Karen Titus
February 2025—Genetic profiling has long had proven winners in oncology: somatic testing of tumors, and germline testing for surveillance and to identify potentially affected relatives. These are molecular pathology’s version of the Beethoven symphonies heard on countless classical programs.
And then there is an emerging dark horse: germline testing to identify variants to direct therapy. If not a complete surprise, it appears to be an undersung approach—Beethoven’s Mass in C Major—even for cancers where universal genetic testing is recommended.
In a relatively recent study from Memorial Sloan Kettering Cancer Center (Stadler ZK, et al. J Clin Oncol. 2021;39[24]:2698–2709), researchers looked at using germline mutation testing to guide therapy for patients with advanced cancers. The authors’ analysis of data from nearly 12,000 patients, across more than 50 malignancies, “goes a long way” toward showing the importance of such testing, says Jonathan Nowak, MD, PhD, associate director of the Center for Advanced Molecular Diagnostics at Brigham and Women’s Hospital.
The therapeutic relevance depends on the tumor type. An alteration in a key gene involved in homologous recombination is more relevant for a patient with breast, ovarian, or pancreatic cancer, for example, than for a patient with colon cancer. But overall, the authors report, eight percent of patients harbored a germline variant of therapeutic importance. “It’s really quite high,” says Dr. Nowak, who is also associate professor of pathology at Harvard Medical School. “That’s a major kind of therapeutic option that you can put on the table for patients.”
“What we have the privilege of doing is providing this testing in an unselected manner at Memorial Sloan Kettering,” says Diana Mandelker, MD, PhD, a coauthor of the MSK article and molecular pathologist and director of the MSK diagnostic molecular genetics laboratory. Patients who have somatic testing performed in the context of tumor-normal sequencing can also consent to germline testing, regardless of whether they meet National Comprehensive Cancer Network or other criteria for genetic testing.
The study cohort—11,947 patients from 2015 to 2019—looked at how many patients were positive for a pathogenic germline variant in a hereditary cancer susceptibility gene. Seventeen percent harbored a likely pathogenic or pathogenic variant.
That itself is of interest. “But then they went further,” Dr. Mandelker says, “because they looked at what patients had actionable, or targetable, germline alterations from a therapeutic perspective”—as noted, eight percent of patients with advanced cancer had such a mutation.
The MSK researchers pushed even further; they looked at which patients received therapy based on their germline findings. As it turns out, 40 percent of patients with a therapeutically actionable variant received germline genotype-directed treatment.
The numbers are doubtless higher now, Dr. Mandelker says, “because there have been more and more approvals and expanded use for targeted therapies for alterations that occur predominantly in the germline.
“Honestly, it’s been a huge paradigm shift in the last 10 years,” she continues, citing approval of the first PARP inhibitor in 2014. “We used to perform germline testing for patient benefit in terms of surveillance or prophylactic measures, and for family members.”
But once a patient developed cancer, “the germline could have been an afterthought,” Dr. Mandelker recalls. “And now it’s at the forefront, because so many of these mutations with targeted therapies—specifically in the homologous recombination pathways, like BRCA1 or BRCA2, PALB2, et cetera—in most cancer types occur much more frequently in the germline than they do as somatic alterations.” For many cancer types, “an up-front standard of care, for us, is to perform germline testing.”
It’s a relatively underused offering, Dr. Nowak says.
“Both on the clinical side, for patient-facing physicians, and also on the laboratory side for molecular pathologists, there’s not a broad recognition of the utility of this type of testing. Many physicians don’t naturally think of performing germline testing to guide cancer therapy even though germline alterations in roughly 20 genes can impact such therapy.”
He echoes the observations of Dr. Mandelker as he con-siders the expanding role of germline testing. “What we’ve seen time and time again with molecular-based therapy selection is that most drugs end up being approved for a specific tumor type first, based on molecular indication.” It’s not infrequent for oncologists to then draw conclusions from an approval, he says, and use it to target the same alteration in a different tumor type.
“We’ve done that since the dawn of molecular testing,” he says. “So I think that’s where a lot of the potential relevance comes in. Could you use the same drug in an off-label indication for a tumor that is driven by the same molecular alteration?” While that doesn’t always work, he acknowledges, often it does. Oncologists used PARP inhibitors in an ad hoc manner to treat pancreatic cancers in patients with germline BRCA1, BRCA2, or PALB2 deficiency before approval was granted, for example, and the sometimes dramatic responses were one factor that led to the trial that brought about approval for olaparib in pancreatic cancer, he says. “So I think the molecular field is comfortable reporting things out as potentially actionable, even though an alteration does not rise to directing approved standard of care therapy yet.”
With multigene germline panels and whole exome analysis becoming more widespread, it’s evident that patients with advanced cancer have a pathogenic or likely pathogenic germline variant in a cancer susceptibility gene—about 17 percent, as the MSK researchers noted. Nevertheless, more than half (55 percent) do not meet genetic testing criteria based on current clinical guidelines and may be excluded from receiving germline-directed treatments. In the study, more than 80 percent of PALB2 carriers did not know about their heritable and potentially actionable alteration.
Moreover, Dr. Mandelker says, there is a constant need for more outreach to expand genetic testing to all eligible populations. Studies have shown lower uptake in genetic testing among certain minority populations; it’s also not offered to cancer patients in general at the rates suggested by current clinical guidelines. It would be worth exploring, she says, whether some groups of patients are offered testing at lower rates than others, or if patients are being offered the testing but then decline it.
“For those of us who are privileged enough to be at these academic institutions, it’s really just part of our standard of care to perform this genetic testing on our cancer patients,” she says. “But I’d like to see that expand further into the community.”
There is, in short, room for growth.
But how much? Pinning down the appropriate rate is more a matter of pliability than precision.
Dr. Nowak keeps two major factors in mind. One is the patient population, he says. “The yield of germline testing is only as valid as the patients you’ve tested.” The other consideration is the gene panel, which also varies by site. A laboratory that considers alterations in 25 genes will identify a higher rate of therapeutically relevant alterations than laboratories that look at, say, 10 genes.
Moreover, he says, identifying mutations is not the same as copy number detection. “So you may be missing out on alterations that are targetable if you’re not optimally designing assays for detecting loss-of-function alterations due to copy number changes as opposed to just mutations.”
Given these variables, he says, current data suggest that the percentage of genes that are therapeutically relevant would be between five and nine percent. But the list will only continue to grow, he predicts. “It’s probably more of a floor right now than a ceiling.”
Broadly speaking, Dr. Nowak says, it’s helpful to think about relevant genes in the context of pathways. “The single biggest pathway [consists of] genes involved in the homologous recombination repair pathway, or that are closely related to that pathway,” which includes BRCA1/2, PALB2, BARD1, CHEK2, and other associated genes, such as ATM. Within that pathway, PARP inhibition is approved in a variety of contexts across different cancer types. In addition, the rates of alteration in those germlines are elevated in patients who have numerous, different types of cancer, including prostate, breast, pancreas, and ovarian cancer.
Mismatch repair deficiency is the second most therapeutically relevant pathway, he continues. “This is just slightly different in the sense that approvals for using MMR deficiency for targeted therapy are typically gated on actual MMR deficiency in the tumor itself, which often can be caused by a germline pathogenic or likely pathogenic alteration in MSH2, MSH6, MLH1, or PMS2.” Simply having the germline alteration might not be sufficient to qualify a patient for treatment. However, the presence of the alteration in the germline does make it more likely that any tumor that develops in the patient is mismatch repair deficient and is likely to respond to immunotherapy.
“The next kind of classic alterations you think about are oncogenes that typically have gain-of-function or activating point mutations,” Dr. Nowak says, including ALK, RET, EGFR, KIT, and MET, where there are germline syndromes that have inappropriate activation of those oncogenes, and for which there are targeted therapies. He adds to that list PTCH1 for patients with Gorlin syndrome, which can be targeted with vismodegib. “This set of genes is a bit more heterogeneous, but the appropriate germline alterations in all of them have implications for the use of targeted inhibitors.”
Keeping up with changes in the field will be a steady challenge. While that’s no different from tracking the expanding array of indications for somatic variants, Dr. Nowak says, the importance of germline alterations is less broadly understood, at least for now. “It’s just something that fewer people pay attention to.”
While most of the focus has been on later-stage disease, that could also evolve, depending on the tumor type.
“In the beginning,” Dr. Mandelker says, “we did limit it more to patients already in the advanced setting, or those who were going to need systemic therapy. We’ve changed that now, where a lot of patients at presentation are getting genetic testing.”
Clinical guidelines have expanded their recommendations for who should receive testing “practically every year,” she says. Accordingly, oncologists and surgeons can order genetic testing for their patients on presentation, even if therapy may be months or a year off. “I personally think it’s beneficial to have this information as early as possible, and not be at a point where they want to start therapy or develop a treatment plan and have to wait weeks for results.”
Says Dr. Nowak: “If you’re at the point where you are sequencing for other targetable genomic alterations, this type of testing should be considered alongside that.” And for some indications—colon cancer, for example—MMR testing is now therapeutically relevant in the first-line setting to identify patients who are likely to benefit from immune checkpoint inhibition.
In general, Dr. Mandelker says, if a large enough panel is done, germline testing should be needed only once, as opposed to sequential testing. “While sequencing may only need to be performed once, clinical genetics labs constantly update their curation of detected variants.” Testing may reveal a variant of uncertain significance, for example, “but as evidence evolves over the course of several years, the variant classification may change to either pathogenic or benign. The lab will amend those cases and issue updated reports when a genetic variant reclassification occurs.”
The tumor-normal sequencing program at MSK began in 2014, when the institution launched its MSK-IMPACT next-generation sequencing assay that identifies mutations, fusions, and copy number alterations in tumor specimens.
From the beginning, Dr. Mandelker says, it was a matched tumor-normal assay but only somatic findings were reported initially. “We had the matched normal essentially to distinguish somatic variants from germline ones.”
In 2015, they launched the germline components of the assay. “So then any patient who was consenting already to receive somatic testing, in the context of tumor-normal sequencing, had the option to also consent to germline analysis.”
Since then the program has been built out even further, she says. “Sometimes germline results are needed more rapidly to inform treatment decisions, so we also allow for blood- or saliva-only testing, if the patient doesn’t have a tumor specimen available.”
Dana-Farber Cancer Institute and Brigham and Women’s launched paired tumor-germline sequencing testing in June 2023.
As one of the key elements of this effort, Dr. Nowak says, “we built a new reporting system that allows us to include information about germline variants that are therapeutically relevant in the somatic report. To reduce confusion, we have a special mechanism for how this works—only actionable germline variants in a set of 19 genes with therapeutic relevance are included in a dedicated section of the somatic report.” Such variants are jointly reviewed by the somatic and germline faculty members on service.
In deciding which genes to include, he says, “we spent a lot of time talking with the adult oncologists, but we also got input from the pediatric oncologists,” who were interested in a slightly different set of genes. Based on those discussions, the lab added three genes with pediatric relevance: TP53, DICER1, and RB1. Although targeted therapies are not yet available for those genes, they can influence pediatric oncologists’ selection of chemotherapy.
The laboratory report is critical. A 2023 editorial (Stadler ZK, et al. JAMA. 2023;330[1]:30–32) suggests that labs can jump-start germline testing, in part by using automated clinician notification of eligibility and by integrating testing criteria into pathology reports for cancers that have inherited susceptibility genes.
Possibly, says Dr. Nowak. “Although I think the first thing to say is that laboratories are only part of the solution.”
The biggest challenge of doing this sort of work is, in many ways, being able to operationalize paired tumor-germline sequencing, he says. Most clinical sequencing—whether at an academic center or in a community setting, or as a send-out to a commercial lab—often entails a tumor-only specimen, he says. “And although you can make a pretty good guess about what might be a germline alteration, the reality is, that’s impossible to do perfectly with somatic-only sequencing.”
The other option is to do germline-only sequencing. Though sometimes done at medical centers internally, Dr. Nowak says, it’s often sent out to reference labs. This poses a different headache: Most of that interpretation and reporting—and often the workflow of those results—is focused on hereditary cancer implications of a variant; indeed, he says, the reference lab may not even be aware of the specifics of a patient’s cancer diagnosis.
As a result, those reports may not mention anything about therapy, he says. Even if they do, the workflow for such reports typically routes positive findings to the genetics department or genetic counselors, who take the lead in returning the information to patients and discussing the results with respect to hereditary cancer. A clear-cut path to oncologists is not always assured.
Both of those pathways are suboptimal for understanding the therapeutic relevance of germline variants. Dr. Nowak says the best approach would be to do paired tumor-germline sequencing, which places the genetic profile of the tumor alongside the germline profile of the patient. “You have those side by side. In your somatic report, you’re putting a list of all the targetable alterations, if there are any, from the somatic side of things: Is this a TMB-high tumor? Does it have a targetable BRAF mutation, or an EGFR alteration, or whatever your favorite targetable alteration is?”
But, he continues, setting up the right reporting system also allows labs to say, for example: Because the patient has a germline X, Y, or Z alteration, here is an additional targeted therapy option. Consider that alongside the list of somatic alterations.
This is the approach he and his colleagues use. He doesn’t mind saying it was not their first choice.
“When we first started planning for tumor-germline sequencing, we thought it would be much cleaner to keep only somatic results in the tumor report, and only germline results in the germline report,” he recalls.
But as history attests, separate but equal is bound to fail.
The push to include germline results in the somatic report came from oncologists. “It makes the interpretation on our side a little more complicated, to be honest,” Dr. Nowak says.
But if writing an integrated report takes more effort, having complete information about therapeutically relevant results on one page helps oncologists, he says. “There is real value to putting information in just one place in the medical record, where you don’t have to scroll through the EMR to look for somatic and germline sequencing results, which often live in different parts of the medical record,” particularly if testing is done in different labs and physicians are trying to track down PDFs or even scanned reports that cannot be searched easily. “The time saved and reducing the chance of missing valuable information, because someone didn’t have the time or even think to look for it in two different places, is a real benefit for patient care.”
The paired reporting, he says, also allows pathologists “to comment on the status of whether there’s a second hit, if you will, in the tumor that goes along with the germline alteration that was found.” Most germline-driven approvals to date don’t yet consider the status of a second hit or loss of another allele in the tumor, Dr. Nowak says. But patients with a completely inactivated gene in their tumor are usually the ones who respond better to therapy, whether the mechanism of a second hit is a second mutation, loss of the wild-type allele, or a copy neutral loss of heterozygosity. With that information, “I can tell the oncologist, not only does this patient have a germline BRCA1 alteration, but there’s also a single copy BRCA1 deletion in the tumor—this is an excellent candidate for PARP inhibition, versus a patient with a germline alteration and no evidence of complete inactivation of BRCA1 within the tumor.” In the latter case, it’s difficult to say with certainty that the tumor is dependent on the loss of BRCA1 function. “We see that all the time. Not every patient who has a germline alteration has a tumor that’s driven by the germline alteration. Patients with hereditary cancer predisposition also just get sporadic tumors.”
Dr. Mandelker seconds that note of caution. “You have to be very careful,” she says. She too cites the example of BRCA findings.
One of the biggest questions from her oncology colleagues comes when there is a germline pathogenic variant in an unexpected tumor type, such as a BRCA mutation in an esophageal cancer. “Because we have the matched tumor, they’ll often say, ‘Hey, does it look like there’s loss of heterozygosity in the tumor? Or is it an incidental finding? Do you think there’s something there we can target?’” In BRCA-associated tumors—breast, prostate, pancreatic, ovarian—the mutation is presumed to be (and in most cases is) driving tumorigenesis and development. When a BRCA germline mutation is found in other cancer types, it’s unclear whether it’s the culprit or an incidental finding. “That distinction is very important in terms of therapeutics. Only if it’s really involved in tumorigenesis and driving the tumor would a targeted therapy potentially be effective,” Dr. Mandelker says.
At the end of the day, Dr. Nowak says, the goal should be doing paired tumor-germline sequencing, as its value goes beyond simply identifying variants with therapeutic relevance. The latter, he says, “just gets you better data for standard somatic indications. Period.”
He recognizes the easier-said-than-done aspect of this. “Honestly, it’s much harder than you would think to get a tumor and germline specimen together in time and space, and then to have the appropriate infrastructure to communicate positive germline results back to patients and make sure they get appropriate follow-up. That’s harder than doing the testing in the lab.” But to have the biggest impact on patient care, “these are the things we need to be working on.”
At Memorial Sloan Kettering, Dr. Mandelker says, “We’re incredibly fortunate to have a large and very involved clinical genetics service. Even though it’s primarily the oncologists who are ordering the genetic testing, we as a lab provide every night a list of all the cases we’ve signed out to our clinical genetics service, so they can follow up on every positive case with the oncologist. These reports don’t get lost in an inbox with a lot of other reports. They are very proactive and do an extraordinary job to make sure these results are transmitted correctly.”
She continues: “Any cancer patient has a lot of test results, and to highlight this and incorporate germline findings with the somatic findings, and any other laboratory values and testing, is a challenge.” And though there’s movement generally toward integrated reports, with interpretive findings, “Not too many places are there yet.”
Occasionally the paired report shows conflicting information.
“You can sometimes run into challenges, which you wouldn’t anticipate unless you’re lining up that data side by side,” Dr. Nowak says. It’s rare—he calls them edge cases—but occasionally pathologists will see a pathogenic variant in the germline that is present in the tumor at a level that’s much lower than expected.
Some of these cases are high-level CHIP—clonal hematopoiesis—or underlying hematologic malignancies, he explains. Though unexpected, “that then allows you to say, ‘One, even though we found this alteration in the patient’s peripheral blood sample, it’s not a germline alteration; it’s a separate hematologic neoplastic process. And then two, it’s not relevant to the solid tumor because it’s not a germline alteration.’ It’s a little complicated, but working through why you have seemingly disparate results is informative.”
Sometimes he and his colleagues will encounter scenarios in which the pathogenic germline alteration leads them to expect a second hit in the tumor. Instead, they’ll see the tumor has rid itself of the alteration. “So there’s been a counterselection of the germline alteration in the tumor. It’s decided that, for whatever reason, it can grow better without the germline alteration.”
In such cases, pathologists can tell their oncologists that even though the patient has, for example, a germline ATM alteration, the tumor is wild-type for ATM. “So don’t consider treating this patient, even though on their germline basis alone they might qualify for therapy, because the ATM protein is not going to be inactivated in the tumor,” Dr. Nowak says.
He adds, “That again is something you could only appreciate if you line up the data from the same sequencing platform side by side.”
One of the bigger challenges he and his colleagues faced internally, he says, was dealing with the lack of clear-cut ways to classify variants. There are ACMG/AMP joint criteria for classifying variants with respect to hereditary cancer predisposition. “It’s a language we broadly speak,” Dr. Nowak says. And while “labs are a little more diffuse in how they do somatic variant classification,” there are AMP/CAP/ASCO guidelines for those interpretations as well.
But germline variants with therapeutic relevance don’t fall into either camp, he says. So he and his colleagues developed their own list of criteria for deciding whether a variant should be reported as therapeutically relevant. “It’s been an interesting exercise, but challenging,” he says, given the lack of broad consensus on how to make that determination.
Precision oncology knowledge bases could eventually play a role in helping physicians sort through the data. OncoKB has been an excellent resource for somatic variants, Dr. Nowak says, and ClinVar is the dominant resource for germline variants. While there’s no equivalent yet for germline variants with therapeutic relevance, Dr. Mandelker says OncoKB is planning to release this year a germline resource that annotates the cancer-type-specific therapeutic implications of germline variants.
Many of the challenges ahead for the field may be more practical than visionary at this point, she says. “A decade ago, it would have been hard to imagine—actually, I don’t think I could have foreseen this—that genetic testing for cancer patients would be so critical, so vital, for their treatments these days. It’s really an incredible time to be in the field.”
Karen Titus is CAP TODAY contributing editor and co-managing editor.