterectomy for endometrial cancer at age 46. Because of the Lynch syndrome mutation, she underwent additional preventative measures — having her ovaries and fallopian tubes removed — and also had a colonoscopy, which identified a tubular adenoma that was surgically removed.
"Lynch syndrome is something she wouldn't have been tested for under the current standard of care," Lincoln said. Lynch syndrome may have also explained why the patient developed endometrial cancer, Lincoln said. "This test, under old way of doing things, would not have been ordered at the time. But had it been, it might have helped catch the endometrial cancer early," he added.
Some of the pathogenic variants, such as the one found to be causative of Lynch syndrome, had clear guidelines about interpretation and explaining risk and recommendations. Other variants however, despite known to be pathogenic, do not have such clear guidelines. Often their exact contribution to cancer risk is unclear and as such it is difficult to make treatment recommendations based on those variants.
Lincoln said that of the panel of 42 genes, about 28 would be considered "diagnostically relevant" by physicians, while the remainder were included to add more data to the literature about how the genes may confer risk.
"A few of the patients were positive for genes that were very interesting from a research point of view and are being followed up on, but wouldn't be used in a diagnostic panel, at least not today," he said.
For example, the authors noted the frameshift mutation in SLX4 that was found in two participants. The mutation was recently found to cause Fanconi's Anemia and is predicted to convey a two-fold increase in breast cancer risk, although "SLX4 mutations appear rare in breast cancer families," the authors wrote.
Because the two participants had a strong family history of breast cancer that included at least three family members diagnosed before age 50, the authors recommended that the participants receive more frequent MRI screenings.
Lincoln said that the SLX4 gene is not currently included on InVitae's diagnostic panel because "the implications of it to a physician, in terms of what they should tell a patient to do next, are not clear."
Aside from identifying the clinically significant pathogenic variants, the InVitae team also found 428 variants of unknown significance in 39 genes among the 175 participants, or about 2.1 VUS per participant. The vast majority, 380, of those variants were novel. Using a prediction algorithm, they concluded that 151 were benign, 65 were probably damaging, and 50 were possibly damaging.
As targeted sequencing panels include more genes, "you will, by definition, encounter more variants in that patient that have unclear interpretations," Lincoln said. The more genes tested for, the more variants of unknown significance that will be found. Part of the reason for including not only the diagnostically relevant genes, but also genes that are interesting from a research point of view and have associations with inherited cancer was to see how many variants of unknown significance would be found and to start to study how much of a burden dealing with those variants would be to physicians, genetic counselors, and the patients themselves, Lincoln said.
Determining the correct number of genes to include on a diagnostic panel can be tricky, Lincoln added, and is a balance between including enough genes so that important information is not missed, but not too many genes as to drown in information of unknown significance.