Paired Tumor-Germline Testing as a Driver in Better Cancer Care

Yap et al¹ provide an analysis of paired tumor-germline next-generation sequencing (NGS) in a large population of patients with cancer. One implication of the data is that paired testing provides important information about a patient’s cancer and its treatment and can address hereditary issues of importance to the patient’s family. This commentary considers the advantages of paired tumor-germline testing in cancer care.

A germline variant, when present, is usually the initiating driver in cancer formation. Driver genes are pathogenic or likely pathogenic (P/LP) variants of normal genes that give selective advantage to a cell, leading to cancer development, progression, or treatment resistance.² Cancer cells usually contain a few to several driver genes, and subsequent driver genes often appear in subclones. An initiating driver is the first gene to start a cell on the path to transformation, and initiating variants are typically present in all tumor cells and their subclones. An initiating driver in a hereditary cancer gene may be present in the patient’s germline and thus be present in the cancer cells or it may exist in the first cell in a sporadic tumor and not in germline.

When tumor-only testing identifies a P/LP in a known hereditary cancer gene, 2 questions arise. First, is the variant present in germline and thus likely to be hereditable, and is the gene the initiating driver? When germline-only testing identifies a P/LP variant in a hereditary cancer gene panel (note that nucleated blood cells are a proxy for germline), there are clear issues for the family, and the second question arises again. Is this gene likely to be the initiating driver of the patient’s cancer? These issues occur frequently in practice, as noted by Yap et al,¹ and suggest that uniform paired tumor-germline testing, which can answer these questions, provides value in patient care.

Paired tumor-germline testing improves understanding of both genomes. Tumor findings can elucidate germline variants. The detection of a P/LP or benign variant in tumor or blood is straightforward. However, a variant of uncertain significance (VUS) poses a barrier to interpretation of germline NGS. Most of the variants returned on germline reports are within the VUS category. When the gene with a VUS is on-tumor, meaning the gene is known to be involved in formation of the tumor under consideration, and when the family history causes concern, substantial uncertainty results. Paired testing can clarify the significance of a germline VUS when relevant additional changes are detected in the tumor tissue. Examples include a second hit, such as copy number loss/loss of heterozygosity for the second allele, or a second variant in the same gene. Other examples include signs of a relevant downstream effect, such as genomic loss of heterozygosity or homologous repair deficiency associated with a BRCA pathway-related VUS, or increased tumor mutation burden and microsatellite instability with a mismatch repair pathway-related VUS. When many genes are included on the panel, raw data files from tumor NGS can be analyzed for mutational signatures that suggest an initiating driver gene, for example, with a POLE or POLD1 variant and elevated tumor mutation burden.³ Data on RNA sequencing in tumor or blood samples reporting gene expression level or splice variants can be informative in VUS resolution.⁴

Tumor testing can also provide confirmation that a germline variant is truly the initiating driver of the patient’s cancer, given the possibility that the germline change is not the underlying culprit. An example would be the recent patient in clinic with NF1 syndrome and endometrial cancer where paired testing revealed clear evidence of a somatic POLE variant as the initiating driver rather than the germline NF1 pathogenic variant, with substantial therapeutic implications.

Paired tumor-germline testing can lead to better treatment choices. Hereditary cancer syndromes are caused by inactivated tumor suppressor genes. By comparison, activated oncogenes are poorly tolerated during embryonic development. Variant tumor suppressor genes and their inactive protein products are more challenging to target than are activated oncogenes. Attention thus turns from the gene itself to the aberrant downstream molecular pathway and subsequent effects of the pathway.⁵ Two notable examples are the use of checkpoint inhibitors to target high tumor mutation burdens from tumors related to a defective mismatch repair pathway (ie, Lynch syndrome when originating in germline) and poly adenosine diphosphate ribose polymerase inhibitors to exploit single-strand breaks from tumors related to the homologous repair/BRCA pathway (ie, hereditary breast and ovarian cancer syndrome when originating in germline). Several US Food and Drug Administration indications are now in place specifically for cancers with a germline variant, including breast, ovarian, and pancreas cancer. The most recent approval is the hypoxia-inducing factor 2α inhibitor belzutifan for any tumor in patients with von Hippel Lindau syndrome (ie, patients harboring a germline P/LP variant in VHL).

However, many trials and approvals now include patients with tumor-only NGS showing a P/LP in a hereditary cancer syndrome gene or downstream effect.^5,6 An example would be the enrollment of a patient in a checkpoint inhibitor trial whose breast cancer shows an elevated tumor mutation burden and microsatellite instability on tumor-only testing. This approach leaves open the possibility of a subclonal P/LP variant and thus a possible lower response rate, especially when the variant allele fraction in a tumor is reported and is less than the 45% to 55% expected with a germline variant. However, in patients with a tumor mismatch repair P/LP variant where germline was known, there were no average differences noted between response rates to checkpoint inhibitors in patients with or without a germline variant.⁷

Tumor-only testing misses an opportunity to reduce risk in family members. When tumor-only testing is performed, the gene underlying monogenic hereditary cancers will be missed.⁸ Family members who are undiscovered carriers may have poor outcomes because their risk is not managed with preventive approaches or surveillance, and family members who are not carriers may face needless worry. Reproductive issues may go unaddressed.

Paired tumor-germline NGS is challenging the classical syndromes and, thus, guideline-based testing. In 2 central points of the study by Yap et al,¹ it is clear that germline testing based on defined hereditary cancer syndromes and their associated guidelines (eg, NCCN guidelines for BRCA testing) misses nearly half of the germline variants in patients with commonly tested tumor types. Germline variants frequently exist in tumors for which the risk is not recognized and guidelines do not exist.⁹

Studies of broader, less-selected populations report the increasingly frequent finding of off-tumor germline changes. This confusing problem can take at least 2 forms. The first issue involves the importance of syndrome-related, germline pathogenic variants in patients with nonsyndromic tumors, such as a BRCA1 pathogenic variant in patients with lung cancer. Tumor types can form from initiating drivers that are rare to them, perhaps dependent on features such as variations in cell-specific patterns of gene expression. However, the reported frequency of these findings seems higher than expected.⁹ Because population-based studies do not allow for detailed analysis of family histories and genetics, this increased frequency will remain a mystery for some time. The second issue involves the importance of frequent germline pathogenic variants in highly penetrant genes, such as TP53, when Li-Fraumeni syndrome is exceedingly rare. These problems indicate that wider testing will change classical gene-based hereditary cancer syndromes into much broader associations and ranges of penetrance. This evolution may enhance the understanding of the role of hereditary cancer genes and will result from wider testing of tumor-germline pairs.

Published: May 20, 2022. doi:10.1001/jamanetworkopen.2022.13077

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2022 Henson JW. JAMA Network Open.

Corresponding Author: John W. Henson, MD, Hereditary Cancer Clinic, Georgia Cancer Center, Medical College of Georgia, Augusta University Health, 1411 Laney Walker Blvd, Augusta, GA 30912 (johenson@augusta.edu).

Conflict of Interest Disclosures: Dr Henson reported receiving fees from Calyx for review of clinical trials data outside the submitted work.