Introduction

The incidence of ovarian cancer is approximately 3,600 cases annually in Poland [1]. It has been shown that ovarian cancer patients from Poland are characterized by a high proportion of a limited number of recurrent mutations in the BRCA1 gene [2,3,4,5,6,7]. In 2003, Menkiszak et al. observed that 13.5% of ovarian cancer patients in the West Pomerania region carry one of these three common founder mutations in BRCA1 (c.5266dupC, c.181T > G, and c.4035delA) gene [8]. A high proportion of carriers with a limited number of recurrent mutations have an impact on test costs reduction and testing effectiveness. Since 2003 some regional differences in BRCA1 and BRCA2 mutation frequency and spectrum have been reported [2,3,4,5,6,7,8]. In our latest study, conducted among consecutive patients with ovarian cancer from the Podkarpacie region (South-Eastern Poland), we observed only 6.3% BRCA1 or BRCA2 founder mutation carriers, and a slightly different spectrum of these mutations than in other regions of Poland [9].

The aim of this study was to define the prevalence and spectrum of BRCA1 and BRCA2 mutations in unselected ovarian cancer patients from the Region of Podkarpacie with the use of NGS, and establish an optimal algorithm for genetic testing of women diagnosed with ovarian cancer from this region.

Methods

Ovarian cancer cases were identified from patients treated at a clinical base in the Department of Obstetrics and Gynecology of Fryderyk Chopin University Hospital No 1 in Rzeszow, Poland between January 2013 and January 2017. All patients were inhabitants of the South-East region of Poland. The study group consisted of 158 consecutive, newly diagnosed cases of ovarian cancer after surgical treatment, unselected for age or family history. The mean age of diagnosis was 58.5 years (range 22–84 years). The reference pathologist reviewed a representative slide from each cancer to confirm the diagnosis. 20% of patients were diagnosed in the I and II clinical stages according to FIGO, and 80% in III and IV. 14.8% of ovarian cancers showed pathological grading G1, 11.1 - G2, and 74.1 – G3. A cancer family history was obtained during an appointment with a clinical geneticist.

DNA was isolated from 5 to 10 ml of blood. We performed BRCA1 and BRCA2 genes Next-Generation Sequencing (NGS) study in all cases.

NGS

BRCA1 and BRCA2 genes were screened using the SureMASTR BRCA Screen Kit from Agilent Technologies. The SureMASTR BRCA Screen (Agilent Technologies) analyses the full coding regions of these genes. In brief, 50ng (2.5ng/µl) of genomic DNA (isolated from a peripheral blood sample from each patient) was used to amplify the target genes in a single-tube multiplex reaction. The obtained amplicon libraries were purified and diluted before single-tube universal PCR reaction to tag all amplicons with specific p5 and p7 adaptors. Each tagged amplicon library was purified to remove small residual DNA fragments, and the DNA library concentration was quantified using a Pre-proof High Sensitivity Qubit quantification kit (Life Technologies). Equimolar quantities of the individually tagged amplicon libraries were pooled, and the final sequencing library was normalized to a concentration of 4nM. Sequencing was performed on a MiniSeq platform (Illumina) using the MiniSeq Mid Output Kit, 2 × 150 cycles, to obtain reads for both strands. All detected pathogenic mutations and variants of unknown significance (VUS) were validated using Sanger sequencing. The conventional Sanger sequencing was performed with the use of BigDye terminator sequencing kit v3.1 (Life Technologies) on the ABI Prism 3130 genetic analyzer (Life Technologies) according to the manufacturer’s protocol.

Bioinformatics analysis

Bioinformatics analysis was performed using the software MASTR Reporter (Agilent Technologies). This analysis included read alignment to the human reference genome (Genome Reference Consortium GRCh37), variant calling, and visualization of the sequence reads. Variants above 40X coverage depth and with a minimum variant allele frequency of 5% for germline analysis are displayed in the software. Further filtering was applied to select germline variants which had a minimum of 100X coverage depth. Each sample passed quality control analysis.

Variant filtering and classification

The clinical significance and the implications of variants were identified based on the annotations in public archives: ClinVar, Breast Cancer Information Core, Leiden Open Variation Database Online Mendelian Inheritance in Man (OMIM), Human Genome Variation Society (HGVS), and VarSome. Germline variants were classified according to the ACMG Standards and Guidelines for the Interpretation of Sequence Variants [10].

Results

A BRCA1 or BRCA2 causative variant was found in 18 of 158 (11.4%) unselected ovarian cancer cases. A BRCA1 mutation was detected in 11 (7.0%) patients. The c.5266dupC mutation was the most common, it was diagnosed in six patients, followed by the c.181T > G mutation observed in three patients, and the c.676delT mutation in one patient. All carriers of these 3 mutations were diagnosed previously with the test based on detection of founder pathogenic variants characteristic for the Polish population [9]. In addition, we diagnosed one BRCA1 c.5346G > A mutation carrier (Table 1) with NGS only. A BRCA2 gene mutation was diagnosed in 7 (4.4%) unselected ovarian cancer cases. None of these BRCA2 mutations was a recurrent mutation characteristic for the Polish population. We also found in 3 patients, variants of unknown significance (VUS), all in the BRCA2 gene (c.9302T > C, c.2063 A > G, and c.8527 A > G). The median age of diagnosis of the 18 hereditary ovarian cancers was 57.50 years (range 41–82 years), compared with a median age of diagnosis of 58.77 years (range 22–84 years) for the 140 cases without a mutation. However, the median age of diagnosis in BRCA1 carriers was lower than BRCA2 carriers − 55.8 vs. 60.1 years, respectively. A BRCA1 mutation was found in 3, and BRCA2 in 1 of 31 (together − 12.9%) women diagnosed with ovarian cancer at or under the age of 50 compared to 8 BRCA1 and 6 BRCA2 carriers of 127 (together − 11.0%) women diagnosed at a later age. Among the 18 women with ovarian cancer and a BRCA1 or BRCA2 mutation, ten reported a first- or second-degree relative with breast or ovarian cancer (55.5%), and there was only a slight difference between BRCA1 and BRCA2 carriers (54.5% vs. 57.1%, respectively). A mutation was present in 25.6% (10/39) of ovarian cancer patients with a positive family history and in 6.7% (8/119) of women with a negative family history. A significant family history, defined as a presence of first- or second-degree relative with breast or ovarian cancer, was observed in 24.7% (39/158) of patients with ovarian cancer (Table 2).

Table 1 The frequency of founder mutations in unselected series of 158 ovarian cancer patients.
Full size table
Table 2 Prevalence of BRCA1and BRCA2 mutations in ovarian cancer patients, by age of onset and family history.
Full size table

Discussion

The region of Podkarpacie is located in the South-East corner of Poland, bordering Ukraine and Slovakia. In our previous study, we identified 10 of 158 (6.3%) of unselected cases of ovarian cancer from this region carried one of 13 founder mutations in the BRCA1 or BRCA2 genes [9]. This is less than in other regions of Poland where the frequency of BRCA1 causative founder variants was observed in about 10-13.5% of ovarian cancer patients [2,3,4,5,6,7,8]. In this study, we performed the NGS study of BRCA1/2 genes in the same group of 158 women affected with ovarian cancer and diagnosed 18 (11.4%) BRCA1/2 mutation carriers. The frequency of individual BRCA1/2 mutations observed in ovarian cancer patients is shown in Table 1. All ten founder mutations were confirmed with our observations performed with a genetic test based on the BRCA1/2 founder mutations characteristic for the Polish population [9]. In addition, we diagnosed 1 carrier of BRCA1 and 7 carriers of the BRCA2 gene. These 8 mutations were detectable by whole sequencing only. Like in other regions of Poland, the most frequent mutation was the BRCA1 c.5266dupC mutation observed in 30% (6/18) of all carriers and the BRCA1 c.181T > G mutation found in 15% (3/18). Other 9 BRCA1/2 mutations were observed in single patients and are rare in the Polish population. In contrary to our former observation the frequency of BRCA1/2 mutation carriers in the group of ovarian cancer patients is only slightly lower than in other regions of Poland. However, we observed a significantly lower frequency of founder mutations, in particular, BRCA1 c.5266dupC and to a lesser extent BRCA1 c.181T > G. This phenomenon can be caused in general by the lower frequency of these mutations in South-East Poland. However, it should be noted that in this region the extensive genetic testing of BRCA1/2 genes has been carried out in patients with ovarian and breast cancer, as well as, in healthy patients since the year 2000. The testing focused mainly on the detection of founder mutations. As a result, several hundred families with BRCA1 founder mutations have been diagnosed so far and several hundred prophylactic adnexectomies have been performed. It should be taken into account that these BRCA1 founder mutations carriers were thus protected against ovarian cancer, and therefore, we observe their lower representation among BRCA1/2 mutation carriers who have now developed ovarian cancer.

The mean age at diagnosis in the 11 cases with BRCA1 mutation was 55.8 years, and of the 7 patients with BRCA2 mutation was 60.1 years. In both groups, the mean age at diagnosis was slightly higher than the observed in BRCA1/2 carriers from other regions of Poland [6,7,8]. Possibly there are lifestyle/environmental factors which may influence the later age of diagnosis in BRCA1/2 carriers. However, for non-carriers, the mean age at diagnosis was similar in the region of Podkarpacie and the rest of Poland (58.77 vs. 56.2–62.3 years) [6,7,8].

We observed strong family history in 10 of 18 (55.5%) mutation carriers, which is slightly more frequent than in other regions [6,7,8]. This applies to families with mutations in the BRCA1 as well as the BRCA2 gene. It can be explained by the relatively larger number of family members in an average family, in this region. However, the frequency of BRCA1/2 mutation carriers with negative family history is so high (44.5%), both in groups with recurrent founder mutations as well as with non-founder mutations, that we cannot recommend limiting performing the BRCA1 and BRCA2 gene testing based on NGS to cases with a burdened family history only. Also, the ovarian cancer age of onset is not a factor facilitating the qualification for this study. Taking into account our observations, it should be stated that performing a test based on the detection of Polish founder mutations in ovarian cancer patients from the Podkarpacie region is associated with relatively low sensitivity (55.5%). In turn, performing the NGS test in all subsequent patients with ovarian cancer is associated with a significant increase in costs. Application of NGS tests only in familial cases is associated also with low sensitivity of 55.5%. One of the compromise solutions would be to perform a standard genetic test based on the detection of founder mutations in all patients, and then if no mutation is detected, perform NGS in cases with a family history. With this algorithm of procedure, the sensitivity of detecting the BRCA1/2 gene mutation in ovarian cancer patients would increase to 77.8%, at a relatively low cost.

However, taking into account the current diagnostic standards of patients with ovarian cancer in the context of determining the optimal treatment and qualifying patients for treatment with PARP inhibitors, the BRCA1/2 gene is routinely tested using the NGS method in DNA extracted from tumour cells. It seems, therefore, that the most justifiable algorithm for detecting a germinal mutation in these patients is to start testing BRCA1/2 genes using the NGS method in DNA extracted from tumour cells. Then, if a mutation is found, the test should be performed in the patient’s peripheral blood to verify whether it is a germline or somatic mutation. If germline mutation is confirmed, the study should be extended to other relatives. However, if such a protocol is used, it should be taken into account that up to nearly 10% of cancer patients with negative NGS results performed in DNA isolated from neoplastic tissue cells, in fact, may carry germline mutation [11]. Lincoln et al. [11] indicated that major reasons for the mutations omissions include: (i) somatic variant interpretation guidelines differ from germline variant interpretation guidelines [12, 13]; (ii) high-quality germline tests can detect a broad spectrum of pathogenic variant types but present technical challenges particularly when analyzing archival specimen types often encountered in oncology e.g., paraffin-embedded formalin-fixed tissues [14,15,16,17], (iii) tumor tests may not include all genes of potential germline relevance in any given patient. Since the material for somatic mutation tests is mainly derived from tissues embedded in paraffin blocks, a significant percentage of the isolated DNA is of poor quality and unsuitable for NGS tests. It should be noted that patients with inconclusive tumor tests may also be carriers of germline mutations [18]. Therefore, the results of NGS tests carried out in somatic tissues should be interpreted with special care.

Conclusions

Approximately 11% of unselected ovarian cancer patients in the region of Podkarpacie carry a BRCA1 or BRCA2 causative variants. We found a significantly higher percentage of BRCA2 gene mutations, which are not detectable with a standard test for founder mutations. Diagnostics based only on testing the BRCA1/2 Polish founder mutations is characterized by relatively low sensitivity in the case of ovarian cancer patients from South-East Poland and should be supplemented by NGS study, in particular of the BRCA2 gene.