Recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients

Recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients Anna Nurmi 1 ,TaruA.Muranen 1 , Liisa M. Pelttari 1 , Johanna I. Kiiski 1 , Tuomas Heikkinen 1 , Sini Lehto 1 , Anne Kallioniemi 2 , Johanna Schleutker 3 , Ralf Bützow 1,4 , Carl Blomqvist 5 , Kristiina Aittomäki 6 and Heli Nevanlinna 1 1 Department of Obstetrics and Gynecology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland 2 BioMediTech Institute and Faculty of Medicine and Health Technology, Tampere University and Fimlab Laboratories, Tampere, Finland 3 Institute of Biomedicine, University of Turku, and Department of Medical Genetics, Genomics, Laboratory Division, Turku University Hospital, Turku, Finland 4 Department of Pathology and University of Helsinki and Helsinki University Hospital, Helsinki, Finland 5 Department of Oncology and University of Helsinki and Helsinki University Hospital, Helsinki, Finland 6 Department of Clinical Genetics, University of Helsinki, and HUSLAB, Helsinki University Hospital, Helsinki, Finland Mutations in BRCA1and BRCA2genes predispose to breast and ovarian cancer (BC/OC) with a high lifetime risk, whereas mutations in PALB2,CHEK2,ATM,FANCM,RAD51Cand RAD51Dgenes cause a moderately elevated risk. In the Finnish population, recurrent mutations have been identified in all of these genes, the latest being CHEK2c.319+2T>A and c.444+1G>A. By genotyping 3,156 cases and 2,089 controls, we estimated the frequencies of CHEK2c.319+2T>A and c.444+1G>A in Finnish BC patients. CHEK2c.319+2T>A was detected in 0.7% of the patients, and it was associated with a high risk of BC in the unselected patient group (OR = 5.40 [95%CI1.58–18.45], p=0.007) and similarly in the familial patient group. CHEK2c.444 +1G>A was identified in 0.1% of all patients. Additionally, we evaluated the combined prevalence of recurrent moderate-risk gene mutations in 2,487 BC patients, 556 OC patients and 261 BRCA1/2carriers from 109 families. The overall frequency of the mutations was 13.3%in1,141 BRCA1/2-negative familial BC patients, 7.5%in1,727 unselected BC patients and 7.2%in 556 unselected OC patients. At least one moderate-risk gene mutation was found in 12.5%ofBRCA1families and 7.1%ofBRCA1 index patients, as well as in 17.0%ofBRCA2families and 11.3%ofBRCA2index patients, and the mutations were associated with an additional risk in the BRCA1/2index patients (OR = 2.63 [1.15–5.48], p=0.011). These results support gene panel testing of even multiple members of BC families where several mutations may segregate in different individuals. Introduction Breast cancer (BC) is the most common cancer in women worldwide. 1 One of the most significant risk factors for BC is a family history of the disease so that the risk increases along with the number of affected first-degree relatives. 2 The major highrisk genes predisposing to BC and ovarian cancer (OC) are BRCA1 and BRCA2, with estimated cumulative risks of 72 and 69% for BC and 44 and 17% for OC, respectively, by the age of 80 years. 3 In Finland, BRCA1/2 mutations have been detected in 21% of BC families. 4 Recurrent BRCA1/2 founder mutations have been found in 1.8% of unselected BC patients, 5 in 26% of OC families, 6 and in 5.6% of unselected OC patients. 7 Several genes with a moderate effect on BC or OC risk have been identified in recent years. Mutations in PALB2 may cause a BC risk comparable to BRCA2, 8 whereas mutations in CHEK2, ATM and FANCM have been associated with a twoto threefold increased risk of BC compared to the general population, FANCM particularly with the risk of a triple-negative BC subtype. 9–15 RAD51C and RAD51D have been identified primarily as OC susceptibility genes, but have recently been linked also to triple-negative BC. 16–19 In addition, genome-wide association studies have reported numerous common risk-modifying alleles. 20 In the Finnish population, the history of geographic isolation and genetic bottlenecks has led to reduced variation and enrichment of deleterious alleles. 21 Consequently, just a few recurrent mutations cover the majority of all pathogenic mutations found in BC and OC susceptibility genes in Key words: breast cancer, ovarian cancer, moderate-risk gene, double heterozygote, CHEK2 Additional Supporting Information may be found in the online version of this article. Conflict of interest: LMP is currently employed by Blueprint Genetics. The other authors declare no conflict of interest. Grant sponsor: The Helsinki University Hospital Research Fund; Grant sponsor: The Sigrid Jusélius Foundation; Grant sponsor: The Cancer Foundation Finland DOI: 10.1002/ijc.32309 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. History: Received 11 Dec 2018; Accepted 18 Mar 2019; Online 30 Mar 2019. Correspondence to: Dr. Heli Nevanlinna, Department of Obstetrics and Gynecology, Helsinki University Hospital, Biomedicum Helsinki, P.O. Box 700, 00029 HUS, Finland, Tel.: +358-9-4717-1750, Fax: +358-9-4717-1751, E-mail: heli.nevanlinna@hus.fi International Journal of Cancer IJC Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics Finland. PALB2 c.1592delT has been observed with a 0.7–0.9% frequency among unselected BC patients and a 2.0–2.7% frequency among familial patients in studies from different regions in Finland. 22,23 CHEK2 c.1100delC has been found in 2.0% of unselected and in 5.5% of familial BC patients. 24 Three FANCM nonsense mutations, c.5101C>T, c.5791C>T and c.4025_4026delCT, were recently discovered in Finnish BC series. 14,15 Of these, c.5101C>T is the most common with a frequency of 2.8% in unselected patients, 3.1% in BC families and 5.6% in patients with triple-negative BC. 14 Two recurrent ATM mutations originally identified in Ataxia-telangiectasia families have been detected also in BC patients: c.6908dupA (previously marked as 6903insA) in 0.4% of unselected and in 0.6% of familial patients, and a pathogenic missense c.7570G>C in 0.2% of both unselected and familial patients. 25,26 Finally, two mutations in RAD51C, c.93delG and c.837+1G>A, and one in RAD51D, c.576+1G>A, were each observed in about 0.5% of unselected OC patients. 27,28 Besides these, a likely pathogenic RAD51C ex1–7 duplication was recently identified in 0.4% of unselected OC patients. 29 Two CHEK2 splicing mutations, an Eastern European founder mutation c.444+1G>A (alternatively IVS2+1G>A) and a novel c.319+2T>A, were recently discovered in Finnish BC patients. 29,30 Here, we estimate the frequencies of CHEK2 c.444 +1G>A and c.319+2T>A among Finnish BC and OC patients and evaluate the BC risk associated with c.319+2T>A. We report the frequencies of ATM c.6908dupA and c.7570G>C in a large series of BC and OC patients from Southern Finland. Furthermore, we define the overall prevalence of known recurrent mutations in PALB2,CHEK2,FANCM,ATM,RAD51C and RAD51D genes in Finnish BC and OC patients and study the cooccurrence of these mutations in BC patients and families, including families carrying BRCA1 or BRCA2 mutations. Materials and Methods Patients The patient series consisted of 3,156 female BC patients from the Helsinki and Tampere University Hospital regions in Southern Finland, as well as of 556 OC patients and 261 BRCA1/2 carriers from Helsinki. The patient series are described in detail below. The study was carried out with informed consent obtained from the patients and with approval by the Ethics committees of the Helsinki University Hospital and the Tampere University Hospital. Helsinki and Tampere BC series The unselected Helsinki BC series, unselected for age and family history, was collected prospectively at the Helsinki University Hospital Department of Oncology in 1997–1998 and 2000 5,31 and Department of Surgery in 2001–2004. 32 The series consisted of 884 and 986 consecutive, newly diagnosed BC patients (79 and 87% of all new cases during the collection periods, respectively). Of these, 1,727 patients with invasive disease were included in the analysis. The familial Helsinki BC series consisted of 1,141 index patients, including 381 familial cases collected among the unselected series and 760 additional familial patients collected at the Departments of Oncology and Clinical Genetics until 2015 as previously described. 24,32,33 Six hundred and nine families had at least three firstor second-degree relatives with BC or OC and 532 families had two affected first-degree relatives. BRCA1/2-positive patients had been excluded from the familial series. In total, 2,487 female BC patients from the Helsinki region were included in the study (Supporting Information Fig. S1). The unselected Tampere BC series was collected at the Tampere University Hospital as previously described. 5,32 The series consisted of 408 consecutive, newly diagnosed patients collected in 1997–1999 (75% of all new cases during the collection period), with additional 336 incident patients collected in 1996–2004. Of these, 669 female patients with invasive tumor were included in the study (Supporting Information Fig. S1). Two hundred and thirty-four of the patients had a positive family history with at least one firstor second-degree relative diagnosed with BC or OC. For the Helsinki and Tampere BC series, genomic DNA was isolated from peripheral blood samples. OC series The unselected Helsinki OC series was collected at the Helsinki University Hospital Department of Obstetrics and Gynecology. Two hundred and thirty-three patients with invasive disease were collected retrospectively during routine follow-up visits to the clinic in 1998 as previously described. 7 Additional patients were collected prospectively in 1998–2006. Altogether 556 cases were included in the series. For 433 patients genomic DNA was What’s new? In the Finnish population, a small number of recurrent mutations account for the majority of deleterious variations in breast and ovarian cancer susceptibility genes. The authors investigated 12 recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients. These mutations were found with similar frequency in unselected patients and BRCA1/2-positive familial index patients, twice as often as in population controls. In addition, a novel association was identified between breast cancer risk and CHEK2variant c.319+2T>A. The findings highlight the relevance of gene panel testing for breast and ovarian cancer risk assessment and its potential use for assessing breast cancer families with members who may carry different mutations. Nurmi et al.2693 Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics isolated from blood and for 123 patients the DNA sample was extracted from tumor tissue. BRCA1/2family series The BRCA1/2-positive series consisted of 126 BRCA1 mutation carriers (including four males) from 56 families and 135 BRCA2 mutation carriers (including 13 males) from 53 families. In 30 BRCA1 and 31 BRCA2 families, only the index patient was genotyped. All index patients had either BC or OC (including one male with BC) or both cancers. The index patients were initially identified from the familial or unselected Helsinki BC series as previously described, 4,5 or in diagnostic testing at the Helsinki University Hospital Department of Clinical Genetics. The BRCA1/2 mutations of the families are listed in the Supporting Information Table S1. Of the BRCA1 carriers, 21 individuals were affected with both BC and OC, 60 with BC, 13 with OC, 6 with other cancer and 26 individuals were healthy. Of the BRCA2 carriers, 10 had BC and OC, 70 had BC, 8 had OC, 18 had other cancer and 29 were healthy. Twentyfour BRCA1/2 carriers belonged also to the Helsinki unselected BC series and nine to the unselected OC series. For all patients, genomic DNA was extracted from peripheral blood. Population controls The geographically matched population controls consisted of 1,273 healthy female blood donors from Helsinki and 816 from Tampere regions. The Helsinki controls were collected in 2002–2003. Genotyping PALB2 c.1592delT, CHEK2 c.1100delC, FANCM c.5101C>T and c.5791C>T, RAD51C c.93delG and c.837+1G>A and RAD51D c.576+1G>A had been previously genotyped in the majority of the samples from the Helsinki region. 14,15,23,27,28,34 FANCM c.4025_4026delCT had been genotyped in one-third of the Helsinki BC cases. 15 Additional cases and controls were genotyped with TaqMan real-time PCR or with Sanger sequencing to obtain comparable sample counts for all mutations (Supporting Information Table S2). All study subjects from Helsinki and Tampere were genotyped for ATM c.6908dupA and c.7570G>C with TaqMan real-time PCR or with Sanger sequencing, for CHEK2 c.319+2T>A with Sequenom MassARRAY or with Sanger sequencing and for CHEK2 c.444+1G>A with Sanger sequencing (see Supporting Information Methods and Supporting Information Table S3 for details). Statistical analyses The statistical analyses were performed using R (version 3.4.4) environment for statistical computing (https://www.r-project. org). Associations were tested with Fisher’s exact test. Combined analysis of the Helsinki and Tampere BC series was performed with region-stratified logistic regression. All pvalues are two-sided. Further details are provided in Supporting Information Methods. Results Frequencies of CHEK2c.319+2T>A and c.444+1G>A CHEK2 c.319+2T>A was identified in 13/1725 (0.8%) unselected BC patients and in 11/1138 (1.0%) familial cases from the Helsinki region in comparison to 1/1271 (0.1%) population controls (Table 1). In the study subjects from the Tampere region, c.319 +2T>A was observed in 5/653 (0.8%) unselected BC patients and in 2/806 (0.2%) population controls. In the region-stratified analysis combining the Helsinki and Tampere BC series, c.319+2T>A was significantly associated with increased risk of BC in the unselected patient group (odds ratio (OR) = 5.40 [95% confidence interval (CI) 1.58–18.45], p= 0.007) and in the familial patient group (OR = 6.04 [1.65–22.10], p= 0.007; Table 2). Of the mutation carriers with available histological data, 19/22 (86.4%) had estrogen receptor (ER)-positive BC. The average diagnosis age was 54.3 years (range 33–81) for CHEK2 c.319+2T>A carriers and 56.5 years (range 21–95) for all BC patients in the combined dataset. In the Helsinki series, 3/18 (16.7%) carriers had bilateral BC and two carriers had breast and other cancer (cervical cancer or basalioma). Three BC patients (3/3147, 0.1%), but none of the controls, were found to carry germline CHEK2 c.444+1G>A (Table 1). The carriers were diagnosed with BC at the ages of 40, 65 and 79 years. One of the patients had bilateral BC, while another patient had both BC and OC. No CHEK2 c.319+2T>A or c.444+1G>A carriers were detected in the unselected OC series. Frequencies of ATM c.6908dupA and c.7570G>C ATM c.6908dupA was identified in two BC patients from the Helsinki region, both being familial index cases (2/1140, 0.2%), and in 2/1271 (0.2%) controls (Table 1). However, the two mutation-positive controls were young at the time of entering the study (ages 19 and 27). In the Tampere series, 5/666 (0.8%) unselected BC patients, but none of the 812 controls, carried the c.6908dupA mutation. The average age of BC diagnosis was 53.7 years (range 30–81) among the carriers. One of the carriers was later diagnosed with bladder cancer. In the Helsinki BC series, ATM c.7570G>C was identified in 2/1727 (0.1%) unselected patients and in 3/1141 (0.3%) familial patients, but in none of the controls (Table 1). In the Tampere BC series, c.7570G>C was not identified among the patients, but it was detected in one out of 815 (0.1%) population controls. The average age of BC diagnosis among the c.7570G>C carriers was 50.3 years (range 28–61). One of the carriers had bilateral BC, one had breast and thyroid cancer and one had BC and basalioma. In the per-gene analysis of the Helsinki and Tampere BC series, combining ATM c.6908dupA and c.7570G>C, the frequency of the mutations was twoto threefold higher in the unselected group (OR = 2.63 [0.69–10.01], p=0.156)andinthe familial group (OR = 2.97 [0.71–12.37], p= 0.136; Table 2) than in the population controls. Ten out of 12 (83.3%) carriers had ER-positive tumors. In the unselected OC series, the c.6908dupA mutation was detected in 2/549 (0.4%) patients (OR = 2.32 2694 Recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics [0.17–32.07], p= 0.589, compared to the Helsinki population controls). No c.7570G>C carriers were identified among the unselected OC cases. Overall frequencies of the moderate-risk mutations in the patients The combined mutation frequency in the moderate-penetrance genes was determined in the Helsinki BC and OC series. At least one of the studied mutations in PALB2,CHEK2,ATM,FANCM, RAD51C or RAD51D genes was detected in 243/2487 (9.8%) BC patients (Table 3). In the unselected BC series, at least one mutation was present in 129/1727 (7.5%) patients, whereas in the familial BC series in 152/1141 (13.3%) patients. The most common mutations among the familial index patients were CHEK2 c.1100delC detected in 67/1141 (5.9%) cases, FANCM c.5101C>T in 36/1141 (3.2%) cases and PALB2 c.1592delT in 27/1141 (2.4%) cases. In the unselected OC series, at least one studied mutation was detected in 40/556 (7.2%) patients. When only germline DNA samples were included, at least one mutation was found in 29/433 (6.7%) patients. FANCM was the most commonly mutated gene among the OC patients with 13 patients carrying the c.5101C>T and four carrying the c.5791C>T mutation (total 17/556, 3.1%). Mutations in the established OC genes RAD51C and RAD51D were identified in total in 9/556 (1.6%) unselected OC patients. Table 1. Frequencies of the CHEK2 and ATM mutations in the Helsinki and Tampere breast cancer series Helsinki 1 Tampere 2 Mutation Study cohort Carriers/total % Carriers/total % CHEK2 c.319+2T>A, rs587782401 Controls 1/1271 0.1 2/806 0.2 All BC 18/2484 0.7 5/653 0.8 Familial BC 11/1138 1.0 1/227 0.4 Unselected BC 13/1725 0.8 5/653 0.8 ER+ 16/1882 0.9 3/499 0.6 ER−2/451 0.4 1/122 0.8 CHEK2 c.444+1G>A, rs121908698 Controls 0/1272 0 0/786 0 All BC 2/2485 0.1 1/662 0.2 Familial BC 1/1140 0.1 0/230 0 Unselected BC 1/1725 0.1 1/662 0.2 ER+ 1/1881 0.1 0/507 0 ER−0/452 0 1/123 0.8 ATM c.6908dupA, p.Glu2304GlyfsTer69, rs773570504 Controls 2/1271 0.2 0/812 0 All BC 2/2486 0.1 5/666 0.8 Familial BC 2/1140 0.2 1/232 0.4 Unselected BC 1/1726 0.1 5/666 0.8 ER+ 2/1882 0.1 4/510 0.8 ER−0/452 0 1/124 0.8 ATM c.7570G>C, p.Ala2524Pro, rs769142993 Controls 0/1272 0 1/815 0.1 All BC 5/2487 0.2 0/669 0 Familial BC 3/1141 0.3 0/234 0 Unselected BC 2/1727 0.1 0/669 0 ER+ 4/1883 0.2 0/513 0 ER−1/452 0.2 0/124 0 ATM c.6908dupA and ATM c.7570G>C combined Controls 2/1271 0.2 1/811 0.1 All BC 7/2486 0.3 5/666 0.8 Familial BC 5/1140 0.4 1/232 0.4 Unselected BC 3/1726 0.2 5/666 0.8 ER+ 6/1882 0.3 4/510 0.8 ER−1/452 0.2 1/124 0.8 1 The cohorts are overlapping with 381 individuals included both in the familial and in the unselected patient group in the Helsinki breast cancer series. 2 The Tampere unselected patient group includes the familial patient group. Nurmi et al.2695 Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics Of the population controls, 47/1273 (3.7%) carried at least one moderate-risk mutation. Double heterozygosity for mutations in the moderatepenetrance genes In the Helsinki BC series, 11 double heterozygotes were identified for the studied PALB2,CHEK2,ATM or FANCM mutations among all 2,487 patients, so that 7/1141 (0.6%) familial and 7/1727 (0.4%) unselected patients carried mutations in more than one gene. Three out of the 11 double heterozygotes were included both in the familial and in the unselected series. Of the mutation-positive patients in the series, 4.6% familial and 5.4% unselected cases carried more than one mutation. All but one identified double heterozygotes carried either CHEK2 c.1100delC or FANCM c.5101C>T, or both of the two mutations (Supporting Information Table S4). Two of the double heterozygotes had bilateral BC, and one had breast and bladder cancer. None of the double heterozygotes in the BC series had RAD51C or RAD51D mutations or were affected with OC. However, one double heterozygote for ATM c.6908dupA and FANCM c.5101C>T was identified among the unselected OC patients (1/556, 0.2%). One healthy control, enrolled at the age of 37 years, was identified as a heterozygote for PALB2 c.1592delT and FANCM c.5101C>T out of 1,273 (0.1%) population controls. The expected number of double heterozygotes based on the mutation frequencies in the unselected series of BC cases was 4.4, less than the observed seven. In the population controls, the expected number of double heterozygotes was 0.7, when one was observed. We saw a statistically significant difference between the frequency of the double heterozygotes in the 7/1141 familial BC patients (OR = 7.85 [1.01–353.57], p= 0.030), but not in the 7/1727 unselected BC patients (OR = 5.17 [0.66–233.26], p= 0.149), when compared to the 1/1273 population controls. Moderate-risk gene mutations in BRCA1/2families We tested a total of 261 BRCA1 and BRCA2 carriers from 109 families for the recurrent mutations in the moderatepenetrance genes. Four of the 56 (7.1%) BRCA1 index patients and 6/53 (11.3%) BRCA2 index patients had at least one of the tested PALB2,CHEK2,ATM,FANCM or RAD51C mutations (Table 4). Two index patients carried three mutations: one OC patient had BRCA1 c.3485delA, ATM c.6908dupA and RAD51C c.93delG, and one BC patient had BRCA2 c.7480C>T, CHEK2 c.1100delC and FANCM c.5101C>T. However, when considering all genotyped BRCA1/2 carriers, including additional family members from 48 of the 109 families, at least one moderate-risk mutation was detected in 7/56 (12.5%) BRCA1 families and in 9/53 (17.0%) BRCA2 families. More than one carrier of a moderate-risk mutation was detected in four families: two BRCA1 families carrying ATM c.6908dupA, one BRCA2 family with CHEK2 c.1100delC and one BRCA2 family with FANCM c.5101C>T. The frequency of the moderate-penetrance gene mutations in BRCA1/2 index patients was comparable to the frequency in the unselected BC patients, and significantly different from the Table 2. Combined analysis of the CHEK2 and ATM mutations in the Helsinki and Tampere breast cancer series Mutation Study cohort OR 95% CI p-value CHEK2 c.319+2T>A, rs587782401 All BC 5.32 1.58–17.97 0.007 Familial BC 6.04 1.65–22.10 0.007 Unselected BC 5.40 1.58–18.45 0.007 ER+ 5.52 1.61–18.93 0.007 ER−4.34 0.84–22.38 0.080 ATM c.6908dupA, p.Glu2304GlyfsTer69, rs773570504 All BC 3.06 0.62–15.11 0.169 Familial BC 2.12 0.34–13.36 0.422 Unselected BC 3.08 0.62–15.42 0.171 ER+ 3.27 0.64–16.65 0.154 ER−1.81 0.16–20.79 0.633 ATM c.7570G>C, p.Ala2524Pro, rs769142993 All BC 3.07 0.35–26.84 0.311 Familial BC 4.71 0.46–48.39 0.192 Unselected BC 1.76 0.16–19.71 0.647 ER+ 3.32 0.36–30.56 0.289 ER−4.31 0.25–73.49 0.313 ATM c.6908dupA and ATM c.7570G>C combined All BC 3.07 0.85–11.09 0.088 Familial BC 2.97 0.71–12.37 0.136 Unselected BC 2.63 0.69–10.01 0.156 ER+ 3.29 0.88–12.23 0.076 ER−2.58 0.42–16.05 0.309 2696 Recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics frequency in the population controls (OR = 2.63 [1.15–5.48], p= 0.011). Overall, our data comply with the multiplicative risk model, which suggests that the risk effects associated with single mutations combine multiplicatively and that the second mutation confers additional increase in risk, which is comparable in magnitude to its nominal risk effect (Table 5). The tumor characteristics of all double heterozygotes with BC or OC are presented in the Supporting Information Table S4. Discussion In our study, we evaluated the overall prevalence of 12 recurrent moderate-risk mutations in Finnish BC and OC patients. These mutations in PALB2,CHEK2,ATM,FANCM,RAD51C and RAD51D genes were observed in 13.3% of the BRCA1/2negative familial BC patients, in 7.5% of the unselected BC patients and in 7.2% of the unselected OC patients. The studied mutations are likely to cover the vast majority of all pathogenic mutations in these genes in the Finnish population. We have not detected any other pathogenic mutations in PALB2, CHEK2,ATM,FANCM,RAD51C or RAD51D among 231 BC or OC patients from 174 families in whole-genome, wholeexome or gene-panel sequencing 14,29 (unpublished data). To the best of our knowledge, no other moderate-risk mutations in these genes have been reported in Finnish BC or OC patients. Furthermore, no other loss-of-function mutations have been identified among over 12,000 Finns included in the GnomAD database in RAD51D gene and only a few very rare or single variants in the other genes. 35 Even if some rare mutations may remain undetected, our results represent the minimum overall frequencies of the moderate-risk mutations in these genes in the BC and OC patients in Finland. Recently, Couch et al. 36 reported the overall frequency of pathogenic variants in 21 BC or OC risk and candidate genes to be 10.2% in a Caucasian BC cohort enriched with familial cases. However, when excluding BRCA1/2 carriers, the combined frequency was 6.2%. 36 We observed the 12 recurrent moderate-risk mutations over twice as often in the BRCA1/2negative familial index cases and the frequency of the mutations was 7.5% even in the unselected series of BC patients. Thus, in a population with a history of isolation, the effect of just a dozen founder mutations could be more substantial than the effect of multiple rare variants in many genes in outbred populations. Hence, the national guidelines of genetic counseling and personal risk stratification could be designed considering the mutation spectrum present in the population. The polygenic risk model for BC suggests that the effects associated with genetic variants combine multiplicatively so that the variants retain their nominal risk effects in the combinations. 37 This has been validated for common low-penetrance variants and for common variants in combination with rare mutations. 38–40 However, as indicated by Sokolenko et al., 41 only few studies have been published on BC patients heterozygous for highand moderate-risk mutations. In the Helsinki Table 3. Frequencies of the moderate-risk mutations in the Helsinki breast cancer series and ovarian cancer series All BC 1 Familial BC Unselected BC Unselected OC Controls Average Mutation Carriers/total % Carriers/total % Carriers/total % Carriers/total % Carriers/total % BC Diagnosis age 2 ATM c.6908dupA 2/2486 0.08 2/1140 0.18 1/1726 0.06 2/549 0.36 2/1271 0.16 41.8 (30.3–53.2) ATM c.7570G>C 5/2487 0.20 3/1141 0.26 2/1727 0.12 0/556 0 0/1272 0 50.3 (28.0–61.1) CHEK2 c.319+2T>A 18/2484 0.72 11/1138 0.97 13/1725 0.75 0/546 0 1/1271 0.08 55.2 (34.2–76.2) CHEK2 c.444+1G>A 2/2485 0.08 1/1140 0.09 1/1725 0.06 0/534 0 0/1272 0 59.5 (39.8–79.2) CHEK2 c.1100delC 99/2487 3.98 67/1141 5.87 49/1727 2.84 10/554 1.81 18/1261 1.43 52.8 (23.3–87.5) FANCM c.4025_4026delCT 1/2487 0.04 1/1141 0.09 0/1727 0 0/556 0 1/1272 0.08 54.7 FANCM c.5101C>T 73/2487 2.94 36/1141 3.16 48/1727 2.78 13/552 2.36 18/1270 1.42 54.2 (31.0–87.5) FANCM c.5791C>T 12/2487 0.48 5/1141 0.44 8/1727 0.46 4/553 0.72 3/1257 0.24 54.6 (28.0–79.9) PALB2 c.1592delT 36/2487 1.45 27/1141 2.37 13/1727 0.75 3/556 0.54 2/1273 0.16 51.6 (33.3–79.8) RAD51C c.93delG 2/2487 0.08 2/1141 0.18 0/1727 0 4/556 0.72 2/1269 0.16 52.3 (49.9–54.6) RAD51C c.837+1G>A 1/2487 0.04 1/1141 0.09 0/1727 0 2/555 0.36 0/1268 0 45.0 RAD51D c.576+1G>A 3/2487 0.12 3/1141 0.26 1/1727 0.06 3/556 0.54 1/1273 0.08 43.7 (30.8–53.8) Total 3 243/2487 9.77 152/1141 13.32 129/1727 7.47 40/556 7.19 47/1273 3.69 1 All BC patients from the familial BC and the unselected BC patient groups after removing the overlap of 381 individuals in these groups. 2 In the Helsinki breast cancer series. 3 Double heterozygotes are included only once in the total frequencies. Nurmi et al.2697 Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics BC series, 0.4% of unselected and 0.6% of BRCA1/2-negative familial patients carried a moderate-risk mutation in more than one gene. The frequency of the double heterozygotes was five times higher in the unselected BC cases (p= 0.149) and over seven times higher in the familial BC cases (p= 0.030) than in the population controls. The studied moderate-risk mutations were detected also in 9.2% of the BRCA1/2-positive index patients, more than twice as often as in the population controls (p= 0.011; Table 5) and in 14.7% of the BRCA1/2 families. These findings comply with the multiplicative risk model, where the second mutation confers additional risk. If it did not, the mutation frequencies would be approximately the same in the carrier patients as in the population controls. However, the exact risk associated with any specific mutation combination remains to be evaluated in larger patient groups. Earlier studies have indicated that CHEK2 or ATM mutations do not cause additional risk of BC in BRCA1/2-positive patients. 42,43 Here, FANCM variants appeared as the most common second mutations with 5.5% frequency in the BRCA1/2 index cases. However, even when excluding FANCM mutations from the double heterozygote analyses, the ORs were similar, even though statistically less significant (Supporting Information Table S5). It is also important to notice that in the BRCA1/2 families, the moderate-risk gene mutation was not always observed in the proband, but in another BRCA1/2 carrier in the family. Even though we pooled the mutations for the double heterozygote analyses, we appreciate that the moderatepenetrance genes have different risk effects and that these effects can be subtype-specific. CHEK2 mutations predispose especially to ER-positive disease 11 and are, like ATM mutations, associated with approximately twoto threefold increased risk of BC. 9–13 However, for patients with a family history of the disease, the risk estimates are often further increased, which is likely due to other risk-increasing variants in the families. The estimated lifetime risk of BC for CHEK2 mutation carriers ranges from about 20% for women without affected relatives to up to approximately 40% for women with positive family history. 9–11 PALB2 mutations are associated with about sixfold increased risk, and the estimated cumulative risk of BC caused by PALB2 mutations is 35% by the age of 70 years, but even 58% for carriers with a strong family Table 4. Frequencies of the moderate-risk mutations in the BRCA1/2 index patients and families Mutation 1 BRCA1index patients (%), n=56 BRCA1 families (%), n=56 BRCA2index patients (%), n=53 BRCA2 families (%), n=53 BRCA1/2index patients (%), n=109 BRCA1/2 families (%), n=109 All BRCA1/2 patients 2 (%), n=206 Healthy BRCA1/2 carriers (%), n=55 ATM c.6908dupA 2 (3.6) 2 (3.6) 0 0 2 (1.8) 2 (1.8) 3 (1.5) 3 (5.5) ATM c.7570G>C 0 0 0 1 (1.9) 0 1 (0.9) 1 (0.5) 0 CHEK2 c.1100delC 0 1 (1.8) 1 (1.9) 3 (5.7) 1 (0.9) 4 (3.7) 4 (1.9) 2 (3.6) FANCM c.5101C>T 1 (1.8) 3 (5.4) 3 (5.7) 3 (5.7) 4 (3.7) 6 (5.5) 7 (3.4) 2 (3.6) FANCM c.5791C>T 1 (1.8) 1 (1.8) 1 (1.9) 1 (1.9) 2 (1.8) 2 (1.8) 2 (1.0) 0 PALB2 c.1592delT 0 0 2 (3.8) 2 (3.8) 2 (1.8) 2 (1.8) 2 (1.0) 0 RAD51C c.93delG 1 (1.8) 1 (1.8) 0 0 1 (0.9) 1 (0.9) 1 (0.5) 0 Total 3 4 (7.1) 7 (12.5) 6 (11.3) 9 (17.0) 10 (9.2) 16 (14.7) 18 (8.7) 7 (12.7) 1 Only the mutations detected in the BRCA1/2 families are presented in the table. 2 BRCA1/2 carriers including the index cases and family members with any cancer type. 3 Triple heterozygotes are included only once in the total carrier frequencies. Table 5. Frequencies of the second moderate-risk mutation in the carriers of BRCA1/2 or moderate-penetrance mutation Second mutation Group Carriers/total % OR 1 95% CI p-value Moderate-risk mutation carrier 2 8/129 3 6.2 1.72 0.69–3.79 0.155 BRCA1 index patient 4/56 7.1 2.01 0.51–5.80 0.164 BRCA2 index patient 6/53 11.3 3.32 1.11–8.34 0.016 BRCA1/2 index patient 10/109 9.2 2.63 1.15–5.48 0.011 Moderate-risk mutation carrier 2 or BRCA1/2 index patient 17/237 3,4 7.2 2.01 1.06–3.65 0.021 1 The frequency of a second mutation in the mutation carrier groups was compared to the overall frequency of moderate-risk mutations in the Helsinki population controls (47/1273, 3.7%). 2 In the Helsinki unselected breast cancer series. 3 Includes one CHEK2 c.1100delC homozygote and one BRCA2 c.7480C>T, CHEK2 c.1100delC and FANCM c.5101C>T triple heterozygote. 4 One BRCA2 index patient included also in the Helsinki unselected breast cancer series is counted only once in the total frequencies. 2698 Recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics history of the disease. 8 RAD51C and RAD51D have recently been suggested as risk genes for triple-negative BC, 19 while their contribution to OC with a fiveto sixfold increased risk has been acknowledged longer. 16–18 We have previously identified FANCM as a BC predisposing gene in the Finnish population with approximately twofold increased risk of BC. 14,15 The highest risk was seen in the patients with the triplenegative subtype. Similar results have been published in other studies. 44,45 FANCM has also been suggested as an OC susceptibility gene. 46 Here, 3.1% of the unselected OC patients carried any of the three FANCM mutations. We report a novel association between CHEK2 c.319 +2T>A and BC risk. This mutation, located at a canonical splice site, was previously observed in a gene panel sequencing of BC patients from Finland 30 and in a Norwegian patient with Cowden-like syndrome, thyroid cancer and bilateral BC. 47 We identified c.319+2T>A in 0.7% of BC patients in the combined Helsinki and Tampere series. The mutation was associated with a 5.4-fold risk of BC in the unselected patient group, higher than the previous estimates for CHEK2 mutations. The risk associated with c.319+2T>A here may be an overestimate and the true risk closer to the approximately threefold risk associated with the truncating CHEK2 mutations. 9,10,13 However, our estimate of sixfold increased risk in the familial patient group is in line with the earlier studies on CHEK2. 9,10 We have previously detected the Eastern European founder mutation CHEK2 c.444+1G>A in a gene panel sequencing of Finnish BC and OC patients. 29 As the c.444+1G>A mutation was rare (0.1%) in our BC series, we were not able to confirm the associated risk in this study. The carrier frequencies of ATM c.6908dupA and c.7570G>C have previously been estimated in a Northern Finnish study. 26 Here, we genotyped these rare ATM mutations in a large series of 3,156 BC patients and 2,089 population controls to establish the carrier frequencies in Southern Finland. ATM c.6908dupA was detected in 0.8% of unselected BC patients from the Tampere region, but only in 0.2% of familial and 0.1% of unselected cases from the Helsinki region. ATM c.7570G>C was identified in 0.3% of familial and 0.1% of unselected BC patients in the Helsinki series, but absent in the Tampere patients. In a study by Pylkäs et al., 26 a haplotype analysis suggested both c.6908dupA and c.7570G>C to be unique founder mutations, the former originating from the Tampere region. The frequency of these pathogenic ATM mutations was about twoto threefold higher in cases in comparison to the population controls, which is consistent with previous results on ATM. 13,48 As the c.6908dupA and c.7570G>C mutations are rare, a family-based penetrance study might be better suited to estimate the risk effect associated with these mutations. Genetic testing allows personalized risk assessment and can lead to disease-preventive actions. An improved diagnostic yield can be achieved with gene panels compared to single gene-based tests. In our study, moderate-penetrance mutations were identified with a relatively high overall frequency in the patients and the results support the utilization of gene panels in clinical testing, possibly for more than one individual in the family. Several pathogenic variants may segregate also in a BRCA-positive family and thus, the family members testing negative for BRCA1/2 mutations may remain at an increased risk of BC or OC due to other predisposing mutations. Furthermore, Manchanda et al. 49 estimated that, in UK and USA, population-based testing of BRCA1/2and moderate-risk genes would be more cost-effective than testing by clinical or family history-based criteria. In a founder population, like the Finns, screening the few recurrent risk mutations could give similar cost-benefits. In conclusion, our study provides overall frequencies for recurrent moderate-risk gene mutations in BC and OC patients in Southern Finland. The frequency of the moderatepenetrance gene mutations was significantly different from the frequency in the population controls also in the BRCA1/2 index patients and our data support the multiplicative risk model, with the second mutation conferring additional increase in risk. Our results underline the benefits of gene panel testing, possibly for multiple members of BC families where several pathogenic variants may segregate in different individuals. Acknowledgements We thank research nurse Outi Malkavaara for the help with the patient data and Dr Katri Pylkäs for providing a positive control sample for genotyping. We express our gratitude to all the patients and their family members for participating in the study. Author contributions AN, TAM, LMP and HN designed the study and drafted the manuscript. AN, LMP, JIK, TAM, TH and SL carried out the mutation screening. AN and TAM performed the statistical analyses with LMP. KA, CB, RB, AK and JS contributed samples and patient information. All authors contributed to and approved the final manuscript. Data availability The data that support the findings of our study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References 1. Bray F, Ferlay J, Soerjomataram I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018;68:394–424. 2. Collaborative Group on Hormonal Factors in Breast Cancer. Familial breast cancer: collaborative reanalysis of individual data from 52 epidemiological studies including 58,209 women with breast cancer and 101,986 women without the disease. Lancet 2001;358:1389–99. 3. Kuchenbaecker KB, Hopper JL, Barnes DR, et al. Risks of breast, ovarian, and contralateral breast Nurmi et al.2699 Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics cancer for BRCA1 and BRCA2 mutation carriers. JAMA 2017;317:2402–16. 4. Vehmanen P, Friedman LS, Eerola H, et al. Low proportion of BRCA1 and BRCA2 mutations in Finnish breast cancer families: evidence for additional susceptibility genes. Hum Mol Genet 1997;6:2309–15. 5. Syrjakoski K, Vahteristo P, Eerola H, et al. Population-based study of BRCA1 and BRCA2 mutations in 1035 unselected Finnish breast cancer patients. J Natl Cancer Inst 2000;92:1529–31. 6. Sarantaus L, Auranen A, Nevanlinna H. BRCA1 and BRCA2 mutations among Finnish ovarian carcinoma families. Int J Oncol 2001;18:831–5. 7. Sarantaus L, Vahteristo P, Bloom E, et al. BRCA1 and BRCA2 mutations among 233 unselected Finnish ovarian carcinoma patients. Eur J Hum Genet 2001;9:424–30. 8. Antoniou AC, Casadei S, Heikkinen T, et al. Breast-cancer risk in families with mutations in PALB2. N Engl J Med 2014;371:497–506. 9. Weischer M, Bojesen SE, Ellervik C, et al. CHEK2*1100delC genotyping for clinical assessment of breast cancer risk: meta-analyses of 26,000 patient cases and 27,000 controls. J Clin Oncol 2008;26:542–8. 10. Cybulski C, Wokolorczyk D, Jakubowska A, et al. Risk of breast cancer in women with a CHEK2 mutation with and without a family history of breast cancer. J Clin Oncol 2011;29:3747–52. 11. Schmidt MK, Hogervorst F, van Hien R, et al. Ageand tumor subtype-specific Breast Cancer risk estimates for CHEK2*1100delC carriers. J Clin Oncol 2016;34:2750–60. 12. Easton DF, Pharoah PD, Antoniou AC, et al. Gene-panel sequencing and the prediction of breast-cancer risk. N Engl J Med 2015;372: 2243–57. 13. Decker B, Allen J, Luccarini C, et al. Rare, protein-truncating variants in ATM, CHEK2 and PALB2, but not XRCC2, are associated with increased breast cancer risks. J Med Genet 2017; 54:732–41. 14. Kiiski JI, Pelttari LM, Khan S, et al. Exome sequencing identifies FANCM as a susceptibility gene for triple-negative breast cancer. Proc Natl Acad Sci U S A 2014;111:15172–7. 15. Kiiski JI, Tervasmaki A, Pelttari LM, et al. FANCM mutation c.5791C>T is a risk factor for triple-negative breast cancer in the Finnish population. Breast Cancer Res Treat 2017;166:217–26. 16. Loveday C, Turnbull C, Ramsay E, et al. Germline mutations in RAD51D confer susceptibility to ovarian cancer. Nat Genet 2011;43:879–82. 17. Loveday C, Turnbull C, Ruark E, et al. Germline RAD51C mutations confer susceptibility to ovarian cancer. Nat Genet 2012;44:475–6; author reply 76. 18. Lilyquist J, LaDuca H, Polley E, et al. Frequency of mutations in a large series of clinically ascertained ovarian cancer cases tested on multigene panels compared to reference controls. Gynecol Oncol 2017;147:375–80. 19. Shimelis H, LaDuca H, Hu C, et al. Triplenegative Breast Cancer risk genes identified by multigene hereditary Cancer panel testing. J Natl Cancer Inst 2018;110:855–62. 20. Michailidou K, Lindstrom S, Dennis J, et al. Association analysis identifies 65 new breast cancer risk loci. Nature 2017;551:92–4. 21. Lim ET, Wurtz P, Havulinna AS, et al. Distribution and medical impact of loss-of-function variants in the Finnish founder population. PLoS Genet 2014;10:e1004494. 22. Erkko H, Xia B, Nikkila J, et al. A recurrent mutation in PALB2 in Finnish cancer families. Nature 2007;446:316–9. 23. Heikkinen T, Karkkainen H, Aaltonen K, et al. The breast cancer susceptibility mutation PALB2 1592delT is associated with an aggressive tumor phenotype. Clin Cancer Res 2009;15: 3214–22. 24. Vahteristo P, Bartkova J, Eerola H, et al. A CHEK2 genetic variant contributing to a substantial fraction of familial breast cancer. Am J Hum Genet 2002;71:432–8. 25. Allinen M, Launonen V, Laake K, et al. ATM mutations in Finnish breast cancer patients. J Med Genet 2002;39:192–6. 26. Pylkas K, Tommiska J, Syrjakoski K, et al. Evaluation of the role of Finnish ataxia-telangiectasia mutations in hereditary predisposition to breast cancer. Carcinogenesis 2007;28:1040–5. 27. Pelttari LM, Heikkinen T, Thompson D, et al. RAD51C is a susceptibility gene for ovarian cancer. Hum Mol Genet 2011;20:3278–88. 28. Pelttari LM, Kiiski J, Nurminen R, et al. A Finnish founder mutation in RAD51D: analysis in breast, ovarian, prostate, and colorectal cancer. J Med Genet 2012;49:429–32. 29. Pelttari LM, Shimelis H, Toiminen H, et al. Genepanel testing of breast and ovarian cancer patients identifies a recurrent RAD51C duplication. Clin Genet 2018;93:595–602. 30. Mantere T, Winqvist R, Kauppila S, et al. Targeted next-generation sequencing identifies a recurrent mutation in MCPH1 associating with hereditary breast cancer susceptibility. PLoS Genet 2016;12:e1005816. 31. Kilpivaara O, Bartkova J, Eerola H, et al. Correlation of CHEK2 protein expression and c.1100delC mutation status with tumor characteristics among unselected breast cancer patients. Int J Cancer 2005;113:575–80. 32. Fagerholm R, Hofstetter B, Tommiska J, et al. NAD(P)H:quinone oxidoreductase 1 NQO1*2 genotype (P187S) is a strong prognostic and predictive factor in breast cancer. Nat Genet 2008;40: 844–53. 33. Eerola H, Blomqvist C, Pukkala E, et al. Familial breast cancer in southern Finland: how prevalent are breast cancer families and can we trust the family history reported by patients? Eur J Cancer 2000;36:1143–8. 34. Weischer M, Nordestgaard BG, Pharoah P, et al. CHEK2*1100delC heterozygosity in women with breast cancer associated with early death, breast cancer-specificdeath,and increased risk of a second breast cancer. JClin Oncol 2012;30:4308–16. 35. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of protein-coding genetic variation in 60,706 humans. Nature 2016;536:285–91. 36. Couch FJ, Shimelis H, Hu C, et al. Associations between Cancer predisposition testing panel genes and Breast Cancer. JAMA Oncol 2017;3:1190–6. 37. Antoniou AC, Pharoah PD, McMullan G, et al. A comprehensive model for familial breast cancer incorporating BRCA1, BRCA2 and other genes. Br J Cancer 2002;86:76–83. 38. Antoniou AC, Beesley J, McGuffog L, et al. Common breast cancer susceptibility alleles and the risk of breast cancer for BRCA1 and BRCA2 mutation carriers: implications for risk prediction. Cancer Res 2010;70:9742–54. 39. Mavaddat N, Pharoah PD, Michailidou K, et al. Prediction of breast cancer risk based on profiling with common genetic variants. J Natl Cancer Inst 2015;107:djv036. 40. Muranen TA, Greco D, Blomqvist C, et al. Genetic modifiers of CHEK2*1100delC-associated breast cancer risk. Genet Med 2017;19:599–603. 41. Sokolenko AP, Bogdanova N, Kluzniak W, et al. Double heterozygotes among breast cancer patients analyzed for BRCA1, CHEK2, ATM, NBN/NBS1, and BLM germ-line mutations. Breast Cancer Res Treat 2014;145:553–62. 42. Turnbull C, Seal S, Renwick A, et al. Gene-gene interactions in breast cancer susceptibility. Hum Mol Genet 2012;21:958–62. 43. Cybulski C, Gorski B, Huzarski T, et al. Effect of CHEK2 missense variant I157T on the risk of breast cancer in carriers of other CHEK2 or BRCA1 mutations. J Med Genet 2009;46:132–5. 44. Peterlongo P, Catucci I, Colombo M, et al. FANCM c.5791C>T nonsense mutation (rs144567652) induces exon skipping, affects DNA repair activity and is a familial breast cancer risk factor. Hum Mol Genet 2015;24:5345–55. 45. Neidhardt G, Hauke J, Ramser J, et al. Association between loss-of-function mutations within the FANCM gene and early-onset familial breast cancer. JAMA Oncol 2017;3:1245–8. 46. Dicks E, Song H, Ramus SJ, et al. Germline whole exome sequencing and large-scale replication identifies FANCM as a likely high grade serous ovarian cancer susceptibility gene. Oncotarget 2017;8:50930–40. 47. Dominguez-Valentin M, Nakken S, Tubeuf H, et al. Potentially pathogenic germline CHEK2 c.319+2T>A among multiple early-onset cancer families. Fam Cancer 2018;17:141–53. 48. Tavtigian SV, Oefner PJ, Babikyan D, et al. Rare, evolutionarily unlikely missense substitutions in ATM confer increased risk of breast cancer. Am J Hum Genet 2009;85:427–46. 49. Manchanda R, Patel S, Gordeev VS, et al. Costeffectiveness of population-based BRCA1, BRCA2, RAD51C, RAD51D, BRIP1, PALB2 mutation testing in unselected general population women. J Natl Cancer Inst 2018;110:714–25. 2700 Recurrent moderate-risk mutations in Finnish breast and ovarian cancer patients Int. J. Cancer: 145, 2692–2700 (2019) ©2019 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of UICC Cancer Genetics and Epigenetics