Patients Having Ovaries Removed Without Clear Genetic, Medical Rationale, Study Suggests

This story has been udpated to include additional comments from a presentation by Jennifer Klemp.

NEW YORK – Women with variants in genes not known to increase the risk of ovarian cancer, and with no personal or family history of the disease, still had surgery to remove their ovaries hoping to reduce their cancer risk, often based on the advice of their doctors, according to a new registry study. 

At the American Society of Clinical Oncology's virtual annual meeting this week, researchers led by Susan Domchek from the University of Pennsylvania Health System presented the results from the study, which found that between 10 and 15 percent of women with pathogenic variants or variants of unknown significance in genes not known to substantially increase ovarian cancer risk reported getting oophorectomies.

Because genetics is not the only reason women decide to have these surgeries, researchers asked the patients about their personal and family history of ovarian cancer in a questionnaire. Most appeared not to be at increased risk for ovarian cancer based on these other factors, raising questions as to whether they should have had their ovaries removed, and suggesting that patients and their doctors do not fully understand the genetic test results.

"These data raise the concern that there are some women who are undergoing preventive oophorectomy, who are often under age 50 [and potentially premenopausal] without clear indications," Domchek said during her presentation at the meeting, adding that if these findings are validated, then it suggests the potential for harm.

The analysis drew on data collected within the Prospective Registry of Multiplex Testing (PROMPT). Back in 2014, when PROMPT launched as a collaboration between several cancer centers and genetic testing firms, next-generation sequencing panel tests were just starting to be rapidly adopted in cancer care, and patients were increasingly finding out they harbored variants in genes with unclear links to cancer. 

Researchers from Memorial Sloan Kettering Cancer Center, the Mayo Clinic, the Abramson Cancer Center at the University of Pennsylvania, and the Dana-Farber Cancer Institute started the registry, hoping to gather information on the penetrance of variants associated with cancer risk and variants of uncertain significance (VUS) detected by multi-gene panels. They collaborated with several genetic testing labs providing such tests — Ambry Genetics, Myriad Genetics, GeneDx, and Quest Diagnostics — which shared registry and enrollment information with patients who had a pathogenic variant or a VUS.

Since 2014, more than 7,600 patients, nearly 94 percent of them women, have enrolled in PROMPT and filled out baseline and annual follow-up questions. The internet registry, in which patients self-enrolled, asked participants in a standardized fashion what their surgical experience was and why they decided to undergo a surgical procedure, said Mark Robson from Memorial Sloan Kettering, who led the study with Domchek. "This is obviously all their own perception of it," he noted.

As Domchek reported at the meeting this week, 1,691 women reported having oophorectomies. Of these, 248 women were excluded because they reported having a pathogenic variant in a gene that is associated with more than 5 percent lifetime risk of ovarian cancer, including BRCA1/2, RAD51C, RAD51D, BRIP or Lynch syndrome genes. In these settings, there are guidelines recommending risk-reducing oophorectomy.

Men were excluded. Another 474 and 494 women were excluded for reporting they had oophorectomies due to benign disease or for the treatment of cancer, respectively. (Of those who reported having cancer, most indicated having ovarian cancer, but there were 66 women who reported having ovaries removed following breast cancer.)

Out of the 4,589 women who received oophorectomy and weren't excluded for these various reasons, Domchek and colleagues identified 457 patients, or 10 percent, who did not fall under expert guidelines for preventive surgery based on their ovarian cancer risk. Specifically, 114, or 16 percent, of 731 women with a VUS in a high-risk ovarian cancer gene reported oophorectomies. Expert guidelines recommend against making clinical decisions based on VUS.

The researchers also looked on oophorectomy rates among women with likely pathogenic and pathogenic variants, and VUS in three genes, CHEK2, ATM, and PALB2. These genes are commonly included in multi-gene panels but "do not have guidelines-based recommendations for oophorectomies," said Robson, a MSK oncologist with expertise in cancer genetic predisposition.

According to the National Comprehensive Cancer Network's guidelines, women with certain pathogenic mutations in CHEK2 are at heightened risk of breast cancer, often at younger ages, but mutations in this gene are not associated with ovarian cancer. Out of 427 women with CHEK2 mutations in the registry, 14 percent said they had oophorectomies, while nearly 12 percent of 246 patients with a VUS in CHEK2 reported having oophorectomies. 

Pathogenic mutations in ATM may "potentially" increase the risk of ovarian cancer, but the NCCN has found insufficient evidence to recommend risk-reducing surgery and advises doctors to manage patients based on their family history. Yet, 11 percent of 288 patients with pathogenic or likely pathogenic mutations in the gene in PROMPT said they got oophorectomies, and 10 percent of 408 women with a VUS in ATM reported having preventive surgery.

Some studies of PALB2 have suggested a small increase in ovarian cancer risk in those with pathogenic mutations but the risk appears to be modified by family history. Therefore, NCCN guidelines don't recommend oophorectomies based on PALB2 mutation status and tell doctors to manage patients based on family history. Of the 179 women with pathogenic mutations in this gene in PROMPT, 19 percent got oophorectomies, and nearly 15 percent of 179 women with PALB2 VUS got surgery.

"Recognizing the limitation of self-reported data, we wanted to be careful not to overinterpret that, because there are reasons why women might have such a surgery apart from the fact they had a mutation in one of these genes," Robson said. "For instance, somebody might have a strong family history of ovarian cancer, and then it wouldn't matter if they had a mutation in the gene, we'd still want to potentially consider preventive surgery."

If patients had breast cancer, then preventive ovary removal surgery could also be part of the interventions. Furthermore, if a woman is at high risk for breast or ovarian cancer and is past the childbearing age, 50 being a common surrogate for post-menopausal status, they might also want to have their ovaries removed to avoid cancer later in life.

"[A]t the end of the day … we found that among the women who had the surgery without any of the genetically-based indications, we couldn't find an alternative explanation," Robson said.

Within the group of 457 women who didn't meet guidelines for oophorectomies, 114 women had a VUS in a gene associated with a high risk of ovarian cancer; 220 women had had breast cancer but no VUS in ovarian cancer genes; and 123 women had no personal history of breast cancer and no VUS in ovarian cancer genes. Past studies have shown that despite expert guidelines advising against making clinical decisions based on VUS, patients and doctors still decide to have risk-reducing surgeries.

Based on a survey of more than 600 women with breast cancer who had undergone genetic testing, for example, Stanford University's Allison Kurian and colleagues reported in the Journal of Clinical Oncology in 2017 that just over half of women who were average risk and had a VUS said they had a bilateral mastectomy. 

In PROMPT, when researchers looked beyond VUS results to see if there was a plausible family history of ovarian cancer to support surgery, they found that around two-thirds of women with a VUS in an ovarian cancer-linked gene reported no family history of ovarian cancer. When they looked in the other categories — women with a personal history of breast cancer but no VUS in an ovarian cancer-linked gene and women without breast cancer and no VUS — 83 percent and 55 percent, respectively, reported no family history of ovarian cancer.

Moreover, across these three groups of women, 63 percent, 80 percent, and 66 percent, respectively, got oophorectomies when they were younger than 50 years old. "Breast cancer patients were more likely to have no family history of ovarian cancer and undergo oophorectomy at an early age," Domchek said in her presentation. The age-related finding, experts said, raises concerns that a substantial number of women had their ovaries removed before the natural onset of menopause, which could cause them to experience menopausal symptoms earlier, osteoporosis, and depression.

"A substantial number of women reported or understood that the reason for doing the surgery was either that there was a physician recommendation directly or presentation of [surgery] as an option," Robson said. Of the 311 women who disclosed in the registry the discussions they had with their providers, 44 percent said that their doctors recommended getting an oophorectomy, and 54 percent of those with a VUS in an ovarian cancer-associated gene said so. Another 206 women, or 66 percent, said that their doctors recommended or discussed oophorectomy has an option. 

The results of PROMPT "are provocative and suggestive of the idea that patients, and probably also their clinicians, do not probably fully understand the meaning of these test results," said Kurian who was not involved in the PROMPT analysis but is well known for her research exploring the impact that wider access to genetic testing is having on cancer patients' treatment and surgery decisions.

Kurian highlighted that the use of surveys to try to tease out the familial cancer history of patients who got oophorectomies, as well as their own understanding of why surgical decisions were made, were strengths of PROMPT. A weakness, she noted, is that patients self-selected to partake in the registry, which raises the question of how representative the study population is.

"We have surveys from the patients, but not the doctors," Kurian added. "These are patients' perceptions of what they were told … but surveys of doctors are also important to get the other side of the story."

In their 2017 JCO study, Kurian and colleagues also queried the women's surgeons. Among 377 doctors who responded, around half of the surgeons who saw a low number of breast cancer patients and a quarter of surgeons who treated a high volume of breast cancer patients said they would manage a woman with a VUS the same way they would manage a woman with a high-penetrance BRCA1 or BRCA2 mutation. Kurian noted that her group recently received funding to dig further into oncologists' rationale for preventive surgeries in these types of settings.

Discussing the results of PROMPT in an ASCO meeting presentation, Jennifer Klemp from the University of Kansas Cancer Center noted that the study may also be capturing the fact that some patients, particularly young women, may overestimate their cancer risk, and their desire to get preventive surgery may be so strong that their doctors cannot dissuade them from having it.

While acknowledging the limitations of self-reported data, Robson said that the findings in PROMPT are in line with the experiences of medical geneticists and genetic counselors, who are increasingly encountering patients who have received or been recommended preventive surgery based on VUS or mutations in genes not clearly association with a particular type of cancer. "Anyone who does cancer genetic counseling has had the experience of patients whose interpretation of their conversation with their provider is that surgery was either strongly or tacitly endorsed," Robson said.

Several cases in a recent series of case studies featured stories of patients who had gotten oophorectomies or were inappropriately recommended for preventive surgery based on a VUS (see here and here). 

Part of the problem is that as use of large NGS genetic testing panels has taken off, not just for guiding cancer treatment but also for gauging cancer risk in the broader population without cancer, oncologists are being asked to navigate complex, evidence-based guidelines around specific genes they're unfamiliar with. Beyond a handful of "famous" genes, such as BRCA1 and BRCA2, many oncologists may not know the nuances of the cancer risk data associated with the dozens or hundreds of other genes now included in NGS panels.

Even more confusing is the fact that experts bodies don't agree on how widely genetic risk testing should be offered. The American College of Medical Genetics and Genomics last year released a statement that all breast cancer patients should be evaluated for their inherited cancer risks based on NCCN guidelines, but did not advocate for genetically testing all breast cancer patients. In contrast, the American Society of Breast Surgeons recommends that all breast cancer patients receive genetic testing for BRCA1/2 and PALB2, and, if appropriate based on clinical factors and family history, for other genes as well. 

"When policymakers and guideline writers consider the wider deployment of multigene panel testing, including those who recommend testing broad, unselected populations of patients with [these tests,] it's incumbent on all of us to keep in mind that these [tests] are not without potential harms," Robson said. "And we, as a community, have to figure out how to mitigate those harms so we can maximize the benefits of the technology."

While the field is trying to improve physician education in this regard, Robson acknowledged it's a difficult task. He is hopeful that studies like PROMPT and others will further invigorate efforts to educate doctors and improve patients' access to genetic counselors and doctors with genetics expertise.

But he also noted that as genetics becomes a part of day-to-day oncology practice, it's important that providers keep up with evolving evidence and changing standards. "This study illustrates just how important that is," Robson said.