Karen Titus
August 2026—Testing for homologous recombination deficiency (HRD) is not for the faint of heart. Even the experts agree.
Tricia Numan, MD, gynecologic pathologist and director of the Women’s Health Initiative, Department of Pathology and Laboratory Medicine, Roswell Park Comprehensive Cancer Center, Buffalo, NY, says that when she first became involved in the topic, “I didn’t realize how many components went into HRD testing,” including germline alterations, somatic alterations, and various classifications of genomic instability.
Moreover, she says, the key clinical trials to date have used different cutoff numbers for their genomic instability score, used to determine HRD status. “I don’t think everyone appreciates that,” she says.
“And the cohort groups had various inclusion criteria,” she adds. “There are different combinations of cohorts, in addition to the different medications, in addition to the different HRD cutoffs. It’s very complex.”
HRD is a deficiency in a certain type of DNA repair; specifically, it addresses how the cell fixes a double-stranded DNA break, which is a type of DNA damage. Homologous recombination deficiency is both a biomarker and a biological driver of ovarian cancer, says Kyle Strickland, MD, PhD, director of medical affairs, Labcorp.
Information about HRD is key to determining which patients might benefit from highly effective PARP inhibitors, including olaparib, rucaparib, and niraparib. And as research continues to unfold, and as new tests and therapies enter the market, homologous recombination proficiency might elbow its way into clinical discussions as well.
Ovarian high-grade serous carcinoma can be broadly divided into two molecular groups: homologous recombination deficient and homologous recombination proficient (HRP) tumors. This distinction is clinically important because HRD tumors are more likely to respond to platinum-based chemotherapy and PARP inhibitors.
HRD most commonly results from alterations in genes involved in homologous recombination DNA repair. These alterations may be germline (inherited) or somatic (tumor-acquired) and include BRCA1, BRCA2, and other homologous recombination repair genes such as PALB2, RAD51C, RAD51D, BRIP1, ATM, and CHEK2. “Whether the alteration is germline or somatic, the biological consequence can be the same: impaired homologous recombination repair at the tumor level,” says Dr. Strickland.
Beyond identifying specific gene alterations, HRD can also be detected by measuring the genomic scars left behind when homologous recombination repair fails. These scars include loss of heterozygosity, which reflects loss of one parental chromosomal region; telomeric allelic imbalance, which represents unequal allelic content extending to chromosome ends; and large-scale state transitions, which are large chromosomal rearrangements indicative of genomic instability. Together, these measures form the basis of genomic instability scoring used in many contemporary HRD assays.
In contrast, HRP tumors lack evidence of homologous recombination repair deficiency and generally demonstrate lower levels of genomic scarring. These tumors typically derive less benefit from PARP inhibition and remain an area of active investigation as researchers seek to identify alternative therapeutic vulnerabilities.
“In healthy somatic cells,” Dr. Strickland says, “homologous recombination functions as a high-fidelity repair mechanism.” But if the repair pathway is defective—for example, if BRCA1 or BRCA2 is mutated or silenced epigenetically, “the cell cannot repair those breaks cleanly.” Instead, it relies on a more error-prone pathway, which results in a genomic disruption such as a deletion, duplication, or translocation. When HR is deficient, the genome accumulates structural damage in the form of genomic scars. “These genomic scars are molecular evidence of HR deficiency.”
“Genomic scar” is the common term for the components of the genomic instability score, Dr. Strickland says, used to describe the physical manifestation of defective DNA repair, which will persist as cells divide. “It’s like scars on the DNA. It’s a very apt description.”
The genomic scars themselves do not drive therapeutic sensitivity. Rather, they serve as evidence of an underlying defect in homologous recombination repair. It is that repair defect, not the scars themselves, that renders tumors vulnerable to platinum agents and PARP inhibition, which exploit the tumors’ inability to use the HR pathway to accurately repair DNA.
“HRD is a driver of cancer,” Dr. Strickland says. “We know this because patients with pathogenic BRCA1 and BRCA2 mutations in the germline are at an increased risk for cancer.” This is, in fact, the source of much of the early HRD data, he says. Not only do the hereditary cancers have BRCA1/2 mutations, but quite often spontaneous cancers do as well.
At a broad molecular level “what you see in the ovarian high-grade serous carcinomas,” he says, “are two types of tumors: those with evidence of homologous recombination deficiency and those that appear homologous recombination proficient.”
That translates into interesting clinical phenomena, he says. First, a germline alteration in BRCA1 or BRCA2 means the cancer may develop earlier in life. While that’s generally considered bad, it also means these cancers have a DNA repair defect that makes them better able to proliferate—and that increased proliferation makes them more susceptible to the platinum-based chemotherapies.
Conversely, he says tumors that are HR-proficient—“a subtype that hasn’t received as much attention”—tend to be relatively resistant to platinum therapy. In contrast to BRCA-associated HRD tumors, which often present at a younger age, HR-proficient tumors are more frequently encountered in older women and may be associated with replication stress phenomenon.
Women with germline BRCA1 or BRCA2 mutations have significantly increased lifetime risk of developing high-grade serous ovarian cancers, Dr. Strickland says; moreover, as noted, HRD extends beyond those mutations and can be driven by mutations in any of the other genes that compromise the HR repair pathway.
“When HR is deficient,” he says, “the cells will accumulate DNA damage and chromosomal rearrangements as a byproduct of the defective DNA repair.” The resulting genomic instability, which can be measured, is a hallmark of HRD-positive tumors.
Information about HRD testing and PARP inhibitors has emerged from a handful of studies in recent years. The landmark trial was the SOLO1 trial (Moore K, et al. N Engl J Med. 2018;379[26]:2495–2505), which studied maintenance olaparib in patients with newly diagnosed high-grade serous or endometrioid ovarian cancer.
The study looked at 391 patients; 388 had a germline BRCA1/2 mutation, and two had a somatic BRCA1/2 mutation. After a median follow-up of 41 months, the study notes, risk of disease progression or death was 70 percent lower with olaparib than with placebo. “That was huge for ovarian cancer treatment,” says Dr. Numan.
Since that initial trial, several others have demonstrated both the importance of PARP inhibitors and the nuances of related testing, she says. The PRIMA trial (González-Martín A, et al. N Engl J Med. 2019;381[25]:2391–2402), which assessed use of niraparib in patients with newly diagnosed high-grade serous or endometrioid ovarian cancer, looked at a subgroup with HR-deficient tumors, identified by the presence of a BRCA deleterious mutation, or an HRD score of at least 42, Dr. Numan says. Of 733 patients, 373 had tumors with HRD; among this group, 223 had tumors with BRCA mutations and 150 did not.
The study’s authors note that the clinical benefit of niraparib in the overall population was driven not only by the subgroup of patients with BRCA mutations; in fact, those with HRD showed a significant clinical benefit in patients with and without BRCA mutations. And in a subgroup of patients with HR proficiency, the longer median duration of progression-free survival suggests the drug has mechanisms of action other than those involved in repairing DNA damage.
Then, says Dr. Numan, there’s the PAOLA-1 trial (Ray-Coquard I, et al. N Engl J Med. 2019;381[25]:2416–2428), which looked at olaparib/bevacizumab as first-line maintenance in high-grade serous, endometrioid, or other nonmucinous epithelial tumors with any BRCA mutation status and any HRD status (score 42 or higher for HRD). Various cohorts included somatic BRCA mutation, BRCA wild-type, and HRD-positive with somatic wild-type BRCA.
Finally, Dr. Numan points to the VELIA trial (Coleman RL, et al. N Engl J Med. 2019;381[25]:2403–2415), again looking at various cohorts and therapy combinations. Particularly notable, she says, is that this trial used an HRD score of 33 or higher as the cutoff; other trials used a cutoff of 42.
“There are several different combinations and nuances involved in these trials,” Dr. Numan says, putting it mildly.
Hence the complexity of the supposedly simple question—Which tests should we order?—from clinicians.
When that question comes up at Roswell Park, Dr. Numan says she and her laboratory colleagues are fortunate. “Our gyn-oncologists are very up to date on what tests they should be ordering, and why.”
At Roswell, they’re typically using the Myriad MyChoice CDx and FoundationOne CDx tests—the only two that are currently FDA approved—but will order additional germline or somatic genetic testing depending on the clinical situation, primarily to capture non-BRCA HRD-associated genes. “At Roswell Park, we have a targeted NGS panel, the Oncomine Precision Assay.” But she says it’s apparent that many colleagues are doing HRD send-outs.
“My impression is that most of them are doing the testing following the surgical staging,” she says. “In suspected high-stage epithelial ovarian cancers, we’ll get a biopsy of the omentum to establish the diagnosis, and then the patient will receive neoadjuvant chemotherapy. The patient will receive three cycles of chemotherapy, followed by the interval debulking surgery, which typically includes the uterus, cervix, bilaterial fallopian tubes and ovaries, omentum, and any peritoneal disease or lymph nodes of concern. And that’s when they’re going to finish primary chemotherapy, followed by a maintenance PARP inhibitor.”
Neoadjuvant chemotherapy can make HRD testing challenging if the tumor has been eradicated. “Sometimes we have to go back to the original biopsy,” Dr. Numan says.
Sample size—always an issue—is particularly important with HRD testing, Dr. Strickland says. “We can do a lot with a little, but we can’t do everything with a little.”
While the genomic instability score is important, BRCA1 and BRCA2 “are the most important biomarkers for HRD right now,” given that they also suggest a hereditary component. But while it’s relatively easy to detect single nucleotide variants and small insertions and deletions, it’s harder to detect copy number losses, he says.
Dr. Numan has advice of her own, should the need arise. “If my gyn colleagues were to ask me what to do, I would probably tell them two things.”
The first would be to remind them that there are only two FDA-approved tests right now. “But that doesn’t cover all the nuances, so ordering germline or somatic genetic testing, depending on the clinical picture and patient wishes, is probably a good step.”
Second, she says, she would delve into the ongoing developments in HR-proficient tumors with RB1, which has suggested compelling differences in patient survival. That points to a potential need for additional testing beyond what’s provided by the currently approved tests for HRD. Using testing to gain information about disease process, and not just HRD status, will likely be valuable.
One other aspect of HRD testing gives Dr. Numan pause. “For better or for worse,” she says, “our clinical colleagues may see the HR result as very black and white.” But as recent studies looking at tumor heterogeneity have shown, there can be cases where one tumor can be tested multiple times with different results, especially those that test around the 42 cutoff. “Three times it may be considered deficient, and the fourth time it’s considered proficient,” she says.
“So what do we do about that gray zone?” she asks. “Should we be lowering the cutoff to 33 instead of 42?
“These are questions still to be answered,” she continues. That’s hardly a surprise, of course, given that the field is still young. “We’re not even a decade in. And we don’t have a lot of data for this gray zone. But we all know that high-grade serous carcinoma is very heterogeneous. If the score results at 41, it might be worth it to test another area of the tumor and to also consider the genetic alterations associated with the tumor.”
Where do other quandaries lie? Dr. Numan suggests taking a more expansive view of the field. “We tend to think of HRD related to high-grade serous carcinoma and BRCA1/2 mutations. But we have to remember that endometrioid carcinomas were included in the SOLO1 trials. Additional HRD research will be beneficial for other tumors that we don’t have a lot of data on right now, such as clear cell carcinomas or maybe carcinosarcomas.” The PAOLA-1 trial, she notes, included patients with other nonmucinous epithelial histology if they had a deleterious germline BRCA1 or BRCA2 mutation.
“We get very fixated on high-grade serous carcinoma, but we should be applying this to other histotypes as well,” she says.
Dr. Strickland agrees. “We’ve seen cases where BRCA is altered in clear cell carcinomas, and where clear cell carcinomas have responded to PARP inhibitors. [These] cases make me go, Hmmm. I think there are lot of open questions regarding ovarian cancer in molecular biology, and how they respond to therapies.”
There are, of course, questions about what to test, and with what, given that there are multiple emerging tests for HRD, loss of heterozygosity, and combinations of LOH and large-scale state transition (LST) scores, some of which also include BRCA testing. Dr. Numan observes that National Comprehensive Cancer Network genetic testing guidelines recommend that patients with ovarian cancer or fallopian tube cancer/primary peritoneal cancer should have genetic risk evaluation with germline and somatic biomarker testing. The American Society of Clinical Oncology recommends germline testing for BRCA1/2 and other ovarian cancer susceptibility genes. In women who have negative germline results, somatic tumor testing for BRCA1/2 pathogenic or likely pathogenic variants should be performed. The Society of Gynecologic Oncology recommends universal germline and somatic testing for ovarian cancer, and the American College of Obstetricians and Gynecologists recommends universal genetic counseling and testing as well as cascade testing.
As he looks over the current testing landscape, Dr. Strickland praises Myriad’s current genomic strategy because it looks for LOH, LST, and telomeric allelic imbalance, combined with BRCA status.
In certain scenarios, he says, perhaps it’s biologically early in the process and therefore there’s not enough tumor present to accurately identify the presence of genomic scars. “You may see some BRCA-altered tumors that don’t have a high genomic instability score.”
Myriad has a “clinically validated cutoff of 42,” Dr. Strickland says, a number that was developed using data sets enriched for BRCA-altered tumors and has served as the basis for several pivotal clinical studies. The test is designed to identify those tumors, which probably respond best to PARP inhibitors. “But we know those are not the only players in the HRD pathway.” Non-BRCA HR-related genes—such as PALB2, CHEK2, RAD51, etc.—can also be defective, as noted.
But there’s the reasonable question of whether non-BRCA alterations have the same penetrance. “They don’t,” says Dr. Strickland. “They don’t occur at the same frequency as the BRCA1 and BRCA2 as far as the HRD-related lesions are concerned. So it turns out that just looking at the genes that are altered is not a good way to evaluate HRD,” he says.
Labcorp has its own comprehensive test, he says. The company’s OmniSeq Insight, an NGS laboratory-developed test for solid tumors, has been validated to report HRD status and detect both the causal mutations and the genomic instability score (GIS). On the DNA side, the assay interrogates the full coding regions of 523 genes, which permits detection of single nucleotide variants, insertions and deletions, and copy number alterations, including amplifications and losses. It also assesses microsatellite instability and tumor mutational burden. On the RNA side, he continues, the test evaluates 55 genes for clinically relevant fusions. PD-L1 immunohistochemistry is included to support immunotherapy decision-making. The genomic profiling “adds a layer of biological insight that HRD testing alone may not capture.” The test incorporates an optimized version of Illumina’s TSO 500 HRD workflow, providing both genomic instability measurements and evaluation of HRD-related genomic alterations.
Dr. Strickland says the GIS is the basis for any testing that might be done, to capture the most patients. “Some labs might just be looking at LOH alone,” while others might be doing something else. “By using a GIS score, you capture more patients who benefit from PARP inhibitors.”
“If I had to think about this as a reader of CAP TODAY,” he says, “you could certainly start with the Myriad MyChoice to learn about GIS and BRCA.”
Then the question becomes, “Is there any utility to looking at these other genomic alterations that appear in high-grade serous ovarian cancer?” he asks. “Because there aren’t many targetable alterations in ovarian cancer. By that, I’m thinking of kinases and other sorts of fusions you might see that are targetable in things like lung cancer. Those just don’t happen very often in high-grade serous carcinoma.”
He says, however, that he and his colleagues have found that reporting out panel-based NGS tests can offer another level of assurance. CCNE1, for instance, is a gene that is often amplified to a high degree in tumors characterized by replication fork stress.
“That’s the direct opposite of the HRD phenotype,” he says. “So if I’m looking at a report, and I see there’s a robust CCNE1 amplification, and the tumor has a borderline GIS, that might lead me to believe this is an HR-proficient tumor.” Seeing BRIP1 and TP53 alteration, on the other hand, and a GIS of 42, “leads me to believe this is a non-BRCA HRD-related tumor.”
Such information should facilitate deeper thinking about these tumors and guide useful research. “Because the genomics is informative,” Dr. Strickland says. “And we’re just now entering an era where that information is becoming available.”
It’s already known that homologous recombination deficiency will deliver genomic scars. What isn’t known is whether those scars will persistently grow after therapy. Little is known about BRCA methylation status with many current tests, which concerns Dr. Strickland. “This is one way that BRCA [function] is lost,” he says, “but it’s lost epigenetically. So patients may recur quickly if that methylation is reversed.” For now, he says, there’s little research being done to look at BRCA methylation in platinum-resistant cases; even if there were, there are few current treatment options for such cases.
His own primary area of research is the immune-related effects from DNA damage repair. Some tumors with DNA repair defects—he speaks specifically about those that are microsatellite unstable—elicit a very strong immune response. “Just due to the nature of the DNA damage repair defects, they’re ‘spitting off’ aberrant nucleotides left and right,” and the creation of neoantigens recruits an immune response, he explains.
This knowledge is being put to use in treating colon and lung cancer with checkpoint inhibitors, which has led to robust responses. Whether this will also work with ovarian cancer remains a question. Early efforts showed no benefits, but in the past year data have emerged that suggest there is a subset of patients with ovarian cancer who might respond.
Naturally, this gives rise to more testing-related questions. “Can we use [the immune-related differences] to our advantage by detecting [them]?” he asks. And if the answer is yes, “How do we detect it?”
The immune-related differences between HR-deficient and HR-proficient tumors are persuasive—the two appear to be operating on different pathways, he says. There are increased tumor-infiltrating lymphocytes in BRCA-altered tumors compared with HR-proficient tumors. And different immune pathways are being activated due to greater expression of PD-1 and PD-L1 in BRCA-altered tumors.
“So you can imagine a scenario where you could detect and potentially treat tumors just based on how those tumors are making the immune system respond,” he says. Moreover, it makes sense to address HR-proficient tumors, which currently have fewer treatment options.
As HRD testing becomes more routine as part of comprehensive genomic profiling, compelling patterns are emerging.
“I would say for the most part, the HR-deficient tumors are characterized by a certain pathway,” he says. HR-proficient tumors are more of a mystery. It’s possible that the HR-proficient tumors may be linked to replication fork stress. “My current mental model—and I may be wrong—is that HR-deficient tumors are causing more deletions, causing LOH,” he says. “Whereas the HR-proficient tumors may be stalling; they may be creating more amplifications.”
“One of the things I’m trying to get a handle on in my research,” he adds, “is what is the origin of HR-proficient tumors? Are they all the same, or are they different?” In a Scientific Reports paper published last year (Strickland KC, et al. Sci Rep. 2025;15[1]:29523), he and colleagues identified HR-proficient tumors that lose RB1 and noted they behaved more akin to HR-deficient tumors. The loss of RB1, which is a cell cycle regulator, may contribute to increased proliferation and thus confer a survival advantage. Another possibility, he says, is that these are borderline HRD tumors, and loss of RB1 is simply a signpost that they’re becoming more HRD-like.
“Emerging data suggest HR-proficient tumors that have retained RB1 are the worst actors,” with survival that is less than HR-deficient tumors.
The field raises more questions than a congressional inquiry. But that also opens the door to opportunity.
“There have been so many surprises,” Dr. Strickland says.
Take, for example, genomic scars. One of the mechanisms of resistance for HRD tumors and PARP inhibitors involves BRCA reversion mutations—nonfunctional BRCA suddenly becomes functional again, and patients become resistant to PARPs. The genomic scars become, for all intents and purposes, static. “They don’t accrue more, or maybe they accrue at a lesser rate,” he says. This has, not surprisingly, launched countless research questions.
It also points to the wild and unpredictable nature of the field. “The whole idea of BRCA reversion mutations, that was just so exciting and almost unbelievable when it was first presented at conferences,” he says.
Every surprising turn, however, is also new knowledge “that we can capitalize on,” he says. “And anything that can help our patients is going to be welcome.”
Karen Titus is CAP TODAY contributing editor and co-managing editor.