Advertisement

High detection rate from genetic testing in BRCA-negative women with familial epithelial ovarian cancer

  • Nicola Flaum
    Correspondence
    Correspondence and requests for materials should be addressed to Nicola Flaum, North West Genomics Laboratory Hub, Manchester Centre for Genomic Medicine, St. Mary’s Hospital, Oxford Road, Manchester, M13 9WL, United Kingdom
    Affiliations
    Division of Evolution, Infection and Genomics, School of Biological Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester Academic Health Science Centre, United Kingdom

    North West Genomics Laboratory Hub, Manchester Centre for Genomic Medicine, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester Academic Health Science Centre, United Kingdom
    Search for articles by this author
  • Emma J. Crosbie
    Affiliations
    Division of Cancer Sciences, School of Medical Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Uunited Kingdom

    Division of Gynaecology, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Search for articles by this author
  • Richard Edmondson
    Affiliations
    Division of Cancer Sciences, School of Medical Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Uunited Kingdom

    Division of Gynaecology, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Search for articles by this author
  • Emma R. Woodward
    Affiliations
    Clinical Genetics Service, Manchester Centre for Genomic Medicine, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Search for articles by this author
  • Fiona Lalloo
    Affiliations
    Clinical Genetics Service, Manchester Centre for Genomic Medicine, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester, United Kingdom
    Search for articles by this author
  • Miriam J. Smith
    Affiliations
    Division of Evolution, Infection and Genomics, School of Biological Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester Academic Health Science Centre, United Kingdom

    North West Genomics Laboratory Hub, Manchester Centre for Genomic Medicine, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester Academic Health Science Centre, United Kingdom
    Search for articles by this author
  • Helene Schlecht
    Affiliations
    North West Genomics Laboratory Hub, Manchester Centre for Genomic Medicine, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester Academic Health Science Centre, United Kingdom
    Search for articles by this author
  • D. Gareth Evans
    Affiliations
    Division of Evolution, Infection and Genomics, School of Biological Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester Academic Health Science Centre, United Kingdom

    North West Genomics Laboratory Hub, Manchester Centre for Genomic Medicine, St Mary’s Hospital, Manchester University NHS Foundation Trust, Manchester Academic Health Science Centre, United Kingdom

    Prevention Breast Cancer Centre and Nightingale Breast Screening Centre, University Hospital of South Manchester, Manchester, United Kingdom

    The Christie NHS Foundation Trust, Manchester, United Kingdom

    Manchester Breast Centre, Manchester Cancer Research Centre, University of Manchester, United Kingdom
    Search for articles by this author
Open AccessPublished:September 28, 2022DOI:https://doi.org/10.1016/j.gim.2022.08.022

      ABSTRACT

      Purpose

      Epithelial ovarian cancer (EOC) is associated with pathogenic variants (PVs) in homologous recombination and/or mismatch repair genes. We aimed to review the testing of women with familial EOC at our center.

      Methods

      Women with familial EOC (≥2 EOC in family, including index case) referred to our center between 1993 and 2021 were included. Genetic testing (BRCA/Lynch syndrome screening, exome sequencing, panel testing, 100,000 Genome Project, and NIHR BioResource genome sequencing) and clinical demographic, diagnosis, and survival data were reviewed.

      Results

      Of 277, 128 (46.2%) women were BRCA heterozygotes (BRCA1: 89, BRCA2: 39). The detection rate in BRCA-negative women was 21.8%; the most commonly affected gene was BRIP1 (5.9%). The non-BRCA detection rate was significantly higher in families with 2 affected members with EOC only (22.4%) than the families with ≥3 (11.1%) affected members (odds ratio = 9.9, 95% CI = 1.6-105.2, P = .0075). Overall, 112 different PVs in 12 homologous recombination/mismatch repair genes were detected in 150 of 277 (54.2%) unrelated women.

      Conclusion

      This is the largest report of women with familial EOC undergoing wider testing to date. One-fifth of BRCA-negative women were heterozygous for a PV in a potentially actionable gene. Wider genetic testing of women with familial EOC is essential to optimize their treatment and prevention of disease in family members.

      Keywords

      Introduction

      One of the most relevant risk factors for epithelial ovarian cancer (EOC) is a family history of breast and/or EOC (hereditary breast and ovarian cancer). Approximately 10% to 15% of ovarian cancer is thought to be hereditary, although, estimates vary.
      • Walsh T.
      • Casadei S.
      • Lee M.K.
      • et al.
      Mutations in 12 genes for inherited ovarian, fallopian tube, and peritoneal carcinoma identified by massively parallel sequencing.
      • Zheng G.
      • Yu H.
      • Kanerva A.
      • Försti A.
      • Sundquist K.
      • Hemminki K.
      Familial risks of ovarian cancer by age at diagnosis, proband type and histology.
      • Stratton J.F.
      • Pharoah P.
      • Smith S.K.
      • Easton D.
      • Ponder B.A.
      A systematic review and meta-analysis of family history and risk of ovarian cancer.
      • Alsop K.
      • Fereday S.
      • Meldrum C.
      • et al.
      BRCA mutation frequency and patterns of treatment response in BRCA mutation-positive women with ovarian cancer: a report from the Australian Ovarian Cancer Study Group.
      There is a 3-fold increase in ovarian cancer risk in women with a first-degree relative with ovarian cancer.
      • Stratton J.F.
      • Pharoah P.
      • Smith S.K.
      • Easton D.
      • Ponder B.A.
      A systematic review and meta-analysis of family history and risk of ovarian cancer.
      Pathogenic variants (PVs) in high risk genes, BRCA1 and BRCA2, moderate risk genes, such as RAD51C/D, and genes involved in mismatch repair (MMR) contribute to approximately 20% to 25% of all EOCs.
      • Walsh T.
      • Casadei S.
      • Lee M.K.
      • et al.
      Mutations in 12 genes for inherited ovarian, fallopian tube, and peritoneal carcinoma identified by massively parallel sequencing.
      • Jervis S.
      • Song H.
      • Lee A.
      • et al.
      Ovarian cancer familial relative risks by tumour subtypes and by known ovarian cancer genetic susceptibility variants.
      Genes identified as associated with an increased ovarian cancer risk include those involved in homologous recombination (HR) (BRCA1, BRCA2, RAD51C, RAD51D, BRIP1, and PALB2) and MMR (MSH2, MSH6, MLH1, and PMS2). The prevalence of MMR-deficiency or microsatellite instability in familial ovarian cancer has been estimated to be 10% to 20%.
      • Pal T.
      • Permuth-Wey J.
      • Sellers T.A.
      A review of the clinical relevance of mismatch-repair deficiency in ovarian cancer.
      • Xiao X.
      • Melton D.W.
      • Gourley C.
      Mismatch repair deficiency in ovarian cancer – molecular characteristics and clinical implications.
      A number of different genetic testing strategies are available to investigate the cause of familial ovarian cancer, and these have evolved over time as techniques such as next-generation sequencing (NGS) were developed. They are also adapted for family history, such as a BRCA1/BRCA2 screen for a family history of breast and ovarian cancer or a screen for Lynch syndrome for women with a family history of ovarian and endometrial and/or colorectal cancer. After the discovery of a number of high and moderate susceptibility genes associated with increased risk of EOC, the use of cancer gene panels began to be introduced in the United Kingdom, initially predominantly through private or research testing. There was some controversy regarding the use of these panels. To address this, in 2018, the UK Cancer Genetics Group supported by the UK Genetic Testing Network (UKGTN) published a consensus statement to be adopted by the UK National Health Service.
      • Taylor A.
      • Brady A.F.
      • Frayling I.M.
      • et al.
      Consensus for genes to be included on cancer panel tests offered by UK genetics services: guidelines of the UK Cancer Genetics Group.
       It was recommended with majority (>75%) vote that ovarian cancer panels include BRCA1, BRCA2, MLH1, MSH2, MSH6, PMS2, RAD51C, and RAD51D, and the inclusion of BRIP1 was made after expert presentation and discussion due to contributing sufficient risk of ovarian cancer that risk-reduction salpingo-oophorectomy could be considered.
      • Taylor A.
      • Brady A.F.
      • Frayling I.M.
      • et al.
      Consensus for genes to be included on cancer panel tests offered by UK genetics services: guidelines of the UK Cancer Genetics Group.
      Because our historical familial EOC cases had not been tested with the full recommended genetic screen and to assess the effect of ovarian cancer panel testing conducted through research or the NHS clinical services, we aimed to review testing outcomes of women with familial ovarian cancer known to our center.

      Materials and Methods

      Study participants

      We reviewed the genetic test results and clinical data of 277 women referred to the Manchester Centre for Genomic Medicine (MCGM) between 1993 and 2021 with a diagnosis of EOC and at least 1 additional first- or second-degree relative with a diagnosis of EOC. Patients were retrospectively identified from within the patient database of women known to MCGM. Clinical information regarding testing results, date of EOC diagnosis, date of sample test, other cancer diagnoses, age at EOC diagnosis, EOC histology, stage at diagnosis, survival data, family history, and Manchester score (MS) were obtained from the MCGM laboratory database; Manchester University NHS Foundation Trust clinical records software, and the Christie NHS Foundation Trust. The MS is a model developed to determine the probability of an index case being a BRCA1/2 heterozygote using the data on an individual’s breast, ovarian, pancreatic, and prostate cancer history and family history.
      • Evans D.G.
      • Harkness E.F.
      • Plaskocinska I.
      • et al.
      Pathology update to the Manchester Scoring System based on testing in over 4000 families.

      Genetic testing

      Genetic testing from 1996 onward included BRCA1/2 screening. Panel testing was introduced in 2016, therefore, most samples received retrospective testing. All women with EOC up to 2017 were tested after referral to genetics, and a minority of women with familial EOC were tested as part of mainstreaming by medical oncologists only for BRCA1/2 and PALB2.
      Screening of BRCA1 and BRCA2 was performed using Sanger sequencing or sequencing using long range polymerase chain reaction, Nextera XT library preparation and MiSeq sequencing and dosage analysis using MRC-Holland multiple ligation-dependent probe amplification (MLPA) probe kits, P002 for BRCA1 and P045 for BRCA2 (MRC-Holland). Testing for each individual is described in Supplemental Table 1. If a BRCA1/2 PV was detected, no further testing was carried out, and it was considered the causal factor in this woman’s familial EOC. If a CHEK2 deletion or duplication was detected using P045, it was further analyzed using probe kit P190. PVs were confirmed using Sanger sequencing. NGS was performed for the Inherited Cancer Panel using the Illumina SureSelect XT next-generation sequencing (NextSeq) (Illumina) using the Agilent BRAVO robot (Agilent). The Inherited Cancer Panel for ovarian cancer included BRCA1, BRCA2, BRIP1, MLH1, MSH2, MSH6, RAD50, RAD51C, and RAD51D.
      The Lynch syndrome screening involved screening for MLH1, MSH2, and MSH6 performed through sequencing using long range polymerase chain reaction, Nextera XT library preparation, and MiSeq sequencing. Variants were confirmed using Sanger sequencing. Dosage analysis was performed using MRC-Holland MLPA probe kits, P003 for MLH1 and MSH2 and P072 for MSH6 (MRC-Holland). MSH6 exon 1 was screened using Sanger sequencing owing to poor coverage of this region through NGS technology. PMS2 was tested using a bespoke sequencing and MLPA kit if suspected.
      Exome sequencing was performed by BGI using the BGISEQ-500 sequencer (BGI). FASTQ files were received from BGI already demultiplexed and mapped against HG19 using Burrows-Wheeler Aligner. Sequence Alignment Map files were converted to Binary Alignment Map files using SAMTools. Duplicates were marked on the Binary Alignment Map file using Picard software. Base recalibration and base quality score recalibration were performed using the Genome Analysis Toolkit. Variants were called using HaplotypeCaller within Genome Analysis Toolkit. Variant quality score recalibration to flag likely false positives and realignment were performed on the variant call format files. For logistic reasons, the samples were processed individually, and joint calling was not performed. Variants were processed on VarSeq (Golden Helix Inc) and exported to Microsoft Excel per sample. Variants with read depths of <30 were excluded. Common variants with an allele frequency of >1% were also excluded.
      Patients recruited through the NIHR BioResource study underwent genome sequencing and cancer panel screening (Illumina TruSight Cancer panel). This is described in detail elsewhere.
      • Whitworth J.
      • Smith P.S.
      • Martin J.E.
      • et al.
      Comprehensive cancer-predisposition gene testing in an adult multiple primary tumor series shows a broad range of deleterious variants and atypical tumor phenotypes.

      Variant classification

      Multiple in silico prediction tools were used to assess potential variant effect prediction. Clinical prediction software Alamut Visual version 2.15 (SOPHiA GENETICS) was used, which incorporates scores from splicing prediction tools MaxEntScan, NNSPLICE, GeneSplicer, Exonic Splicing Enhancer (ESE) tools, and missense prediction tools Sorting Intolerant From Tolerant,
      • Sim N.L.
      • Kumar P.
      • Hu J.
      • Henikoff S.
      • Schneider G.
      • Ng P.C.
      SIFT web server: predicting effects of amino acid substitutions on proteins.
      MutationTaster, Polymorphism Phenotyping v2,
      • Adzhubei I.A.
      • Schmidt S.
      • Peshkin L.
      • et al.
      A method and server for predicting damaging missense mutations.
      and data from ClinVar,
      • Landrum M.J.
      • Lee J.M.
      • Riley G.R.
      • et al.
      ClinVar: public archive of relationships among sequence variation and human phenotype.
      database for Nonsynonymous SNPs’ Functional Predictions,
      • Liu X.
      • Wu C.
      • Li C.
      • Boerwinkle E.
      dbNSFP v3.0: a one-stop database of functional predictions and annotations for human nonsynonymous and splice-site SNVs.
      and Catalog of Somatic Mutations in Cancer (COSMIC). Missense variants were further assessed using the in silico tool Rare Exome Variant Ensemble Learner (REVEL).
      • Ioannidis N.M.
      • Rothstein J.H.
      • Pejaver V.
      • et al.
      REVEL: an ensemble method for predicting the pathogenicity of rare missense variants.
      Variant classification according to the American College of Medical Genetics and Genomics/Association for Molecular Pathology criteria was checked and confirmed.

      Statistical analysis

      Independent t test, Fisher exact test, and odds ratios were calculated using GraphPad Prism 8.4.3.

      Results

      Demographics

      A total of 277 unrelated women with familial EOC were identified, and they underwent testing as shown in Figure 1. Most of these women had high grade serous ovarian cancer (HGSOC) or what was historically classified as poorly differentiated cystadenocarcinoma, the majority of which would likely be classified as HGSOC under modern histology assessment (Table 1). Data regarding age at diagnosis were available for 270 women, and mean age was 56 (range: 27-86) years.
      Figure thumbnail gr1
      Figure 1Flowchart of variants detected through BRCA screening and wider testing. EOC, epithelial ovarian cancer; HR, homologous recombination; MCGM, Manchester Centre for Genomic Medicine; MMR, mismatch repair; MPS, Multiple Primary tumours Study.
      Table 1Breakdown of cases by histologic subtype
      Histologic SubtypeNumber (%)Heterozygous for PV Detected (%)Heterozygous for VUS Detected (%)
      High grade serous ovarian carcinoma156 (56.3)86 (55.1)10 (6.4)
      Poorly differentiated cystadenocarcinoma42 (15.2)31 (73.8)5 (11.9)
      Endometrioid ovarian carcinoma31 (11.2)18 (58.1)1 (3.2)
      Clear cell ovarian carcinoma6 (2.2)5 (83.3)0
      Carcinosarcoma6 (2.2)2 (33.3)0
      Low grade serous ovarian carcinoma4 (1.4)01 (25)
      Mixed endometrioid/clear cell ovarian carcinoma3 (1.1)1 (33.3)0
      Serous tubal intraepithelial carcinoma1 (0.4)00
      Borderline1 (0.4)00
      Mucinous1 (0.4)00
      Unknown26 (9.4)7 (26.9)2 (7.7)
      Serous200 (72.2)113 (56.5)16 (8)
      Total277150 (54.1)20 (7.2)
      PV, pathogenic variant; VUS, variant of uncertain significance.
      In total, 47 (17.0%) individuals also had breast cancer (BC) in addition to EOC; 38 (80.9%) of these were in women with clinically significant PVs, and there were 23 other cancer diagnoses in 21 women (Supplemental Table 1). Of the 47 women with BC, mean age at BC diagnosis was 48.5 (range: 32-68) years, 47.8% had ER-positive tumors, and 9 women (19.1%) had triple-negative BCs. A PV was detected in 37 (78.7%) of these women, and 35 (74.5%) were detected in BRCA1/BRCA2. Detection rate by personal and family history is summarized in Table 2.
      Table 2Detection rate by personal and family cancer history
      Personal Cancer HistoryTotal (n)BRCA1/2 (%)Non-BRCA HR Gene PVs (%)Non-BRCA MMR Gene PVs (%)
      OC only20986 (41.1)9 (4.3)6 (2.9)
      OC + BC4333 (76.7)1 (2.3)0
      OC + CRC52 (40)02 (40)
      OC + EN7002 (28.6)
      OC + other137 (53.8)1 (7.7)0
      Family cancer history
      OC only11239 (34.8)5 (4.5)6 (5.4)
      OC + BC only13679 (58.1)6 (4.4)1 (0.7)
      OC + BC + other115 (45.4)01 (9.1)
      OC + CRC105 (50)02 (20)
      OC + EN3001 (33.3)
      OC + other (non-BC/CRC/EN)92 (22.2)00
      BC, breast cancer; CRC, colorectal cancer; EN, endometrial carcinoma; HR, homologous recombination; MMR, mismatch repair; OC, ovarian carcinoma; PV, pathogenic variant.

      Detection rates

      PVs were detected in 128 of 277 (46.2%) women (BRCA1: 89, BRCA2: 39). Of the remaining 149 BRCA-negative women, further testing was carried out in 101 women. All testing by individual is described in Supplemental Table 2. This included a Lynch syndrome screen or ovarian cancer panel screen through the NHS or testing via research consisting of exome sequencing of HR/MMR genes, genome sequencing through the 100,000 Genome Project, or genome sequencing through the NIHR Bioresource MPS study (Figure 1).
      In total, 48 BRCA-negative women were unable to undergo further testing owing to absent or poor quality DNA samples. There was no significant difference between the average age at diagnosis and death of women who underwent further testing and those who did not (57.6 years vs 57.7 years, P = .97 and 65.5 years vs 64.1 years; P = .63, respectively). There was also no significant difference in the proportion of women alive 5 years after diagnosis (χ21, n=116 = 3.8, P = .051). However, there was a significant difference in MS, with women who underwent further testing scoring significantly higher (25.7 vs 20.6, P = .0002).
      In the 101 BRCA-negative women who underwent further testing, the detection rate was 22 of 101 (21.8%). The PVs detected were equally split between MMR and HR genes, with BRIP1 being the most commonly affected gene (n = 6; 5.9%).
      In total, 52 BRCA-negative women were from families with at least 3 family members with EOC. Of these women, 39 (75%) had a BRCA1/2 PV (BRCA1: 30, BRCA2: 9) (Table 3). One woman (1.9%) had a RAD51D PV with no PVs in other genes identified in this subset.
      Table 3Detection rates grouped by family history and testing method
      Affected RelativesTotalBRCA1BRCA2BRCA NegativeNon-BRCA PV Detected/Tested Using Specified Testing MethodAdjusted Prediction of BRCA Negative (n)Adjusted Prediction of BRCA Negative (%)
      2 OC only11118 (16.2%)11 (9.9%)8213/58 (22.4%)18.416.6
      Testing performedBRCA screen1031811740
      Lynch syndrome screen1000107/10 (70%)
      Ovarian cancer panel2302203/20 (15%)
      Research exome/genome sequencing3401326/32 (18.8%)
      2 OC + BC11441 (36.0%)19 (16.7%)548/36 (22.2%)12.010.5
      Testing performedBRCA screen1134119530
      Lynch syndrome screen40031 (33.3%)
      Ovarian cancer panel1200123 (25%)
      Research exome/genome sequencing80061/6 (16.7%)
      3+ OC5230 (57.7%)9 (17.3%)131/9 (11.1%)1.40.7
      Testing performedBRCA screen51308131
      Lynch syndrome screen0NANANANA
      Ovarian cancer panel50141/4 (25%)
      Research exome/genome sequencing50050
      BC, breast cancer; NA, not available; OC, ovarian cancer; PV, pathogenic variant.
      The detection rate of variants in non-BRCA genes was highest in families with 2 affected members with EOC only (22.4%) and lowest in families with 3 or more affected members with EOC (11.1%) (Table 3). The difference between adjusted detection rates in families with 2 compared with that in families with 3 affected members with EOC was statistically significant (odds ratio = 9.9, 95% CI = 1.6-105.2, P = .0075). To see if this was owing to the histology of EOC predominant in each group, we compared the proportions of HGSOC in each group, however, there was no significant differences.
      Overall, 153 PVs, 112 different PVs, in 12 genes, including BRCA1/2, were detected in 150 of 277 (54.2%) women (Supplemental Table 3). Variants of uncertain significance (VUS) were also detected in 20 women (Supplemental Table 4).

      Concurrent variants

      Two women with BRCA1/2 PVs were also heterozygotes for a concurrent PV in another HR gene. In one woman with a BRCA1 PV, a PALB2 PV was also detected on a BRCA1/BRCA2/PALB2 screen. This was requested because BRCA1/2 testing on the pathology sample identified the BRCA1 c.4625_4626delCT variant, but the protocols at the time determined that complete testing of all 3 genes should still be performed on blood samples. The other woman was found to be a heterozygote for a BRCA2 and concurrent CHEK2 PVs, detected from the BRCA1/2 screen. Dosage analysis of BRCA1 and BRCA2 and further analysis with the P190-C1 (CHEK2) MLPA probe mix showed this copy number variant to extend from CHEK2 exons 3 to 15. MLPA does not provide the location of the extra copy of CHEK2 exon 3 to exon 15, and it is possible that the CHEK2 gene was not disrupted. Neither CHEK2 nor PALB2 were considered causative PVs in these cases and were not counted in the BRCA-negative women analysis.

      PVs

      Of the 62 different BRCA1 and 30 BRCA2 PVs detected, most were frameshift-causing variants (Supplemental Table 3). Variants that occurred in 3 or more unrelated individuals can be seen in the Supplemental Table 3. Of note was the detection of missense variant BRIP1 c.1045G>C; p.(Ala349Pro) in 3 unrelated individuals. This variant has been described previously by our group.
      • Flaum N.
      • van Veen E.M.
      • Smith O.
      • et al.
      Dominant-negative pathogenic variant BRIP1 c.1045G>C is a high-risk allele for non-mucinous epithelial ovarian cancer: a case-control study.

      Clinical outcomes

      We analyzed differences between age at diagnosis, stage at diagnosis, and survival data in women with different PVs (Table 4). The proportion of women diagnosed at stage 1/2 with a BRCA1 or MMR PV was higher than the proportion with a BRCA2 or HR PV. Data were limited for the MMR PV group, however, the finding of an earlier age at diagnosis and an earlier stage at diagnosis was not unusual for this group.
      • Ryan N.A.J.
      • Evans D.G.
      • Green K.
      • Crosbie E.J.
      Pathological features and clinical behavior of Lynch syndrome-associated ovarian cancer.
      • Ryan N.A.J.
      • Bolton J.
      • McVey R.J.
      • Evans D.G.
      • Crosbie E.J.
      BRCA and lynch syndrome-associated ovarian cancers behave differently.
      There was no significant difference in survival at 5 years post-diagnosis between any of the groups categorized by PV.
      Table 4Stage, age at diagnosis, and survival outcomes
      Genes/AffectedStaging Data Available/Total NumberStage at DiagnosisAlive 5 Years After Diagnosis
      Data available for 241 women, 25 women were diagnosed within 5 years of analysis cutoff date and therefore were ineligible for analysis.
      Mean Age at Diagnosis (Median)Mean Age at Death (Median)
      Data available for 240 women, 148 women passed away during the analysis period.
      Mean Manchester Score
      1/234
      All198/27749 (24.7%)114 (57.6%)35 (17.7%)121/216 (56%)55.8 (55.5)62.5 (62)28.9
      BRCA174/8922 (29.7%)39 (52.7%)13 (17.6%)38/68 (55.9%)53.3 (53)57 (58)36.1
      BRCA232/395 (15.6%)22 (68.8%)5 (15.6%)19/32 (59.4%)63.2 (65)64.8 (63)30.6
      BRCA1/BRCA2106/12827 (25.4%)61 (57.5%)18 (17.0%)57/100 (57.0%)54.8 (53)59.6 (58)34.4
      HR PV115/13929 (25.2%)67 (58.3%)19 (16.5%)62/109 (56.9%)55.5 (54)60.4 (58.5)33.9
      MMR PV3/113/3 (100%)NANA6/7 (85.7%)43.5 (47.5)45.5
      Data only available for 2 women.
      27
      Any PV118/15032 (27.1%)67 (56.8%)19 (16.1%)68/116 (58.6%)55.5 (54)60 (58.5)33.4
      No PV80/12717 (21.3%)47 (58.8%)16 (20.0%)53/100 (53.0%)56.1 (56)65.2 (64.5)23.6
      HR, homologous recombination; MMR, mismatch repair; NA, not available; PV, pathogenic variant.
      a Data available for 241 women, 25 women were diagnosed within 5 years of analysis cutoff date and therefore were ineligible for analysis.
      b Data available for 240 women, 148 women passed away during the analysis period.
      c Data only available for 2 women.

      MS

      We examined differences between MS in different groups. The mean MS in the study population was 28.9 (range: 10-71), and the mean results by variant are shown in Table 4. There was a statistically significant difference between women with and women without a PV (independent t test: P < .0001) but not between heterozygotes for HR PVs and heterozygotes for MMR PVs (P = .0505). The MS is designed to predict the probability of BRCA PVs on the basis of personal and family history of EOC and BC. These results are therefore not surprising.
      To assess if the MS was helpful at predicting non-BRCA PV, we compared the mean MS in women with any BRCA PV (34.4) with all other women (24.1) (unpaired t test: 2-tailed P < .0001). The mean MS of women with any PV, including BRCA (33.4), was statistically significantly different from women without a detected PV (23.6) (unpaired t test: 2-tailed: P < .0001).

      Discussion

      In our study, a different HR or MMR PV was detected in 1 in 5 BRCA-negative women with EOC. A number of international studies have been undertaken to assess detection rates from cancer gene panels/NGS in women with EOC. These studies predominantly assessed women with hereditary breast and ovarian cancer and included varying levels of data regarding family history. Cancer gene panels/NGS in women with familial ovarian cancer in the United Kingdom have not been previously reported. As can be seen in Supplemental Table 5, study detection rates in BRCA-negative ovarian cancer gene panels vary between 0% to 10% internationally. Most of these studies were small and did not report detection rates by family history or reported only in very small numbers, although, this is an easily measurable statistic monitored in both oncology and genetics clinics. We found a detection rate of 21.8% out of 101 BRCA-negative women with familial ovarian cancer. This is substantially higher than similar cancer gene panels in BRCA-negative patients with BC in whom detection rates range from 1% to 12%.
      • Catana A.
      • Apostu A.P.
      • Antemie R.G.
      Multi gene panel testing for hereditary breast cancer – is it ready to be used?.
      This shows the utility of testing in patients with familial ovarian cancer for HR and MMR genes beyond BRCA1 and BRCA2.
      All women in this study came from families with 2 or more relatives with EOC. We are unaware of any previous reports that have concentrated on this relatively rare situation. The PV detection rates in the BRCA-negative population were significantly lower in families containing 3 family members with EOC than those with only 2 members with EOC, and it seems that further testing is most important in women with EOC and only 2 affected relatives in the family. Further analysis showed that differences in histology between groups could not explain this difference. This could be because lower and moderate penetrance genes are more likely to cause EOC in women in this group. As such, BRCA2, which is associated with a 15% to 30% lifetime ovarian cancer risk, is found in families with 3 ovarian cancer cases, but this is clearly much less likely for the remaining genes with penetrance rates of <15%.
      Our study found the highest prevalence of non-BRCA PV in BRIP1 driven by the dominant negative BRIP1 c.1045G>C missense variant.
      • Flaum N.
      • van Veen E.M.
      • Smith O.
      • et al.
      Dominant-negative pathogenic variant BRIP1 c.1045G>C is a high-risk allele for non-mucinous epithelial ovarian cancer: a case-control study.
      Identification of this variant alone would justify inclusion of BRIP1 on ovarian cancer gene panels because the penetrance of this variant is likely to be >10%. The degree of increase in risk for BRIP1 for all EOCs was estimated from a large case-control study to be 11.22-fold (95% CI = 3.22-34.10). Despite this, the penetrance was estimated to be only 5.8% by age 80 years on the basis of testing in unaffected women from the UK Ovarian Cancer Screening Study. This represents only a 3- to 4-fold risk at the bottom end of 95% CIs of the case-control study.
      • Ramus S.J.
      • Song H.
      • Dicks E.
      • et al.
      Germline mutations in the BRIP1, BARD1, PALB2, and NBN genes in women with ovarian cancer.
      This relatively low lifetime risk meant that BRIP1 did not reach the 75% consensus for inclusion of BRIP1 in the United Kingdom ovarian panel, although, it was added later.
      • Taylor A.
      • Brady A.F.
      • Frayling I.M.
      • et al.
      Consensus for genes to be included on cancer panel tests offered by UK genetics services: guidelines of the UK Cancer Genetics Group.
      It is clearly time to revisit this penetrance estimate given that case-control evidence for BRIP1 is much more in keeping with risks equivalent to RAD51C and RAD51D, (5- to 6-fold relative risk),
      • Lilyquist J.
      • LaDuca H.
      • Polley E.
      • et al.
      Frequency of mutations in a large series of clinically ascertained ovarian cancer cases tested on multi-gene panels compared to reference controls.
      and it is arguable whether testing of unaffected women (UK Ovarian Cancer Screening Study) gives a robust estimate.
      At present, the detection of BRCA1 or BRCA2 PVs in germline or somatic testing in a woman with EOC enables the use of PARP inhibitors, which offer a survival advantage.
      • Mirza M.R.
      • Monk B.J.
      • Herrstedt J.
      • et al.
      Niraparib maintenance therapy in platinum-sensitive, recurrent ovarian cancer.
      • Pujade-Lauraine E.
      • Ledermann J.A.
      • Selle F.
      • et al.
      Olaparib tablets as maintenance therapy in patients with platinum-sensitive, relapsed ovarian cancer and a BRCA1/2 mutation (SOLO2/ENGOT-Ov21): a double-blind, randomised, placebo-controlled, phase 3 trial.
      • Coleman R.L.
      • Oza A.M.
      • Lorusso D.
      • et al.
      Rucaparib maintenance treatment for recurrent ovarian carcinoma after response to platinum therapy (ARIEL3): a randomised, double-blind, placebo-controlled, phase 3 trial.
      The ENGOT-OV26/PRIMA phase III trial described improved survival in women with HR-deficient EOC beyond BRCA PVs.
      • González-Martín A.
      • Pothuri B.
      • Vergote I.
      • et al.
      Niraparib in patients with newly diagnosed advanced ovarian cancer.
      In 2017, the US Food and Drug Administration approved the use of PD-1 inhibitor pembrolizumab for the use of microsatellite-high or MMR deficient solid tumors that had progressed with first-line treatment.
      • Marcus L.
      • Lemery S.J.
      • Keegan P.
      • Pazdur R.
      FDA approval summary: pembrolizumab for the treatment of microsatellite instability-high solid tumors.
      It is clear that maximizing the use of targeted treatments based on genetic testing is an increasingly used tactic, optimizing EOC management, in particular, treatments targeting HR and MMR pathways.
      This study had several limitations. The numbers were small relative to the larger national studies and also geographically limited to patients referred to the MCGM. However, family history taking and confirmation of family history is likely to have been more complete than the studies from gene testing companies. Nevertheless, this is the largest series to our knowledge to assess testing of familial EOC with at least 2 cases in the family. Women reviewed at MCGM are from high risk families, introducing an element of ascertainment bias. There is also the potential for survivor bias because women with very aggressive disease may not have time to be referred for genetic testing. To assess this aspect, we reviewed the time from EOC diagnosis to sample taken. This ranged from –232 months (19.3 years) (as a result of an alternative cancer diagnosis or cascade testing) to +565 months (47 years), with a mean of 40 months and median of 12 months. Our patients therefore seem to reflect a wide range of referral times to clinical genetics. Because these patients were referred over a 28-year time period, a range of genetic analysis techniques were used, and patients were tested for different genes over different periods. The Lynch syndrome screening did not include PMS2. However, because newer technologies and clinical studies became available, retrospective samples were eligible for newer testing, therefore, this bias was minimal. The clinical reports from MCGM testing also had a practice of not reporting VUS in routine reports; therefore, Supplemental Table 4 will be an underestimate of the VUS in this group. A total of 7 VUS were also reclassified as benign under current reporting guidelines.
      In conclusion, in this article, we describe the largest series of women with familial ovarian cancer undergoing cancer gene panel/NGS screening to date and the only one to our knowledge in the United Kingdom. As expected, almost half of the women tested positive for a BRCA1/2 PV, however, 1 in 5 of the remaining women who underwent testing were found to be heterozygotes for another EOC-associated gene. This is a substantial proportion of potentially actionable genes that could affect patient care with the increasing use of targeted therapies in women with HR- or MMR-gene associated EOC as well as use of risk-reduction surgery. Our previous identification of BRIP1 as the most prevalent gene in this higher penetrance EOC setting should overturn any question relating to its recommended inclusion in EOC panel,
      • Taylor A.
      • Brady A.F.
      • Frayling I.M.
      • et al.
      Consensus for genes to be included on cancer panel tests offered by UK genetics services: guidelines of the UK Cancer Genetics Group.
      and we recommend BRIP1 testing for all cases of nonmucinous epithelial cancer particularly in BRCA1/2-negative women with HR-deficient ovarian cancers and women with a family history of EOC in which an affected relative is not available. Wider genetic testing of women with familial ovarian cancer is essential to both optimize their treatment and enable prevention of disease in family members.

      Data Availability

      Data are available from the corresponding author on request.

      Conflict of Interest

      The authors declare no conflicts of interest.

      Acknowledgments

      D.G.E., M.J.S., and E.J.C. are supported by the NIHR Manchester Biomedical Research Centre (IS-BRC-1215-20007). E.J.C. is an NIHR Advanced Fellow (NIHR300650). N.F. is supported by Cancer Research UK via the funding to Cancer Research UK Manchester Centre (C147/A18083 and C147/A25254). The authors like to thank all the staff working in the laboratory at Manchester Centre for Genomic Medicine.

      Author Information

      Conceptualization: D.G.E., N.F.; Data Curation: D.G.E., N.F.; Formal Analysis: N.F.; Funding Acquisition: D.G.E.; Investigation: N.F., D.G.E.; Methodology: N.F., D.G.E.; Project Administration: N.F.; Resources: D.G.E., H.S., F.L., E.R.W.; Supervision: D.G.E., E.J.C., R.E., M.J.S.; Validation: N.F.; Visualization: N.F.; Writing-original draft: N.F., D.G.E.; Writing-review and editing: N.F., D.G.E., E.J.C., F.L., E.R.W., M.J.S., H.S.

      Ethics Declaration

      Informed consent was obtained from patients who underwent BRCA screening, Lynch syndrome screening, Manchester Centre for Genomic Medicine epithelial ovarian cancer panel sequencing, and exome sequencing to have their samples used in future anonymized research (FH-Risk approved by the NHS North Manchester Research Ethics Committee [08/H1006/77] and the University of Manchester Ethics Committee [08229] and a later substantial amendment to incorporate the study, “Investigation of genetic modifiers in BRCA1/2 breast cancer and non-BRCA1/2 high risk families” [Reference 08/H1006/77] was approved by Greater Manchester West [GM West] Research Ethics Committee). Women involved in the 100,000 Genome Project and NIHR BioResource study were consented through their clinical trial protocol.

      Supplementary Material

      References

        • Walsh T.
        • Casadei S.
        • Lee M.K.
        • et al.
        Mutations in 12 genes for inherited ovarian, fallopian tube, and peritoneal carcinoma identified by massively parallel sequencing.
        Proc Natl Acad Sci U S A. 2011; 108: 18032-18037https://doi.org/10.1073/pnas.1115052108
        • Zheng G.
        • Yu H.
        • Kanerva A.
        • Försti A.
        • Sundquist K.
        • Hemminki K.
        Familial risks of ovarian cancer by age at diagnosis, proband type and histology.
        PloS One. 2018; 13 (Published correction appears in PLoS One. 2018;13(10):e0206721)e0205000
        • Stratton J.F.
        • Pharoah P.
        • Smith S.K.
        • Easton D.
        • Ponder B.A.
        A systematic review and meta-analysis of family history and risk of ovarian cancer.
        Br J Obstet Gynaecol. 1998; 105: 493-499https://doi.org/10.1111/j.1471-0528.1998.tb10148.x
        • Alsop K.
        • Fereday S.
        • Meldrum C.
        • et al.
        BRCA mutation frequency and patterns of treatment response in BRCA mutation-positive women with ovarian cancer: a report from the Australian Ovarian Cancer Study Group.
        J Clin Oncol. 2012; 30 (Published correction appears in J Clin Oncol. 2012;30(33):4180): 2654-2663
        • Jervis S.
        • Song H.
        • Lee A.
        • et al.
        Ovarian cancer familial relative risks by tumour subtypes and by known ovarian cancer genetic susceptibility variants.
        J Med Genet. 2014; 51: 108-113https://doi.org/10.1136/jmedgenet-2013-102015
        • Pal T.
        • Permuth-Wey J.
        • Sellers T.A.
        A review of the clinical relevance of mismatch-repair deficiency in ovarian cancer.
        Cancer. 2008; 113: 733-742https://doi.org/10.1002/cncr.23601
        • Xiao X.
        • Melton D.W.
        • Gourley C.
        Mismatch repair deficiency in ovarian cancer – molecular characteristics and clinical implications.
        Gynecol Oncol. 2014; 132: 506-512https://doi.org/10.1016/j.ygyno.2013.12.003
        • Taylor A.
        • Brady A.F.
        • Frayling I.M.
        • et al.
        Consensus for genes to be included on cancer panel tests offered by UK genetics services: guidelines of the UK Cancer Genetics Group.
        J Med Genet. 2018; 55: 372-377https://doi.org/10.1136/jmedgenet-2017-105188
        • Evans D.G.
        • Harkness E.F.
        • Plaskocinska I.
        • et al.
        Pathology update to the Manchester Scoring System based on testing in over 4000 families.
        J Med Genet. 2017; 54: 674-681https://doi.org/10.1136/jmedgenet-2017-104584
        • Whitworth J.
        • Smith P.S.
        • Martin J.E.
        • et al.
        Comprehensive cancer-predisposition gene testing in an adult multiple primary tumor series shows a broad range of deleterious variants and atypical tumor phenotypes.
        Am J Hum Genet. 2018; 103: 3-18https://doi.org/10.1016/j.ajhg.2018.04.013
        • Sim N.L.
        • Kumar P.
        • Hu J.
        • Henikoff S.
        • Schneider G.
        • Ng P.C.
        SIFT web server: predicting effects of amino acid substitutions on proteins.
        Nucleic Acids Res. 2012; 40: W452-W457https://doi.org/10.1093/nar/gks539
        • Adzhubei I.A.
        • Schmidt S.
        • Peshkin L.
        • et al.
        A method and server for predicting damaging missense mutations.
        Nat Methods. 2010; 7: 248-249https://doi.org/10.1038/nmeth0410-248
        • Landrum M.J.
        • Lee J.M.
        • Riley G.R.
        • et al.
        ClinVar: public archive of relationships among sequence variation and human phenotype.
        Nucleic Acids Res. 2014; 42: D980-D985https://doi.org/10.1093/nar/gkt1113
        • Liu X.
        • Wu C.
        • Li C.
        • Boerwinkle E.
        dbNSFP v3.0: a one-stop database of functional predictions and annotations for human nonsynonymous and splice-site SNVs.
        Hum Mutat. 2016; 37: 235-241https://doi.org/10.1002/humu.22932
        • Ioannidis N.M.
        • Rothstein J.H.
        • Pejaver V.
        • et al.
        REVEL: an ensemble method for predicting the pathogenicity of rare missense variants.
        Am J Hum Genet. 2016; 99: 877-885https://doi.org/10.1016/j.ajhg.2016.08.016
        • Flaum N.
        • van Veen E.M.
        • Smith O.
        • et al.
        Dominant-negative pathogenic variant BRIP1 c.1045G>C is a high-risk allele for non-mucinous epithelial ovarian cancer: a case-control study.
        Clin Genet. 2022; 101: 48-54https://doi.org/10.1111/cge.14068
        • Ryan N.A.J.
        • Evans D.G.
        • Green K.
        • Crosbie E.J.
        Pathological features and clinical behavior of Lynch syndrome-associated ovarian cancer.
        Gynecol Oncol. 2017; 144: 491-495https://doi.org/10.1016/j.ygyno.2017.01.005
        • Ryan N.A.J.
        • Bolton J.
        • McVey R.J.
        • Evans D.G.
        • Crosbie E.J.
        BRCA and lynch syndrome-associated ovarian cancers behave differently.
        Gynecol Oncol Rep. 2017; 22: 108-109https://doi.org/10.1016/j.gore.2017.11.007
        • Catana A.
        • Apostu A.P.
        • Antemie R.G.
        Multi gene panel testing for hereditary breast cancer – is it ready to be used?.
        Med Pharm Rep. 2019; 92: 220-225https://doi.org/10.15386/mpr-1083
        • Ramus S.J.
        • Song H.
        • Dicks E.
        • et al.
        Germline mutations in the BRIP1, BARD1, PALB2, and NBN genes in women with ovarian cancer.
        J Natl Cancer Inst. 2015; 107: djv214https://doi.org/10.1093/jnci/djv214
        • Lilyquist J.
        • LaDuca H.
        • Polley E.
        • et al.
        Frequency of mutations in a large series of clinically ascertained ovarian cancer cases tested on multi-gene panels compared to reference controls.
        Gynecol Oncol. 2017; 147: 375-380https://doi.org/10.1016/j.ygyno.2017.08.030
        • Mirza M.R.
        • Monk B.J.
        • Herrstedt J.
        • et al.
        Niraparib maintenance therapy in platinum-sensitive, recurrent ovarian cancer.
        N Engl J Med. 2016; 375: 2154-2164https://doi.org/10.1056/NEJMoa1611310
        • Pujade-Lauraine E.
        • Ledermann J.A.
        • Selle F.
        • et al.
        Olaparib tablets as maintenance therapy in patients with platinum-sensitive, relapsed ovarian cancer and a BRCA1/2 mutation (SOLO2/ENGOT-Ov21): a double-blind, randomised, placebo-controlled, phase 3 trial.
        Lancet Oncol. 2017; 18 (Published correction appears in Lancet Oncol. 2017;18(9):e510): 1274-1284
        • Coleman R.L.
        • Oza A.M.
        • Lorusso D.
        • et al.
        Rucaparib maintenance treatment for recurrent ovarian carcinoma after response to platinum therapy (ARIEL3): a randomised, double-blind, placebo-controlled, phase 3 trial.
        Lancet. 2017; 390 (Published correction appears in Lancet. 2017;390(10106):1948): 1949-1961
        • González-Martín A.
        • Pothuri B.
        • Vergote I.
        • et al.
        Niraparib in patients with newly diagnosed advanced ovarian cancer.
        N Engl J Med. 2019; 381: 2391-2402https://doi.org/10.1056/NEJMoa1910962
        • Marcus L.
        • Lemery S.J.
        • Keegan P.
        • Pazdur R.
        FDA approval summary: pembrolizumab for the treatment of microsatellite instability-high solid tumors.
        Clin Cancer Res. 2019; 25: 3753-3758https://doi.org/10.1158/1078-0432.CCR-18-4070