With continued advances in next-generation sequencing, germ-line mutations are increasingly being identified in tumors and normal DNA of patients. The significance of these findings, including their relevance and whether results arising from germ-line testing should be returned to patients, was the focus of three presentations during the Monday, June 1, Clinical Science Symposium “A Tale of Two Genomes: Interpreting Somatic and Germ-Line Sequencing Results Together.”
Dr. Kasmintan A Schrader
The Memorial Sloan Kettering Cancer Center Study
Massively parallel sequencing of tumor and normal tissues from 1,570 unselected cancer cases identified several germ-line mutations across multiple malignancies, with a degree of overlap with those recommended for screening by the American College of Medical Genetics and Genomics (ACMG; Abstract 1509).
“The aim of the study was to determine the burden of germ-line mutations in individuals with advanced cancers,” Kasmintan A. Schrader, MBBS, PhD, of Memorial Sloan Kettering Cancer Center (MSKCC), said.
The study was initiated at MSKCC to identify patients for appropriate targeted therapy. A 341-gene MSKCC panel (now expanded to a 410-gene panel) was used to probe tumor and matched normal DNA. Of the 68 different tumor types, 20% of the samples were from patients with breast cancer and 14% were from patients with non–small cell lung cancer.
Paraffin-fixed tumor samples were used to sequence tumor DNA; DNA from blood was used to sequence normal DNA. Subtracting the sequences obtained from normal DNA from those obtained from tumor DNA provides insights into tumor genotype associated with tumor phenotype.
Comparing normal DNA of patients with a reference from an “accepted standard” normal provides insights into mutations harbored in the patient’s germ line.
A survey of variants present in the normal germ line showed that each person harbored, on average, 63 variants; of these, 16% had at least one presumed pathogenic variant based on genes in the panel that have been linked to single-gene disorders.
Twenty-six of the 341 genes in the panel overlapped with the 56-gene panel recommended by ACMG to all physicians ordering clinical sequencing. Dr. Schrader indicated that for these 26 genes, patients in their cohort had, on average, 11.7 variants in the germ line.
Whereas 101 patients had at least one pathogenic variant, 807 patients (51%) carried one or more variants of uncertain significance. Dr. Schrader reported that the variant was retained in the tumors in 179 of 198 cases with a germ-line variant in a known cancer susceptibility gene, suggesting it may have contributed to the tumor phenotype. Of these 179 cases, 39 were associated with loss of heterozygosity or mutations in one of the alleles. However, 33% of these 39 cases involved mutations that did not explain tumor phenotype; for example, germ-line BRCA1 mutations were seen in patients with gastric cancer, which is typically not associated with mutations in BRCA1.
“The study showed that, not surprisingly, germ-line mutations are found in a subset of patients not selected based on family history or the type of tumor. Moreover, the finding that a susceptibility gene is present in a tumor indicates that it may contribute to the tumor phenotype in ways that are currently not understood completely,” Dr. Schrader said.
“Even when the goal is anticancer therapy only, and the panel is cancer focused, there will be germ-line findings,” discussant Robert McWilliams, MD, of Mayo Clinic, said. Although some of these mutations have implications for treatment, potential downsides include cost and time to discuss potential implications.
The MD Anderson Cancer Center Study
A similar study undertaken at the University of Texas MD Anderson Cancer Center showed that 999 out of 1,000 patients with cancer had at least one germ-line variant in 19 potentially actionable genes based on next-generation sequencing of tumor and matched normal DNA (Abstract 1510).
“Even when only tumor analysis is performed, it creates an opportunity for identifying pathogenic germ-line mutations incidentally; when tumor-normal pairs are used, it is an opportunity for directed germ-line analysis to identify pathogenic germ-line mutations,” Funda Meric-Bernstam, MD, of the University of Texas MD Anderson Cancer Center, said.
Dr. Meric-Bernstam presented data obtained using their Clearinghouse and Germline Companion protocols. Clearinghouse uses standardized somatic mutational analysis with the goal of identifying genomic alterations that may be targeted through appropriate clinical trials. The Germline Companion protocol is a research tool “to determine the frequency of germ-line polymorphisms and deleterious mutations in different tumor types and patient preferences for return of incidental test results,” she said.
Using a 202-gene panel in Clearinghouse, Dr. Meric-Bernstam reported that 999 patients had at least one variant in 19 actionable genes and 43 (4.3%) of these had a pathogenic germ-line cancer-associated variant. Mutations were reported in several genes, including BRCA1 (11 patients), BRCA2 (10 patients), TP53 (10 patients), MSH6 (4 patients), and PALB2 (2 patients). The 19 actionable genes included 18 from the 57 ACMG-recommended list and PALB, which has recently emerged as an actionable gene from a breast cancer gene panel.
Dr. Meric-Bernstam showed that germ-line mutations in BRCA1, BRCA2, MSH16, and PALB2 were present in more than 50% of the tumors analyzed. In addition, of the 43 patients with pathogenic germ-line variants, 16 did not meet requirements for mutational testing based on tumor type or family history. This raises the question of whether all patients with cancer should receive genetic testing.
Of the 1,167 patients in the Germline Companion protocol, 99% (1,157 patients) consented to the return of their results. Returning results to patients was recommended in the ACMG requirements in March 2013; in April 2014, ACMG amended this recommendation and provided an opt-out option for patients.
Dr. Meric-Bernstam reported that the 23 previously unknown actionable alterations found in this data set showed 100% concordance when the samples were also analyzed in a Clinical Laboratory Improvement Amendments–certified lab where clinical testing is performed routinely on a separate platform. Seven patients have undergone genetic counselling, and testing was initiated on new samples, for which results were confirmed in five patients and are pending in two.
In her discussion of Abstract 1510, Stacy W. Gray, MD, AM, of the Dana-Farber Cancer Institute and Harvard Medical School, said, “We are moving into a new era where even tumor-only sequencing uncovers incidental germ-line mutations.”
She indicated that germ-line mutations have ramifications for patients and potentially for their family members. She explained that studies to date show that many patients with cancer want to learn about incidental findings, especially when it comes to actionable information about cancer risk and the risk of developing noncancer conditions. She concurred with Dr. Meric-Bernstam that “most patients are interested in return of relevant germ-line results.”
In addition, Dr. Meric-Bernstam said, “Consideration should be given to soliciting patient preferences at the time of consent for trials with next-generation sequencing, and to developing a plan for return of results.” To contextualize the MD Anderson findings, Dr. Gray discussed some of the ways in which obligations related to the return of incidental findings may be different in a research setting compared with a clinical context.
Similar to the ACMG recommendations for clinical care, most groups advocate for the return of incidental findings in the research setting when they are analytically and clinically valid, have important health implications, and are actionable. Most groups also suggest that the preferences of research participants regarding the receipt of incidental information are elicited as part of the informed consent process.
Unlike the ACMG recommendations, however, many groups do not endorse a "duty to hunt" for genomic findings beyond what would be identified in the course of the primary research.
Dr. Gray said that there are significant patient and provider knowledge gaps when it comes to cancer genetics. “In order to ensure the success of cancer genomic medicine, education and decision support systems need to be in place both for patients and cancer providers as we move forward,” she concluded.
The University of Pennsylvania Study
Kara Noelle Maxwell, MD, PhD, of the University of Pennsylvania Perelman School of Medicine, reported that 11% of patients with high-risk breast cancer had deleterious mutations identified by massively parallel sequencing (Abstract 1511).
As multiplex panels become more complex, published studies of approximately 6,000 patients with BRCA1- and BRCA2-negative breast cancer have indicated that 3.7%-13.0% of patients have deleterious mutations in cancer-susceptibility genes. “What clinical features account for this range?” Dr. Maxwell asked.
In this study, massively parallel sequencing was performed on germ-line blood DNA from 810 patients using one of three custom Agilent panels. Results were reported for 20 genes associated with cancer susceptibility. Patients in the study were all diagnosed with BRCA1- and BRCA2-negative breast cancer with high-risk features: 325 had early-onset breast cancer; 326 had multiple primary cancers; and 422 had a significant family history of breast cancer.
Of the 810 patients in this study, 11% (87 patients) had deleterious or likely deleterious mutations. These included high-penetrance genes (TP53, PTEN, STK11, CDH1), moderate-penetrance genes (ATM, CHEK2, NBN, PALB2), possible breast cancer genes, and other genes. Deleterious mutations were detected in 14% of patients with multiple primary cancers, and in 6% of both patients with early-onset breast cancer and patients with a family history of breast cancer without a second primary malignancy. ATM mutations were more frequent in patients without multiple cancers, and other mutations were more evenly distributed.
Dr. Maxwell indicated that sequencing was underway for 275 additional patients with multiple primary malignancies to confirm these initial findings.
Discussant Allison W. Kurian, MD, MSc, of Stanford University School of Medicine, said that this was a large, well-conducted study that asked an important question about the prevalence of newly-identified mutations, such as ATM, CHEK2, and PALB2, according to a patient’s cancer diagnosis, family history of cancer, and age at cancer onset. The study contributes to the understanding of the spectrum of cancers associated with mutations in newly-identified cancer susceptibility genes, she said.
Watch the session: Visit the ASCO Virtual Meeting website.