Article Text
PDF
Abstract
Background Since 2015, Dutch guidelines have recommended BRCA1/2 pathogenic variant testing for all patients with epithelial ovarian cancer. Recently, recommendations shifted from germline testing to the tumor-first approach, in which tumor tissue is tested first, and subsequent germline testing is performed only in those with BRCA1/2 tumor pathogenic variants or a positive family history. Data on testing rates and on characteristics of patients missing out on testing remain scarce.
Objective To evaluate BRCA1/2 testing rates in patients with epithelial ovarian cancer and compare testing rates of germline testing (performed from 2015 until mid-2018) versus tumor-first testing (implemented mid-2018).
Methods A consecutive series of 250 patients diagnosed with epithelial ovarian cancer between 2016 and 2019 was included from the OncoLifeS data-biobank of the University Medical Center Groningen, the Netherlands. Testing rates were analyzed for the overall study population and for germline testing (period I) and tumor-first testing (period II) separately. Characteristics of tested and untested patients were compared and predictors for receiving testing were assessed with multivariable logistic regression.
Results Median age was 67.0 years (IQR 59.0–73.0) and 173 (69.2%) patients were diagnosed with high-grade serous carcinoma. Overall, 201 (80.4%) patients were tested. In period I, 137/171 (80.1%) patients were tested and in period II this was 64/79 (81.0%). Patients with non-high-grade serous carcinoma were significantly less likely to receive BRCA1/2 testing than patients with high-grade serous carcinoma (OR=0.23, 95% CI 0.11 to 0.46, p<0.001).
Conclusions The results show that BRCA1/2 testing rates are suboptimal and suggest that clinicians may not be choosing to test patients with epithelial ovarian cancer with non-high-grade serous ovarian carcinoma, although guidelines recommend BRCA1/2 testing in all patients with epithelial ovarian cancer. Suboptimal testing rates limit optimization of care for patients with epithelial ovarian cancer and counseling of potentially affected relatives.
- ovarian cancer
- BRCA1 protein
- BRCA2 protein
Data availability statement
Data may be obtained from a third party and are not publicly available. This study used data from the OncoLifeS databiobank. The authors do not have the authority to share these data, but researchers can apply to use OncoLifeS data. Information about access to OncoLifeS data is given on their website: www.umcgresearch.org/w/oncolifes.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, an indication of whether changes were made, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
BRCA1/2 pathogenic variant testing provides valuable information on poly-(ADP-ribose)-polymerase inhibitor (PARP) inhibitor treatment options for patients with epithelial ovarian cancer and potential prevention in family members and is therefore recommended by Dutch guidelines for all patients with epithelial ovarian cancer.
WHAT THIS STUDY ADDS
BRCA1/2 testing rates are suboptimal and our analysis showed that patients diagnosed with non-high-grade serous ovarian cancer are significantly less likely to receive testing than those with high-grade serous ovarian cancer.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Patients who do not receive the recommended testing may miss out on optimal treatment, and their family members potentially miss out on genetic counseling and effective preventive measures.
INTRODUCTION
Ovarian cancer is the most lethal gynecological malignancy.1 The lack of early symptoms and effective screening methods require alternative mitigation strategies to reduce its impact. Currently, detecting BRCA1/2 pathogenic variants is considered a key component in epithelial ovarian cancer management for two main reasons. First, it provides opportunities for breast cancer prevention or surveillance in patients with a BRCA1/2 germline pathogenic variant and for surveillance or risk-reducing prophylactic surgery in female relatives who are potentially at high lifetime risk of ovarian and breast cancer. Second, it identifies patients with BRCA1/2 pathogenic variants in the germline or in tumor cells only, who will benefit greatly from poly-(ADP-ribose)-polymerase inhibitor (PARPi) therapy.2–5
Since 2015, BRCA1/2 testing has been recommended by Dutch guidelines for all patients with epithelial ovarian cancer, regardless of age and histotype, and is covered by health insurances. Originally, this included counseling by clinical geneticists and then germline testing for all patients.6 With reported improved outcomes in PARPi-treated women with BRCA1/2 pathogenic variants, regardless of origin (germline or tumor), guidelines recently changed to recommend BRCA1/2 tumor testing and subsequent germline testing only in those with tumor pathogenic variants or a positive family history.7 This tumor-first approach reduces the number of patients being referred to clinical geneticists to only those with a BRCA1/2 tumor pathogenic variant: approximately 20% of all patients.8 9 This lessens the burden for patients, decreases costs of genetic counseling and tests, and potentially reduces socioeconomic driven inequalities previously reported in genetic testing.2 3
Previous research has reported poor BRCA1/2 testing rates,4 5 and information on patients missing out on BRCA1/2 testing remains scarce. However, insight into characteristics of untested patients may allow identification of barriers to testing, and thus can be of great value in addressing suboptimal testing. The purpose of this study was to evaluate BRCA1/2 testing rates in patients with epithelial ovarian cancer diagnosed between January 2016 and December 2019 in the Netherlands, and to compare germline testing rates with the testing rates of the recently implemented tumor-first approach. Moreover, this study aimed to provide insight into characteristics of patients who were less likely to receive the recommended BRCA1/2 testing.
METHODS
Participants and Data Collection
The University Medical Center Groningen is an academic, medical, tertiary referral center in the north of the Netherlands, covering an area with 3.4 million inhabitants. The medical center has been appointed an expertise center for familial breast and ovarian cancer by the European Reference Network.10 The current study included patients from the OncoLifeS data-biobank, which is embedded within the structure of the University Medical Center Groningen to prospectively collect real-world data on clinical and lifestyle factors of patients with cancer.11 The OncoLifeS data-biobank received approval by the medical center’s medical ethical committee. All patients provide written informed consent to the OncoLifeS data-biobank before their data can be used.
Patients with epithelial ovarian cancer of adult age (≥18 years) were included in a consecutive series from January 2016 to December 2019. Patients with borderline ovarian tumors were not included. Data on patient demographics were collected from self-report questionnaires completed shortly after diagnosis. Collection of clinical data was embedded in routine care, and data on BRCA1/2 tests were abstracted from patient files. Variable descriptions are included in Online supplemental box S1.
Supplemental material
Data Definition
On July 1, 2018, BRCA1/2 testing protocols in the University Medical Center Groningen changed from germline testing for all newly diagnosed patients with epithelial ovarian cancer to testing tumor tissue first, and subsequently referring only those with BRCA1/2 tumor pathogenic variants or a positive family history to clinical geneticists for a germline test. To investigate testing rates for each testing guideline separately, patients were divided into two groups based on their date of diagnosis. July 1, 2018, was used as a cut-off date to split the study population into two groups, referred to as period I, including patients diagnosed before July 1, 2018, and period II, including patients diagnosed on or after July 1, 2018. Patients diagnosed at the end of period I were likely to receive testing in period II according to the tumor-first approach. However, this intermediate group was included in period I based on date of diagnosis.
Data Analysis
Patient and disease characteristics were described for the total study population and by period of diagnosis (period I and period II). Continuous variables were presented as median (IQR), and categorical variables as frequency (percentage). Groups were compared using the Mann-Whitney U test, Χ2 test, or Fisher’s exact test.
The primary outcome was the percentage of patients with epithelial ovarian cancer with a ‘known BRCA1/2 germline pathogenic variant status’—that is, those who completed the testing pathway. A patient’s germline pathogenic variant status was considered known in two ways. First, a germline test can confirm the absence or presence of a BRCA1/2 germline pathogenic variant. Second, a tumor test can confirm the absence of any BRCA1/2 tumor pathogenic variants, which indicates the absence of germline pathogenic variants. In the case of a tumor pathogenic variant, a germline test is required to analyze whether the pathogenic variant is of germline or somatic origin. Patients with a known BRCA1/2 germline pathogenic variant status will hereafter be referred to as ‘tested’ and those with an unknown BRCA1/2 germline pathogenic variant status as ‘untested’.
The primary outcome was analyzed for the total study population and by period of diagnosis (period I and period II). Patient and disease characteristics were also summarized by testing status: tested versus untested patients, and these groups were compared using the Mann-Whitney U test, Χ2 test or Fisher’s exact test. Furthermore, patient and disease characteristics of tested and untested patients were compared by period of diagnosis (period I and period II), to investigate if potential differences existed for both the germline testing guideline (period I) and the tumor-first testing guideline (period II). Multivariable logistic regression analysis was performed to evaluate histotype, age at diagnosis, and period of diagnosis as predictors of receiving BRCA1/2 testing. All statistical analyses were performed using Statistical Package for the Social Sciences (SPSS), version 23 (IBM) and two-sided p values <0.05 were considered statistically significant.
RESULTS
Patient Characteristics
A total of 250 newly diagnosed patients with epithelial ovarian cancer were included, 171 patients diagnosed in period I and 79 patients diagnosed in period II (Table 1). Median age of all patients was 67.0 years (59.0–73.0). Most patients were diagnosed with high-grade serous carcinoma (n=173, 69.2%) and diagnosed in International Federation of Gynecology and Obstetrics (FIGO) stage III (n=132, 52.8%). The proportion of patients with a low educational level was significantly greater in period I than in period II (period I: n=56, 32.7%; period II: n=14, 17.7%; p=0.049). Other characteristics did not significantly differ between period I and period II.
Patient and disease characteristics at baseline for all patients and by time of diagnosis (period I and period II)
BRCA1/2 Testing Rates and Test Outcomes
Of all 250 patients, 201 (80.4%) patients were tested (Table 2). A total of 63.6% of all patients were tested within 1 year after diagnosis (Online supplemental table S1). Of the patients diagnosed in period I, 80.1% was tested, and in period II this was 81.0% (Table 2), showing similar figures in both periods. For three patients (1.2%) the tumor test was unsuccessful and there was no test result available. These three patients were not referred for germline testing for unknown reasons and their BRCA1/2 germline pathogenic variant status remains unidentified. One patient (0.4%) had a BRCA1 tumor pathogenic variant but decided not to proceed with genetic counseling. Therefore, this patient was omitted from further analyses.
BRCA1/2 testing rates for all patients and by period of diagnosis (period I and period II)
In period I, most patients received a germline test only (n=93, 54.4%). In period II, the proportion of patients receiving a germline test only decreased to 5.1% and most patients received a tumor test only (51.9%). The overall proportion of patients receiving a germline test decreased from 69.0% in period I (germline test only: 54.4%; tumor+germline test: 14.6%) to 29.2% in period II (germline test only: 5.1%; tumor+germline test: 24.1%).
BRCA1 pathogenic variants were detected in 8.0% of the tested patients (n=16), including 10 germline pathogenic variants (5.0%), and BRCA2 pathogenic variants were detected in 4.0% of the tested patients (n=8), including five germline pathogenic variants (2.5%) (Table 3). In period I, BRCA1/2 germline pathogenic variants were detected in 8.0% (n=11) of the tested patients and in period II this was 6.3% (n=4). In period II, somatic BRCA1/2 pathogenic variants were detected in an additional 7.8% of the tested patients with no germline pathogenic variants (n=5). There were no statistically significant differences in BRCA1/2 germline pathogenic variants detected in period I vs period II. All BRCA1/2 pathogenic variants were detected in patients with high-grade serous carcinoma (Online supplemental table S2).
BRCA1/2 test results for all tested patients and by period of diagnosis (period I and period II)
Comparing Characteristics by Testing Status
Tested patients were relatively younger than untested patients (median age: 66.0 years (59.0–72.0) vs 70.0 years (61.3–75.0), p=0.023: Table 4). Furthermore, testing rates were significantly higher in patients with high-grade serous carcinoma than in patients with non-high-grade serous carcinoma (n=150, 74.6%; n=48, 23.9%, respectively; p<0.001). Also, testing rates were significantly higher in patients diagnosed with an advanced FIGO stage (III/IV), as compared with testing rates of patients with an early FIGO stage (I/II) (n=155, 77.1%; n=45, 22.4%, respectively; p=0.018).
Patient and disease characteristics at baseline for all patients and by testing status (tested and untested)
Characteristics of tested and untested patients were compared per period of diagnosis to investigate whether the group differences existed regardless of the active guideline (Table 5). Untested patients were significantly older than tested patients in period I (median age: 70.0 years (66.0–75.0) vs 66.0 years (58.0–72.0), p=0.009). In period II, the median age of untested and tested patients was similar (untested: 65.0 years (56.0–78.0); tested: 66.0 years (59.0–72.0)). In both periods, testing rates were significantly higher in patients with high-grade serous carcinoma than in those with non-high-grade serous carcinoma (period I: n=103, 75.2%; n=31, 22.6%, respectively; p=0.019; period II: n=47, 73.4%; n=17, 26.6%, respectively; p=0.006).
Patient and disease characteristics at baseline for tested and untested patients by period of diagnosis (period I and period II)
Multivariable Regression Analysis
In multivariable regression analysis adjusted for period of diagnosis and histotype, increasing age was significantly associated with lower chances of receiving BRCA1/2 testing (OR=0.96, 95% CI 0.93 to 0.99, p=0.012; Online supplemental table S3). Furthermore, patients diagnosed with non-high-grade serous carcinoma were significantly less likely to receive BRCA1/2 testing than patients with high-grade serous carcinoma (OR=0.23, 95% CI 0.11 to 0.46, p<0.001, adjusted for age at diagnosis and period of diagnosis).
DISCUSSION
Summary of Main Results
In the current study, 80.4% of all patients received BRCA1/2 testing. The proportion of tested patients was similar for period I and II: 80.1% and 81.0%, respectively. Overall, older patients and patients with non-high-grade serous carcinoma were less likely to receive testing. Also, patients diagnosed at an early stage (FIGO I/II) were less frequently tested, which is likely to be a result of the greater proportion of non-high-grade serous carcinoma in this group.
Results in the Context of Published Literature
Overall, BRCA1/2 pathogenic variants were detected in 12.0% of all tested patients, which is relatively low compared with the prevalence of 16–19% reported by others.8 9 This lower percentage could partly be explained by an active performance of risk-reducing salpingo-oophorectomies in our region. The percentage of BRCA1/2 germline pathogenic variant carriers undergoing risk-reducing salpingo-oophorectomies increased from 81% to 95% in our medical center after stopping ovarian cancer screening in 2009, and unpublished data from our hospital indicate that this now reaches 99%.12 With more BRCA1/2 germline pathogenic variant carriers opting for risk-reducing salpingo-oophorectomies, the proportion of women with BRCA1/2 germline pathogenic variants among newly diagnosed patients is likely to be smaller.
The percentage of BRCA1/2 germline pathogenic variants detected by germline testing in period I did not differ significantly from those detected by the tumor-first approach in period II (8.0% vs 6.3%). This suggests that the effectiveness of both testing approaches is comparable regarding the proportion of germline pathogenic variants detected. However, changing testing strategies to the tumor-first approach greatly benefits both the medical centers and the patients. In this study, the proportion of patients receiving a germline test decreased from 69.0% in period I to 29.2% in period II. Considering the emotional stress, travel expenses, waiting time, and uncertainty that is related to germline testing, the tumor-first approach significantly alleviates the burden for a large proportion of patients and their relatives. These factors support the implementation of the tumor-first approach.
Suboptimal BRCA1/2 testing rates have been reported in previous studies, ranging from 10% for genetic testing13 14 to 80% for tumor testing.8 The relatively high tumor testing rates of almost 80% were reported by Vos et al,8 who performed an implementation study of the tumor-first approach. The pathologists involved were explicitly instructed to perform tumor-first testing for all patients, and the study reported an overall tumor testing rate of 77.6%. The real-world testing rates we reported after the implementation of the tumor-first approach were even higher: 81.0% of all patients was tested. With the University Medical Center Groningen being a recognized expert center for hereditary gynecological cancers, it is likely that our percentage of tested patients is relatively high compared with other medical centers. Nonetheless, our study suggests that up to almost 20% of the patients are not currently tested, and consequently a proportion of these women and their relatives may be missing out on optimal treatment and preventive strategies. Moreover, the gene panels used for tumor tests have been expanded during the past years by also including the moderate ovarian cancer risk genes RAD51C, RAD51D, BRIP1, and PALB2, and testing rates of these genes are also likely to be suboptimal.
To our knowledge, no research has been published investigating differences between tested and untested patients in the context of tumor-first testing. While the age of patients seemed to influence the likelihood of getting tested in period I by means of germline testing, the age of tested and untested patients did not differ in period II when tumor-first testing was performed. Nevertheless, testing rates were significantly lower in patients diagnosed with non-high-grade serous ovarian carcinoma in both periods of diagnosis, also in the currently active tumor-first approach. Since BRCA1/2 pathogenic variants are most frequently detected in high-grade serous carcinoma, we expect a relatively smaller proportion of positive tests in the untested patients compared with the proportion reported in tested patients.
Disparities in general cancer predisposition testing have been reported previously. Van der Giessen et al2 reported significant differences in educational level and migrant status between 700 newly referred counselees and the overall Dutch population. Our study did not report significant differences in educational level between tested and untested patients. However, the proportion of patients with a medium/higher education was notably greater among tested patients than among untested patients in period I with germline testing (44.5% vs 33.3%). The proportion of patients with a medium/higher education was more similar between tested and untested patients in period II with the tumor-first approach (54.7% vs 53.4%). This could indicate that educational level, like age, seems to affect the likelihood of receiving a germline test, but not the likelihood of receiving a tumor test.
Strengths and Weaknesses
Strengths of this study include the selection of a consecutive series of patients and the comparison of characteristics of tested versus untested patients in a real-world clinical setting. Differences in histotype and age were observed between tested and untested patients in the analysis combining period I and II. Differences in histotype were also present when analyzing period I and II separately. These analyses were limited by the relatively small sample sizes of the groups in each period (n=171 and n=79), with only 15 untested patients in period II.
Implications for Practice and Future Research
Some clinicians may not offer BRCA1/2 testing to all patients due to the assumption that BRCA1/2 pathogenic variants are ‘only’ detected in high-grade serous carcinoma and that BRCA1/2 germline carriers generally develop cancer at a younger age.15 16 While all pathogenic variants in this study were detected in high-grade serous ovarian carcinoma, a recent meta-analysis by Witjes et al15 indicated that BRCA1/2 somatic and germline pathogenic variants can also be found in other histotypes. In addition, restricting BRCA1/2 testing to histotypes in which pathogenic variants are most prevalent is highly dependent on the accuracy of histological classification and could lead to undesirable exclusion of misclassified patients. Furthermore, age of disease onset is reported to only slightly differ between BRCA1/2 germline carriers and non-carriers. Besides age and histotype, a family history of breast or ovarian cancer is also not a suitable criterion for selecting patients for testing as approximately 30% of patients with a BRCA1/2 germline pathogenic variant do not have a family history of breast or ovarian cancer.17 18 For these reasons, all patients with epithelial ovarian cancer should be offered a tumor test, regardless of histotype, age, and family history. Nonetheless, the current study shows that adherence to these guidelines is not yet optimal and future research should focus on increasing test uptake.
CONCLUSION
Our analyses show suboptimal BRCA1/2 testing rates in patients with epithelial ovarian cancer, even after implementation of the tumor-first approach. This can be attributed to suboptimal adherence of clinicians to guidelines recommending BRCA1/2 testing in all patients with epithelial ovarian cancer as clinicians may deliberately choose not to test patients with non-high-grade serous carcinoma. Our findings can be used to optimize the care of patients with epithelial ovarian cancer and counseling of potentially affected family members. Awareness should be raised among clinicians on current guidelines and ways to increase uptake should be further investigated.
Data availability statement
Data may be obtained from a third party and are not publicly available. This study used data from the OncoLifeS databiobank. The authors do not have the authority to share these data, but researchers can apply to use OncoLifeS data. Information about access to OncoLifeS data is given on their website: www.umcgresearch.org/w/oncolifes.
Ethics statements
Patient consent for publication
Ethics approval
This study used data from the OncoLifeS databiobank (date of registration: June 28, 2019; International Clinical Trial Registry Platform, identifier: NL7839), which has been approved by the local Institutional review board (UMCG METc2014/109). All adult patients with a diagnosis of cancer or with a genetically increased risk of cancer are included in the OncoLifeS databiobank after written informed consent. Participants gave informed consent to participate in the study before taking part.
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors Conceptualization: LL, MJEM, NA; GHdB; methodology: LL, MJEM, JB, AtE, GHdB; statistical analysis: LL, GHdB; writing – original draft: LL; writing – review and editing: MJEM, JB, AtE, LPVB, AHvdH, NA, GHdB; guarantor: GHdB.
Funding This study was supported by AstraZeneca and Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, New Jersey, USA, who are co-developing olaparib.
Competing interests At the time of conducting this research, NA was an employee of, and held stock in, AstraZeneca LP. GdB and LL were financially supported by a grant from AstraZeneca for the purpose of this research project.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.