Introduction

Breast cancer is by far the most common cancer in women worldwide, with more than 1.6 million new cases diagnosed each year. Ovarian cancer is substantially less common, with 240 000 new cases each year, but with higher mortality.1 Most cases of breast and ovarian cancer are sporadic, but a minor fraction (2–8% and 8–15%, respectively) is caused by inheritance of pathogenic germline variants in BRCA1 or BRCA2, with variation in prevalence and relative contribution of BRCA1 and BRCA2 in different populations.2, 3, 4, 5, 6, 7, 8 It is important to identify these patients because the presence of such germline variants affects treatment, follow-up and further cancer prevention in patients with breast or ovarian cancer.9, 10 In addition, it may strongly influence upon their close relatives, as BRCA1/2 testing can identify healthy BRCA1/2 mutation carriers at high risk and thereby prevent cancer and cancer-related deaths through increased surveillance and prophylactic surgery.10, 11, 12, 13, 14, 15, 16

The most common current practice of BRCA1/2 testing is based on referral of suspected high-risk patients to clinical genetics services for specialized face-to-face genetic counselling. This procedure traditionally includes collection and confirmation of family history, risk assessment and eventually BRCA1/2 testing followed by a post-test counselling with dissemination of test results and advice concerning surveillance and follow-up.17, 18, 19 Based on family history, BRCA1/2-negative families with increased risk of familial breast cancer can also be identified.18, 20

However, this traditional approach is time consuming and resource demanding for both the patient and the health-care system, with an inherent risk of focusing too much on healthy relatives and not reaching all the cancer patients in question. Moreover, the discovery that BRCA1/2 status can inform treatment decisions in breast and ovarian cancer patients has led to an increased demand for BRCA1/2 testing at the time of cancer diagnosis.9, 21 New approaches to BRCA1/2 testing and genetic counselling may be needed to meet this situation. The aim of this project was therefore to assess the feasibility and impact of offering BRCA1/2 testing to all newly diagnosed patients with breast or ovarian cancer without prior face-to-face genetic counselling. We here report the uptake of BRCA1/2 testing, the incidence of pathogenic BRCA1/2 variants and the individual risk profiles among these unselected breast and ovarian cancer patients. As the psychosocial impact of such BRCA1/2 testing in newly diagnosed cancer patients without prior genetic counselling is scarcely described,22 we also examined the symptoms of anxiety and depression at inclusion and during the follow-up period of 6 months.

Patients and methods

Recruitment of patients

The patients were recruited from four hospitals in Western Norway (Haukeland University Hospital, Stavanger University Hospital, Haugesund Hospital and Førde Central Hospital), including three surgical departments and two gynecological departments, from September 2012 to April 2015. All patients with newly diagnosed breast or ovarian cancer were invited to participate in the study (for overview, see Figure 1). The patients received written information on the project and general information on hereditary breast and ovarian cancer, including the mode of inheritance and the potential consequences of a positive test results; such as the elevated cancer risk, recommended follow-up and risk-reducing strategies for the patient and healthy relatives. They were also informed that a positive test result could affect the surgical treatment of breast cancer patients, whereas specific information on novel therapies, like PARP-inhibitors, was not given. In addition, the patients had the opportunity to contact a genetic counselor on telephone for any further questions. All participants signed informed consent and filled in a structured questionnaire on personal and family medical history. The patients could choose BRCA1/2 testing with or without participating in an associated study of psychosocial aspects (see below). A blood sample was then collected at the local hospital and sent to a central laboratory for BRCA1/2 analysis. The study protocol was approved by the Regional Committee for Medical and Health Research Ethics (reference number REK Vest 2012-60).

Figure 1
figure 1

Flowchart showing the inclusion of patients and reporting of results in the DNA-BONus study.

DNA isolation, BRCA mutation analysis and clinical assessments

Genomic DNA was purified from EDTA-anticoagulated blood using the QiaSymphony instrument (Qiagen, Hilden, Germany). Genotyping of a panel of 20 pathogenic BRCA1 and 10 pathogenic BRCA2 variants that are recurrent in the Norwegian population was carried out using TaqMan Low-Density Arrays on the ABI 9700 instrument (Applied Biosystems, Foster City, CA, USA) as recommended by the manufacturer. An overview over the variants and sequences for the corresponding primers and probes is given in the Supplementary Table 1. In addition, the BRCA1 and BRCA2 genes were analyzed for deletions and insertions by Multiplex Ligation-dependent Probe Amplification (MLPA) technology (P002 BRCA1 and P045 BRCA2 MLPA probe mixes; MRC-Holland, Amsterdam, The Netherlands).

The result of the BRCA1/2 testing was given to the patient by a genetic counselor within 3 weeks after blood sample collection (Figure 1). In addition, the result was reported to the clinician who was responsible for treating the patient, to be filed in the patient’s medical record at the hospital. If the test result was negative and there was no increased familial cancer risk, the patient received the result by letter. Patients with a positive test result or with a personal or family history indicative of a high risk of hereditary cancer were contacted over the phone by a genetic counselor and were offered traditional face-to-face genetic counselling and further investigations in one of our outpatient clinics.

Based on collection of family history and confirmation of cancer diagnoses in relatives, selected patients were then offered extended genetic testing, with Sanger sequencing of all exons and flanking intron sequence in both BRCA1 and BRCA2. We used the following reference sequences: BRCA1: NG_005905.2 (gene), NM_007294.3 (mRNA), NP_009225.1 (protein); BRCA2: NG_012772.3 (gene), NM_000059.3 (mRNA) and NP_000050.2 (protein)).

To classify the sequence variants we followed the recommendations given by the International Agency for Research on Cancer (IARC).23 Pathogenic (class 5) and likely pathogenic (class 4) variants were regarded as positive genetic test results and have been submitted to the Leiden Open Variation Database (LOVD 3.0 shared installation; www.databases.lovd.nl/shared/genes). In this article we use the term BRCA1/2 mutation carrier for patients in whom a pathogenic or likely pathogenic variant was found.

All patients were categorized before BRCA1/2 testing depending on the presence of increased familial cancer risk or not. Increased risk was defined as personal at risk cancer history (eg, patients with young age at diagnosis, bilateral breast cancer or both breast and ovarian cancer) or positive family history (eg, close relative with breast cancer before 50 years of age or ovarian cancer at any age, two or more relatives with breast cancer or both breast and ovarian cancer in relatives) or a combination of personal at risk cancer history and positive family history, according to the current national clinical criteria for BRCA1/2 testing (see also legend to Table 1). The participants were in addition rated by the Manchester scoring system for BRCA1/2 testing.24, 25

Table 1 DNA-BONus study population

Psychological measurements

Participants who gave informed consent for the psychosocial part of the project were asked to fill in questionnaires at baseline when they were offered genetic testing (T1), at 1 week after disclosure of the BRCA1/2 test result (T2) and 6 months after disclosure of the BRCA1/2 test result (T3). In the present study, we have used data from the Hospital Anxiety and Depression Scale (HADS).26 HADS comprises two subscales for symptoms of anxiety and depression, respectively, each with 7 items to be scored on a 4-point (0–3) scale, giving a range of subscores from 0 to 21. The reliability of the HADS subscales in this study, as estimated with Cronbach’s α, had a range of 0.83–0.88 for HADS anxiety and 0.80–0.86 for HADS depression at the three assessments. Subscale scores of ≥8 were used as cutoff for defining higher, caseness-relevant levels of anxiety and depression.27

Statistical methods

Descriptive statistics were used for psychological and clinical variables, reporting the mean values, SD and range. To analyze the changes over time in HADS anxiety and depression scores, we used a paired sample t-test and McNemar’s exact test. Independent sample t-test was used to compare the means of two independent groups and χ2 test was used to analyze dichotomous variables for independent groups.

Missing values were replaced by the individual’s own average score for HADS if 60% or more of the items were filled in by the respondents. All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 22.0 (Armonk, NY, USA).

Results

A total of 1015 patients with either breast cancer (N=893) or ovarian cancer (N=122) were offered BRCA1/2 testing at the time of cancer diagnosis, of whom 405 (45.4%) of the breast cancer patients and 83 (68.0%) of the ovarian cancer patients completed the genetic testing. The mean age of the participants was 56.9 years (SD 12.4, range (min–max) 23–89) in the patients with breast cancer and 60.5 years (SD 11.9, range 24–88) in the patients with ovarian cancer (Table 1). Among the participants, 202 (49.9%) of the patients with breast cancer and 70 (84.3%) of the patients with ovarian cancer were eligible for BRCA1/2 testing according to current national clinical guidelines (Table 1). The median time from diagnosis to blood sampling was 34 days (mean 68, range 0–1402) and the median time from diagnosis to the patient received initial test result was 52 days (mean 87, range 12–1423) (data not shown). For 13 patients, the interval between diagnosis and blood sampling exceeded 1 year.

A pathogenic BRCA1/2 variant was identified in 7 (1.7%) of the 405 breast cancer patients (mean age of 50.6 (SD 15.8, range 32–76) years; Table 1), of whom 6 carried a BRCA1 and 1 a BRCA2 pathogenic variant (Table 2). Three BRCA1 and one BRCA2 mutation carriers had a breast cancer that was triple negative (Er-/Pr-/HER2-) and all seven breast cancers were HER2 negative (Table 2). Interestingly, as many as 19 (22.3%) of the 83 ovarian cancer patients (mean age 56.5 (SD 9.1, range 44–72) years) were BRCA1/2 mutation carriers (Table 1), including 15 with a pathogenic BRCA1 variant and 4 patients with a pathogenic BRCA2 variant (Table 2). Most ovarian cancers were serous carcinomas, apart from one poorly differentiated carcinoma and one endometroid adenocarcinoma. The majority of the mutation carriers (N=21; 80.8%) were identified by the standard test panel of recurrent mutations (Table 2), where 3 of the most frequent Norwegian pathogenic founder variants in BRCA1 (c.1556del, c.697_698del and c.3228_3229del) were detected in 15 (57.7%) of the mutation carriers. Four additional pathogenic variants were identified by Sanger sequencing of selected breast cancer (N=94) or ovarian cancer (N=31) patients with a particularly high risk of carrying a pathogenic BRCA1/2 variant, based on the personal and family history (see Table 2). During the first (years 2012–2013) and second (years 2014–2015) half of the DNA-BONus study period, 26.1% (55 out of 211) and 25.3% (70 out of 277) of the participants were selected for Sanger sequencing, respectively. Out of the total population of 488 patients, no one had BRCA1/2 alterations that could be detected by MLPA.

Table 2 Clinical and genetic data on the BRCA1/2 mutation carriers that were identified in the DNA-BONus study

Among the 272 patients fulfilling the current national criteria for diagnostic BRCA1/2 testing at inclusion, 6 out of 202 breast cancer patients (3.0%) and 18 out of 70 ovarian cancer patients (25.7%) were found to be mutation carriers (Table 1). Among 216 patients not meeting current clinical test criteria at inclusion, the corresponding numbers of BRCA1/2 mutation carriers were 1 of the 203 breast cancer patients (0.5%) and 1 of the 13 ovarian cancer patients (7.7%). However, it should be noted that the breast cancer patient with a pathogenic BRCA1 variant and the ovarian cancer patient with a pathogenic BRCA2 variant, who apparently had negative family histories upon inclusion, were both subsequently reclassified as having familial risk, based on extended pedigrees obtained through the genetic counselling (see below, Discussion section).

The mean combined Manchester score at inclusion was 8.9 (range 2–71) (data not shown), with 67 out of 488 patients (13.7%) having a score of ≥15 (Table 1). A pathogenic BRCA1/2 variant was found in 13 out of 67 patients (19.4%) with a score of ≥15 and in 13 out of 421 patients (3.1%) with a score <15 (Table 1; summarized numbers). Among the 26 BRCA1/2 mutation carriers, the mean combined Manchester score at inclusion was 19.5 (range 4–71) (Table 2; summarized numbers). After genetic counselling and collection of additional clinical information, including pathology reports, the scores could be recalculated for 25 of the 26 mutation carriers (Table 2). The mean combined score increased to 27.7 (range 14–81) (data not shown), with 24 mutation carriers having a score of ≥15, whereas the remaining mutation carrier had a score of 14 (Table 2).

All 26 BRCA1/2 mutation carriers accepted the offer of traditional face-to-face post-test genetic counselling. Among participants with a negative result on the initial BRCA1/2 panel and MLPA analysis, genetic counselling was offered for 188 patients (40.3% of total) with a personal at risk cancer history indicating further genetic testing (eg, young age at diagnosis or more than one primary cancer) or with a positive family history indicative of either familial breast cancer (eg, two or more breast cancer cases in first-degree relatives) or another hereditary cancer syndrome. The acceptance rate for genetic counselling in this group was 93.6% (N=176).

Because of the potential risk of imposing additional psychosocial burden by offering and performing BRCA1/2 testing in the newly diagnosed cancer patients, we measured the level of anxiety and depression scores before testing and at 1 week and 6 months after disclosure of the test result in a subset of participants (Table 3). Among these 215 patients, the median time from diagnosis to blood sampling was 32 days (mean 56, range 0–436) and median time from diagnosis to received result was 50 days (mean 75, range 12–456) (data not shown). The mean HADS subscale score for anxiety symptoms was 6.84 (SD 4.28) at baseline (ie, time of inclusion), with a significant decrease to 4.88 (SD 3.86) 6 months after disclosure of the BRCA1/2 test result (P<0.001). The percentage of patients with higher levels of anxiety symptoms, defined as scores ≥8, decreased significantly from inclusion (39.9%) to 1 week (23.6%, P<0.001) and 6 months (19.8%, P<0.001) after disclosure of the test result, respectively. During the observation period there was no significant change in depression symptoms, with a mean HADS score of 3.32 (SD 3.07) at baseline and 2.65 (SD 3.04) at 6 months. Approximately 10% of the patients showed higher levels of depression symptoms with a score of ≥8, both at baseline and follow-up measurements (Table 3). There were no significant differences in HADS scores between patients with breast (N=138) and ovarian (N=29) cancer, or between mutation carriers (N=8) and noncarriers (N=159) (data not shown).

Table 3 HADS anxiety and depression subscale scores at various time points for a subset of DNA-BONus participants

To explore the effect of time after diagnosis on the HADS scores, we divided the sample in two groups, with N=171 (83.0%) having less than and N=35 (17.0%) having more than 90 days from cancer diagnosis to blood sampling. There were no significant differences in HADS scores between the two groups (data not shown).

Compared with the participants who only agreed to genetic testing (mean age 61.6 years), the patients who also took part in the psychosocial study were significantly younger (P< 0.001), with a mean age of 56.2 years (data not shown). There were no significant differences between the two groups regarding educational level or type of cancer diagnosis (breast or ovarian).

Discussion

The main findings in this study are that: (1) most patients with newly diagnosed ovarian cancer accept germline BRCA1/2 testing, with significantly lower uptake among breast cancer patients; (2) there is a high prevalence of BRCA1/2 mutation carriers in the group of ovarian cancer patients; (3) all patients who were identified with a pathogenic BRCA1/2 variant fulfill our current clinical criteria for diagnostic BRCA1/2 testing; and (4) the level of anxiety and depression symptoms in the participants at inclusion was comparable to what can be found in cancer patients in general.28, 29

Ovarian and breast cancer patients with pathogenic BRCA1/2 variants are candidates for targeted drug therapy, such as PARP inhibitors.21 Recently, the US Food and Drug Administration (FDA) approved a PARP inhibitor for use in ovarian cancer (http://www.fda.gov/newsevents/newsroom/pressannouncements/ucm427554.htm). Our study shows that, even before such treatment options became available, BRCA1/2 testing was well accepted among newly diagnosed patients, with 68% participation rate among the women with ovarian cancer, whereas 45% of patients with breast cancer chose to undergo BRCA1/2 testing. There may be a selection bias among the participants because, on average, patients with breast cancer and ovarian cancer in our study were younger (mean age 56.9 and 60.5 years, respectively) as compared with patients with these cancers in the Norwegian population in general. According to national numbers, the mean age of all cases with breast cancer and ovarian cancer diagnosed between 2008 and –2012 was 61.5 and 65.4 years, respectively,30 thereby indicating that older patients may have declined participation in our study. This could be particularly relevant for breast cancer patients with low a priori risk of carrying a pathogenic BRCA1/2 variant. The assumption of a certain degree of risk-based selection in the uptake is further supported by the fact that among the participants, 50% of the patients with breast cancer and >80% of the patients with ovarian cancer were eligible for diagnostic BRCA1/2 testing according to the current clinical guidelines. For obvious reasons the uptake will be higher when the result of BRCA1/2 testing influences treatment options.21

In total, we identified 26 patients with a pathogenic BRCA1/2 variant and by that identified 22 new BRCA1/2 families. This finding supports a need for increased availability and use of such BRCA1/2 testing, as a supplement to the existing referral systems and service in cancer genetics. Our study also reports a high prevalence (22.3%) of pathogenic BRCA1/2 variants in ovarian cancer patients, substantially higher than reported by others.3, 4, 5 This may be caused by a high prevalence of pathogenic founder variants in our population, but surprisingly the prevalence among patients with breast cancer is rather low (1.7%) compared with international data.2, 6 The highest prevalence of BRCA1/2 mutation carriers has been reported in populations with very strong founder effects, and most studies on the frequency of pathogenic BRCA1/2 variants in patients with sporadic breast cancer have had some form of selection criteria, for example, young age at onset or triple-negative histology.2, 7, 8 In the DNA-BONus study, we offered genetic testing to all patients with newly diagnosed breast cancer that, in combination with a rather low prevalence of pathogenic BRCA2 variants in the Norwegian population,6 at least in part may explain the rather low frequency of pathogenic BRCA1/2 variants among our patients with breast cancer.

At inclusion, all but two of the 26 BRCA1/2 mutation carriers fulfilled the current clinical recommendations for diagnostic BRCA1/2 testing in Norway. One patient with breast cancer after the age of 75 years apparently had a negative family history according to the information forwarded at inclusion. However, further examination revealed that her sister died from ovarian cancer before the age of 50 years. The other patient was a woman with ovarian cancer after the age of 70 years. We were informed at inclusion that she had two first-degree relatives with abdominal cancer and cervical cancer, respectively, both after the age of 70 years. During the genetic counselling these diagnoses were both confirmed to be ovarian cancer cases. Thus, all mutation carriers in this study fulfilled current national criteria for diagnostic BRCA1/2 testing when a proper personal and family history had been taken.

The Manchester scoring system is a frequently used tool to identify individuals and families at high risk of having a pathogenic BRCA1/2 variant.24 In this study, we found that the Manchester scores obtained at inclusion were markedly lower than the real values (see below). In retrospect, all BRCA1/2 mutation carriers had combined Manchester scores at ≥14 points, demonstrating that the hereditary breast and ovarian cancer families identified through testing of patients with incidental breast or ovarian cancer do not differ significantly from families identified through the traditional route. These findings indicate that most BRCA1/2 mutation carriers can be identified through evidence-based clinical criteria, also within a group of incidental patients.

In order to identify patients at risk of non-BRCA1/2 familial breast cancer and other causes of hereditary cancer, we systematically collected structured family history from the participants before BRCA1/2 testing and employed a low threshold for our genetic counselors to contact the participants for additional information. Indeed, the importance of family history should not be neglected when the availability of BRCA1/2 testing increases and more patients with breast cancer are tested in routine clinical practice. Most familial breast cancer risk is not caused by pathogenic BRCA1/2 variants, and women belonging to BRCA1/2-negative breast cancer families are also at increased risk for breast cancer.20 The importance of obtaining a structured family history was illustrated by the fact that BRCA1/2 mutation carriers in our study scored significantly higher in the Manchester scoring system when taking into account the information collected during the genetic counselling procedure, as compared with the rating based on the initial self-reported information. In this regard, oncologists and surgeons may need additional support and training to extract a structured and relevant family history.31, 32

The traditional genetic counselling procedure has obvious benefits with respect to high-quality family history collection, and it has been shown to increase cancer-related knowledge and decrease distress in newly diagnosed cancer patients with an elevated risk of hereditary cancer.33 However, because this procedure is resource demanding, alternative approaches are needed when treatment-driven genetic testing is offered to larger patient groups with lower probability of carrying a pathogenic BRCA1/2 variant. Written, telephone-based or digital information provided by a clinical geneticist or genetic counselor, together with adequate information from the oncologist or surgeon, could be considered as an alternative for some patients.22 Patients at increased risk of psychosocial distress should have easy access to genetic counselling. An open telephone line to a genetic counselor might not be optimal for patients newly diagnosed with cancer, as we experienced that <20 patients actually contacted the genetic counselor for more information before testing throughout the whole DNA-BONus study period of two-and-a-half years. In order to discuss the consequences of the BRCA1/2 test results for the patient and other family members, as well as to explain complex test results and other hereditary causes of cancer, we also advise genetic counselling in case of a positive BRCA1/2 test result and in case of a personal or family history suggestive of hereditary cancer.

As the most common current practice of BRCA1/2 testing is based on referral of selected high-risk subjects to extensive face-to-face procedures of genetic counselling before BRCA1/2 testing,17, 18 we investigated whether our new simplified approach could lead to increased anxiety or depression in the newly diagnosed patients. Interestingly, the level of anxiety symptoms was comparable to those reported for patients with breast cancer and gynecological cancer in general,28, 29 but higher than normal population values.34 Approximately 40% of the patients had a HADS subscale score above the defined threshold for symptoms of anxiety27 at inclusion, and the level of anxiety decreased significantly during the 6-month follow-up period that also included the dissemination of the BRCA1/2 test result. The drop in the level of anxiety symptoms during the observation period may simply reflect the adjustment to the cancer diagnosis and treatment, and genetic testing in our study did not appear to influence on this expected drop.

There are some limitations to our study. Because of ethical regulations, we had no information about the patients who declined participation in the study. Another limitation is that Sanger sequencing of the BRCA1/2 genes was only performed on selected high-risk patients, implying that some of the lower-risk patients could be carriers of rare BRCA1/2 variants that were not covered by the BRCA1/2 panel test. In this respect, it should be noted that the methods and two-step procedure for BRCA1/2 testing (ie, multiplex panel test for recurrent variants, plus optional BRCA1 plus BRCA2 Sanger sequencing) remained unchanged during the whole inclusion period, and that the fraction of patients who were sequenced was almost the same in the first and second half of the DNA-BONus study. Another potential weakness is that patients with previously known pathogenic BRCA1/2 variants, who were diagnosed with cancer during the DNA-BONus study period, might have declined participation because of low relevance, thereby reducing the total count of BRCA1/2 mutation carriers among the participants. Finally, some of the psychosocial results are limited by a small number of participating BRCA1/2 mutation carriers and should therefore be interpreted with caution.

In conclusion, we show that BRCA1/2 mutation testing is well accepted among patients with newly diagnosed breast or ovarian cancer. We further conclude that current clinical guidelines are sufficient to identify the majority of the BRCA1/2 mutation carriers among patients with breast cancer. Because of the high prevalence of pathogenic BRCA1/2 variants, we recommend that all patients with epithelial ovarian cancer are offered germline BRCA1/2 testing, irrespective of age or family history of cancer.