Background Multigene panel testing by next-generation sequencing (MGP-NGS) enables the detection of germline pathogenic or likely pathogenic variants (PVs/LPVs) in genes beyond those associated with a certain cancer phenotype. Opportunistic genetic screening based on MGP-NGS in patients with suspicion of hereditary cancer reveals these incidental findings (IFs).
Methods MGP-NGS was performed in patients who fulfilled the clinical criteria to undergo genetic testing according to the Catalan Health Service guidelines. Variants were classified following the American College of Medical Genetics and Genomics-Association for Molecular Pathology guidelines and the Cancer Variant Interpretation Group UK guidelines.
Results IFs were identified in 10 (1.22%) of the 817 patients who underwent MGP-NGS. The mean age at cancer diagnosis was 49.4±9.5 years. Three IFs (30.0%) were detected in PMS2, two (20.0%) in ATM and TP53 and one (10.0%) in MSH6, NTHL1 and VHL. Seven (70.0%) IFs were single-nucleotide substitutions, two (20.0%) were deletions and one (10.0%) was a duplication. Three (30.0) IFs were located in intronic regions, three (30.3%) were nonsense, two (20.0%) were frameshift and two (20.0%) were missense variations. Six (60.0%) IFs were classified as PVs and four (40.0%) as LPVs.
Conclusions Opportunistic genetic screening increased the diagnostic yield by 1.22% in our cohort. Most of the identified IFs were present in clinically actionable genes (n=7; 70.0%), providing these families with an opportunity to join cancer early detection programmes, as well as secondary cancer prevention. IFs might facilitate the diagnosis of asymptomatic individuals and the early management of cancer once it develops.
- congenital, hereditary, and neonatal diseases and abnormalities
- medical oncology
Data availability statement
Data are available on reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
WHAT IS ALREADY KNOWN ON THIS TOPIC
Hereditary cancer syndromes are caused by inherited alterations in >200 cancer-related genes, most of which are involved in cell cycle control and DNA repair; however, in only 5%–10% of cases with a hereditary cancer suspicion, the disease-causing alteration is identified.
Opportunistic genetic screening through multigene panel testing by next-generation sequencing (MGP-NGS) can unveil unexpected pathogenic or likely pathogenic variants (PVs/LPVs) initially inconsistent with the personal and familial cancer history of the patient.
WHAT THIS STUDY ADDS
Our results confirm that MGP-NGS performed on patients with personal and/or familial histories of cancer uncovers genetic alterations inconsistent with personal and familial history of cancer, known as incidental findings.
Interestingly, we report that opportunistic genetic screening increases the diagnostic yield by 1.22% in our cohort.
In addition, 70% of the identified incidental findings in this study are present in clinically actionable genes.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The identification of incidental findings in clinically actionable genes can provide clinical benefits for patients harbouring the variant and their relatives.
Patients can benefit from personalised oncological treatments.
In the case of asymptomatic individuals, opportunistic genetic screening could yield the discovery of an as-yet clinically unrecognised disorder for its early management.
Finally, opportunistic genetic screening provides an opportunity to join cancer early detection programmes as well as to secondary cancer prevention to those healthy relatives harbouring the alteration.
The development of next-generation sequencing (NGS) marks a turning point in the genetic diagnosis field. Compared with the gold standard technique for mutation analysis in cancer diagnosis, the Sanger sequencing method, NGS can sequence a large number of DNA regions while being time-effective and cost-effective at the same time. NGS offers considerable benefits in clinical settings since it allows for molecular characterisation of rare diseases, individualisation of oncological treatments and population screening for disease risk, among other abilities. Moreover, these cutting-edge technologies have significantly improved assay sensitivity and enabled multigene panel (MGP) testing, that is, simultaneous sequencing of multiple genes.1
More than 200 genes have been associated with hereditary cancer syndromes, most of which are implicated in cell cycle control and DNA repair. However, only in 5%–10% of patients with clinical suspicion of hereditary cancer is the disease-causing variant identified.2 Most MGPs undertaken in these patients are phenotype-driven since they include those genes associated with a certain cancer phenotype, while some of them also comprise additional genes associated with increased hereditary cancer risk. In this sense, opportunistic genetic screening through MGP-NGS testing can unveil unexpected pathogenic or likely pathogenic variants (PVs/LPVs), known as incidental findings (IFs). IFs have been defined by the American College of Medical Genetics and Genomics (ACMG) as the results of a deliberate search for PVs/LPVs in genes that are not apparently relevant to a diagnostic indication for which the sequencing test was ordered but that could nonetheless be medically valuable to the patient and the ordering clinician.3
The identification of germline IFs might have considerable implications for the clinician, the patient and relatives.4–6 When a PV/LPV is identified in clinically actionable genes (online supplemental table S1), the patient can benefit from personalised treatment selection and monitoring programmes. Therefore, the role of IF in managing the patients’ and their relatives’ health and in correctly assessing the risks of developing pathologies is of paramount importance. However, some potential drawbacks could emerge from IFs discovery, especially in patients with no familial history of hereditary cancer. The ACMG has recently published an update of its policy statement providing recommendations for reporting IFs in clinical contexts. Despite adherence to this statement being voluntary, these recommendations provide high-quality clinical and laboratory genetic services.7
In the present work, we aimed to evaluate the presence of PVs/LPVs in genes that are not related to the primary clinical indications for MGP-NGS ordering through an opportunistic genetic screening approach.
Materials and methods
Participants and DNA obtaining
In this population-based, retrospective chart review, patients were selected from the Oncology Institute of South Catalonia (IOCS). These patients underwent genetic testing according to the Catalan Health Service guidelines (table 1, online supplemental appendix S1).8 DNA was extracted from peripheral blood lymphocytes using the Gentra PureGene DNA Isolation Kit (Qiagen).
MGP-NGS library preparation was based on the Imegen Hereditary OncoKitDx kit (Health In Code), which targets coding exons and 20 bp of the flanking intronic regions of 50 genes (185.56 kb) relevant for hereditary cancer syndromes (online supplemental table S2). Products were analysed by NGS using the Illumina Platform MiSeq. Data were analysed with the Datagenomics platform (Health In Code). Only target regions with a minimum depth of 20× were considered. Variants were assumed to be of germline origin if found with a variant allele frequency (VAF) ≥20%.
DNA Sanger sequencing
PVs/LPVs detected by NGS were confirmed by Sanger sequencing with BigDye Terminator V.3.1 kit (Life Technologies). Long-range PCR was performed to discard PMS2 pseudogene contamination.9 Long-range PCR products were used as the template for nested PMS2 amplification. In all cases, capillary electrophoresis was conducted on a SeqStudio sequencer (Applied Biosystems) and analysed using Sequencher V.5.0 software (Gene Codes).
The clinical significance of variants was examined following the ACMG and the Association for Molecular Pathology standards and guidelines,10 the ClinGen Variant Expert Curation Panel specifications for TP53, MMR genes and ATM11–13 and the Cancer Variant Interpretation Group UK guidelines (CanVIG-UK), following a 5-tier classification system.14 In silico predictive studies were performed with those tools recommended elsewhere.10 15 Variant frequencies were analysed through the Genome Aggregation Database (gnomAD) browser and the 1000 Genomes Project. Variants were examined in databases such as ClinVar, OMIM, Leiden Open Variation Database and BRCA Share and by reviewing updated bibliographies. The predicted consequences of splice variants were conducted mainly by Splice AI. The analyses of the consequences of missense variants were examined by REVEL and PRIORS.16
Fibroblasts from skin punch biopsy culture
A 5 mm skin biopsy was fragmented under sterile conditions and cultured using the explant technique.17 DNA from cell culture was extracted manually with the Qiagen Gentra Puregene blood kit (Qiagen).
Between July 2020 and July 2022, a total of 817 individuals referred to the Cancer Genetic Counselling Unit of the IOCS met the Catalan Health Service clinical criteria to undergo MGP-NGS testing (table 1 and online supplemental appendix S1). Eighty-six (10.53%) of these patients harboured a PV/LPV in a gene consistently associated with their particular diagnosed cancer (online supplemental appendix S1). IFs were identified in 10 additional subjects (1.22%) and therefore comprised the study population of the present work. Among these 10 subjects, the majority were women (n=9; 90.0%), and the mean age at cancer diagnosis was 49.4±9.5 years. The most prevalent cancer type was breast cancer (BC; MIM:114480; n=7; 53.84%), followed by colorectal (MIM:114500; n=4; 30.77%), ovarian (MIM:167000; n=1; 7.69%) and cervical (MIM:603956; n=1; 7.69%) cancers (see table 2 for summarised information and online supplemental table S3 for complete information).
NGS features, confirmation of the results and general characteristics of the IFs identified
Most of the 10 IFs included in this study were detected in patients referred for hereditary breast and ovarian syndrome suspicion (MIM:PS604370; n=7; 63.63%), followed by Lynch syndrome (MIM:120435; n=4; 30.77%). The IFs identified in our study were located in six different genes: three IFs (30.0%) were detected in PMS2, two (20.0%) in ATM, two (20.0%) in TP53 and one (10.0%) in the MSH6, NTHL1 and VHL genes. All IFs were identified in heterozygosis except for the variant identified in NTHL1. Five IF (50.0%) were classified as PV (see table 2 for summarised information and online supplemental table S3 for complete information).
IF in patients with breast and ovarian cancer
The NM_000546.6:c.743G>A variant in TP53 was identified in patient 1 diagnosed with BC at 45–50 years (VAF=0.427; see online supplemental figure S1, table 2 for summarised information and online supplemental table S3 for complete information). This SNV in a hotspot region of the TP53 gene leads to a missense change in the protein (p.(Arg248Gln)).18 This particular variant was classified by the ClinGen TP53 Variant Curation Panel as a PV based on their specific classification guidelines (PM1, PP3, PS3, PS4 and PS2).11 This patient had a confirmed family history of BC in a first-degree relative (FDR) and in a third-degree relative (TDR), who underwent a mastectomy and died in her 30s because of cancer. None of the relatives tested were carriers of this variant. The variant was identified in the buccal swab (VAF=20%) and in tumorous and non-tumorous breast tissues (VAFs≈50%), but not in skin fibroblasts, as confirmed using Minor Variant Finder Software (Applied Biosystems).
The variant NM_000546.6:c.783-1G>A in TP53 was identified in patient 2 (VAF=0.569), who was diagnosed with bilateral BC (aged 45–50 years), cervical cancer (aged 45–50 years) and colorectal cancer (aged 65–70 years) with conserved immunohistochemistry (IHC) for mismatch repair genes (MMR) (online supplemental figure S2). This variant is a single nucleotide substitution variant (SNV) located in a splicing consensus region of TP53 that results in aberrant transcripts and a truncated protein with compromised function, as experimental studies have confirmed.19–21 This variant is not found in population databases, but it is reported in the National Cancer Institute of the US TP53 database in Li-Fraumeni syndrome and Li-Fraumeni syndrome-like families (last accessed October 2022). This variant has been classified as LPV according to the ClinGen TP53 Variant Curation Panel specific guidelines (PS4, PP1, PM2 and PP3).11 Patient 2 had a long familial history of cancer: three FDRs, six second-degree relatives (SDRs) and two TDRs. Cosegregation analysis allowed us to confirm that this LPV in TP53 has a germline origin.
The variant NM_000179.3:c.762dup in MSH6 (VAF=0.420) was identified in patient 3, who was diagnosed with BC at the age range of 35–40 years (online supplemental figure S3). This variant is a duplication with a nonsense coding effect (p.(Glu255*)), surely resulting in the loss of function of the MSH6 protein by premature protein truncation or nonsense-mediated mRNA decay (NMD). The variant is not found in the gnomAD database, and some authors have reported this variant as pathogenic in ClinVar. This variant has been classified as a PV according to the ClinGen ISiGHT Hereditary Colorectal Cancel/polyposis Variant Curation Expert Panel-specific classification guidelines (PVS1, PP5 and PM2).12 The patient had a confirmed family history of breast and uterine cancer in an FDR (subject IV:2, both diagnosed at age 50–55 years, respectively) and two SDRs who suffered from mouth cancer (subject IV:8, aged 55–60 years) and both colorectal (aged 55–60 years) and prostate (aged 75–80 years) cancers (subject III:1). Moreover, posterior IHC studies performed in the endometrial tumorous tissue of subject IV:2 confirmed the lack of expression of MSH6.
The variant NM_000535.7:c.24-2A>G in PMS2 was found in patient 4 (VAF=0.513; online supplemental figure S4), a woman diagnosed with asynchronous breast and bilateral BC (at age 35–40 and 45–50 years) and with a confirmed familial history of BC in two FDRs (subjects III:12 and II:10, diagnosed at age 60–65 and 80–85 years, respectively, the latter deceased from BC) and in an SDR (III:5, diagnosed at age 60–65 years), as detailed in online supplemental figure S4. IHC studies showed that the expression of MMR proteins was conserved in the tumorous tissue. This intronic variant is an SNV of two nucleotides upstream from coding exon 2 in the PMS2 gene. Despite the lack of direct evidence, splice site prediction tools predict that this variant abolishes the canonical splice acceptor site. Thus, it is expected to result in aberrant transcripts subject to NMD. However, further analyses are warranted to discern whether this variant impairs the normal splicing process and compromises protein function. Moreover, this variant is not described in population databases (gnomAD). Therefore, it is classified as an LPV according to the ClinGen ISiGHT Hereditary Colorectal Cancel/Polyposis Variant Curation Expert Panel-specific classification guidelines (PVS1 and PM2).12
The variant NM_000535.7:c.1579_1580del in PMS2 was found in patient 5 (VAF=0.480; online supplemental figure S5), diagnosed with BC at age 50–55 years. IHC studies confirmed the lack of PMS2 expression in breast tumorous tissue. This two-nucleotide deletion presumably results in a frameshift coding effect (p.(Arg527Glyfs*14)) and may create a stop codon producing a disrupted or absent protein by NMD, but no functional evidence is reported. Loss-of-function variants in PMS2 are a known mechanism of disease.22 23 However, no functional evidence of this particular variant has been published to date. This variant has been previously found in subjects with HBOC and in population databases at a very low frequency.24 25 Therefore, this variant is classified as a PV according to the ClinGen ISiGHT Hereditary Colorectal Cancel/Polyposis Variant Curation Expert Panel-specific classification guidelines (PM2 and PP5).12 No blood or tissue samples of relatives were available to perform further studies.
The variant NM_000535.7:c.989-2A>G in PMS2 (VAF=0.500) was identified in a woman affected by ovarian cancer diagnosed at age 50–55 years (patient 6; online supplemental figure S6), who underwent HBOC and Lynch syndrome MGPs-NGS. The NM_000535.7:c.989-2A>G variant in PMS2 is an intronic SNV located two nucleotides upstream from coding exon 10 in PMS2, affecting an acceptor site in intron 9. This particular variant is not found in population databases (gnomAD). This variant is classified as an LPV according to the ClinGen ISiGHT Hereditary Colorectal Cancel/polyposis Variant Curation Expert Panel-specific classification guidelines (PM2 and PP5).12 The proband had a long familial history of cancer, particularly on the paternal side. Two FDRs were affected by colorectal cancer and myeloma; four SDRs by liver, prostate, colon and gastric cancers and three TDRs by brain, ovarian and both ovarian and BC. Most of these relatives died of the oncological process. We could perform carrier studies, as well as access the clinical records, of the IHC of some of the relatives. In this sense, the IHC of subject II:2 showed that tumorous colon tissue had conserved expression of MMR proteins. In contrast, there was a lack of expression of the proteins PMS2 and MSH6 in the ovarian tissue of subject III:1. Genetic studies of subject III:10 showed that this individual harboured the same variant in PMS2, but no IHC analysis of this patient has been accessible. In an attempt to discern the familial origin of the variant in PMS2, carrier studies of available blood samples were performed. Subjects III:5 and III:3 did not harbour this variant. Interestingly, subject III:6 was diagnosed with ovarian and BC, but genetic studies showed that she carried a PV in BRCA2 inherited from the maternal side. The ovarian IHC for the MMR proteins of subject III:6 was normal. We cannot assure the maternal or paternal origin of the variant identified in patient 6, as it was not detected in the paternal side and genetic studies on the maternal side were not possible due to the lack of contact of the proband with her maternal family. The paternal family history of cancer may lead us to think that the variant NM_000535.7:c.989-2A>G in PMS2 was inherited from the paternal side, but we cannot guarantee it. Taking this into consideration, we can neither assure nor refuse that subjects II:2 and III:1 are carriers of the same variant in PMS2, as discussed in the ‘Discussion’ section.
The variant NM_000551.4:c.341G>A in VHL was found in patient 7 (VAF=0.501) (online supplemental figure S7), a woman diagnosed with BC at the age of 45–50 years. This SNV leads to a missense coding effect, resulting in a change in the amino acid in the VHL protein (p.(Gly114Asp)). To our knowledge, this variant has not been previously reported or described in gnomAD. This variant is located in a hotspot region where 25 PVs/LPVs, 8 variants of uncertain significance and no benign variants have been described. In addition, four pathogenic alternative changes have been reported in the same residue. The in silico tool used (Revel) predicts altered function of the VHL protein, but no functional studies have been performed to confirm it. Therefore, this variant is classified as an LPV according to the ACMG and CanVIG-UK guidelines (PM1, PM5, PP3, PM2).10 22 Carrier studies performed in available specimens showed that subject II:4 also harboured this variant, but no cancer was diagnosed in this subject despite being aged 90–95 years. One can assume that this variant has probably a low penetrance but follow-up tests recommended to this patient showed that abdominal echography was normal, no retinal hemangioblastoma was detected and normal values of blood metanephrines were found, endorsing the low penetrance of this variant.
IFs in patients with colorectal cancer
The variant NM_000051.4:c.5908C>T in ATM was found in patient 8, diagnosed with colorectal cancer at age 35–40 years (VAF=0.448) (see online supplemental figure S8, table 2 for summarised information and online supplemental table S3 for complete information). Microsatellite instability analysis showed no evidence of MMR deficiency. This SNV in exon 39 has been predicted to have a nonsense coding effect (p.(Gln1970*)). The predicted creation of a premature stop codon could result in a truncated or absent ATM protein due to NMD. This variant has a low frequency in gnomAD and has been classified as a PV according to the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel (PVS1, PP5 and PM2).10 13 On the maternal side of patient 8, a large deletion of exon 14 of BRCA1 was identified in several relatives, but the ATM variant was absent on this maternal side. On the paternal side, the proband had one FRD diagnosed with bladder and lung cancers and one SDR diagnosed with colorectal cancer. According to this long history of familial cancer, the proband underwent HBOC MGP-NGS analysis, but no additional PVs/LPVs were identified. Carrier studies confirmed that the proband did not harbour the deletion in BRCA1. No conclusive results were obtained in the cosegregation studies performed.
The variant NM_000051.4:c.7670_7674del was identified in ATM (VAF=0.442) in patient 9 diagnosed with colorectal cancer at age 65–70 years (online supplemental figure S9). This deletion in exon 52 leads to a frameshift coding effect (p.(Leu2557Tyrfs*12)), in turn, resulting in the creation of a premature stop codon, and it is predicted that it causes a loss of function of the ATM protein. In addition, this variant is not reported in gnomAD and it is classified as LPV according to the ClinGen Hereditary Breast, Ovarian and Pancreatic Cancer Expert Panel (PVS1 and PM2).10 13 Several digestive cancers are present in the familial history of patient 9: one FDR and three SDRs. Patient 9 had conserved IHC for the MMR proteins. Similarly, medical records revealed that the IHC of the two SDRs diagnosed with colorectal cancer were also normal. Cosegregation studies were also performed (online supplemental figure S9).
We have identified the variant NM_002528.7:c.244C>T in NTHL1 in homozygosis (VAF=0.997) in patient 10, diagnosed with colorectal cancer at age 45–50 years (online supplemental figure S10). This nonsense single base substitution in exon 2 of the NTHL1 gene results in a premature stop codon in the protein, which could lead to a truncated or absent protein by NMD ((p.Gln82*)). This variant is classified as a PV according to the guidelines (PVS1, PP5 and PM2).10 23 Patient 10 was diagnosed with colorectal cancer with conserved IHC for MMR proteins. Gastroscopy and colonoscopy revealed that this patient presented 6–7 gastric polyps and 10 colonic polyps. The familial history of cancer of the proband includes only one SDR. Carrier studies are being performed on the siblings and sons.
We report here that 10.53% of our population harbours a PV/LPV in a gene consistently associated with their diagnosed cancer (online supplemental appendix S1, table 1). Interestingly, we report that opportunistic genetic screening has increased the diagnostic yield by 1.22% in our cohort. In addition, 70.20% of the identified IFs are present in clinically actionable genes providing these families with an opportunity to join cancer early detection programmes, as well as secondary cancer prevention. IFs might facilitate the diagnosis of asymptomatic individuals and the early management of cancer once it develops.
Conventionally, in up to 50% of families with HBOC a deleterious germline alteration is found in high-penetrance genes (BRCA1, BRCA2 and PALB2) or moderate-penetrance genes (ATM and CHEK2 in BC and RAD51C, RAD51D and BRIP1 in ovarian cancer).27 However, the development of NGS tools has allowed for the identification of PVs/LPVs in other genes also related to BC. In the present work, we identified IFs in the TP53, MSH6, PMS2 and VHL genes in patients undergoing HBOC panel testing. Our group has previously reported that 8% of patients with HBOC criteria not harbouring PVs/LPVs in the BRCA genes were carriers of PVs/LPVs in BARD1, BRIP1, CDH1, CHEK2, PALB2, RAD50 or TP53.28 Hauke et al studied a cohort of non-carriers in the BRCAs genes with a familial or personal history of BC and found that the diagnostic yield increased by 1.66%, mainly in the ATM gene.29 Maani et al found 24 IFs in 6060 MGP-NGS performed (0.39%) in patients on suspicion of a hereditary cancer syndrome harbouring multiple PVs with a low allele fraction.1 In HBOC, hereditary non-polyposis colorectal cancer, Li-Fraumeni syndrome or adenomatous polyposis, extending the genetic analysis to 24 genes beyond the patients’ phenotype increased the diagnostic yield by 2.07%. Interestingly, these authors observed that opportunistic screening of BRCA1, BRCA2, MSH1, MSH2 and MSH6 increased the diagnostic yield by only a modest 0.58% compared with 2.07% when the full 24-gene panel was analysed.30
TP53 is included in panel testing when BC is diagnosed in patients aged ≤45 years or those fulfilling the Chompret criteria,31 which was not the case for either of patients 1 or 2 in our study. The variant NM_000546.6:c.783-1G>A in TP53 identified in our study has previously been found in patients with adrenocortical carcinoma, ovarian cancer and BC,32 agreeing with our results. However, no evidence has been published regarding the role of NM_000546.6:c.743G>A variant identified in our study in HBOC patients. As reviewed by Batalini et al, distinguishing a germline from a somatic PV/LPV in TP53 has enormous clinical implications.33 Clonal haematopoiesis (CH) is a common phenomenon in TP53 consisting of abnormal expansion of a haemopoietic stem cell clone harbouring a somatic variant that provides these cells with survival advantages. In contrast, in classic mosaicism, the mutation occurs at a postzygotic stage so it is present only in those cells derived from this affected cell.33 The presence of NM_000546.6:c.783-1G>A in a relative of patient two allowed us to confirm its germline origin. However, the absence of the variant NM_000546.6:c.743G>A in the relatives of patient 1 (online supplemental figure S1) allowed us to confirm that this PV in TP53 does not have a germline origin. The presence of a PV in non-tumorous tissue (such as fibroblasts from skin punch biopsy and hair follicles from eyebrows) may discard the diagnosis of CH and confirm the diagnosis of classic mosaicism. We found that the variant NM_000546.6:c.743G>A was present in the buccal swab (VAF≈20%), breast tumorous and non-tumorous paraffined tissue (VAF≈50%) but not in fibroblasts from skin punch biopsy, allowing us to confirm the diagnosis of classic mosaicism in patient 1.
PVs/LPVs in the MSH6 and PMS2 genes confer an increased risk throughout the life of Lynch syndrome and BC.34 35 Dominguez-Valentin et al found that BC risks were similar across all MMR genes.35 Interestingly, Roberts et al reported that PV in only MSH6 and PMS2 increased the cumulative incidence of BC.34 This evidence endorses the inclusion of these MMR genes in genetic tests undertaken in patients with a personal or familial history of BC. Concordantly, the Catalan Health Service recommends offering opportunistic screening for the MSH6 gene, among others, in all hereditary MGP-NGS tests performed.8 The variant NM_000535.7:c.24-2A>G in PMS2 identified in our cohort has been previously reported in homozygosis in a Spanish woman diagnosed with lymphoma, endometrial cancer and colorectal cancer.35 The variant NM_000535.7:c.1579_1580del in PMS2 has previously been identified in patients with BC and ovarian cancer.24 25 The variant NM_000535.7:c.989-2A>G in PMS2 has been previously identified in patients with Lynch syndrome36 and in a subject affected by two BC.19 Conflicting results aroused when IHC and genetic studies of family members of patient 6 were investigated. Although the family history of cancer on the paternal side may lead us to think that the variant NM_000535.7:c.989-2A>G in PMS2 was inherited from the paternal side, we cannot assure it. Taking this into consideration, we can neither assure nor refuse that subjects II:2 and III:1 are carriers of the same variant in PMS2. Some explanations for the discordances found in these subjects include that subject III:1 could be a carrier of a different MMR variant of somatic origin or germline origin, inherited from her paternal side and, therefore, not familiarly connected with pour proband (III:9). Sanger sequencing of the tumour or blood sample of subject III:1 was not possible because we had no access to the sample, but only to the electronic medical records stating the lack of expression of PMS2 and MSH6 proteins. Regarding the subject II:2, we ignore if she was carrier of the variant in PMS2. If so, one could expect an IHC with lack of expression of the proteins PMS2 and/or MSH6 in the colorectal tissue, which was not the case. IHC screening of Lynch syndrome is usually done for only MSH6 and PMS2 proteins (two-stains method) to reduce costs. However, as reported by Pearlman et al37 and recently reviewed by Leclerc et al,38 it is recommended to test the four MMR proteins to diagnose Lynch syndrome, since the two-stain immunohistochemical screening may fail to detect mismatch repair deficiency in some Lynch syndrome tumours. Screening for MMR protein expression by IHC is a standard clinical practice to identify patients with Lynch syndrome but not in those diagnosed with BC. However, several authors have demonstrated that BC in women with Lynch syndrome is more prone to exhibit loss of MMR protein expression compared with sporadic BC.34 Patient 4 with BC harbouring a PV/LPV in PMS2 presented conserved MMR protein expression in the tumorous tissue, while patient 5 showed a lack of PMS2 protein expression.
Patients with Von Hippel-Lindau disease develop hemangioblastomas in the central nervous system and retina, renal clear cell carcinomas, renal and pancreatic cysts and pheochromocytomas, and it is caused by PVs/LPVs in the VHL gene.39 To our knowledge, this is the first time that the variant NM_000551.4:c.341G>A in VHL has been reported (subject 7, diagnosed with BC).
Three IFs (two in ATM and one in NTHL1) were found in patients with a personal history of colorectal cancer, all of whom underwent Lynch syndrome MGP-NGS testing. Lynch syndrome predisposes patients to colorectal and endometrial cancer, among others.40 It is mainly caused by germline PVs/LPVs in the MMR genes (MLH1, MSH2, MSH6 and PMS2) and the EPCAM gene and, as a consequence, tumours from patients with Lynch syndrome display loss of expression of MMR proteins, microsatellite instability and increased hypermutation phenotype.38 40 41 With this evidence, screening for Lynch syndrome includes tumour IHC of MMR proteins and/or microsatellite instability analysis in all colorectal and endometrial cancers. According to the Spanish Society Of Medical Oncology (SEOM) guidelines, genetic studies must be undertaken when MSH2, MSH6 or PMS2 protein expression is affected. Moreover, if IHC analysis shows an absence of MLH1 staining, and BRAF/MLH1 methylation testing is normal, germline MMR testing is recommended.8 40 The variant NM_000051.4:c.5908C>T in ATM (patient 8) has been identified in patients with ataxia-telangiectasia in a homozygous and compound heterozygous state,42 but the variant NM_000051.4:c.7670_7674del 9 in ATM (patient 9) has never been reported. Carrier studies performed in this family led us to believe that this latter variant does not segregate with the disease and thus this variant has low penetrance. We also identified the variant NM_002528.7:c.244C>T in NTHL1 in homozygosis in a patient suffering from colorectal cancer. In the literature, this mutation has been reported in homozygosis or compound heterozygosity in several individuals affected with colorectal cancer, colorectal adenomas or polyposis.38 43 44
The identification of IFs can provide clinical benefits for patients harbouring the variant and their relatives. However, a major concern is how to integrate the identified IFs into routine clinical settings, as variants non-amenable to medical interventions could emerge.4–6 The European Society of Human Genetics recommends focusing on the identification of the underlying cause of a particular disease excluding deliberate searches for additional variants.45 In contrast, the ACMG endorses the intended and routine searching for variants not consistent with personal and family histories in cancer-related genes with known clinical actionability (online supplemental table S2).3 7 Nevertheless, the ACMG states that the option ‘not to know’ must be offered to those subjects willing to be informed only of those findings associated with the initial indication, according to the principle of autonomy. Some other societies have adhered to these ACMG recommendations, with few modifications. The FSPPM compiled a curated list of 60 cancer-related genes (online supplemental table S2) according to their risk and clinical actionability. Remarkably, the FSPPM considers a two-step informed consent: a first one about opportunistic genetic screening during the initial medical procedure and a second one once primary findings are disclosed to the patient, when the patient is more likely to differentiate the extent of the information to be revealed from opportunistic genetic screening.26
In the present work, 8 of the 11 IFs were identified in genes considered actionable from a clinical perspective by the ACMG and the FSPPM statements (MSH6, PMS2, TP53 and VHL). Nevertheless, the SEOM, the European Society for Medical Oncology (ESMO) and the National Comprehensive Cancer Network (NCCN) have published guidelines for the screening and management of ATM carriers and the SEOM has also published actionability recommendations for NTHL1 carriers (online supplemental table S4).27 40 46–50
MSH6 and PMS2 have similar actionability according to the SEOM, ESMO and NCNN guidelines, which include colonoscopies on an annual/biennial basis (online supplemental table S4).48 No available evidence exists regarding prophylactic colectomy in healthy individuals diagnosed with Lynch syndrome.40 Regarding gastric cancer surveillance in these patients, the ESMO recommends the surveillance with upper endoscopy in families with a history of gastric neoplasms and testing for Helicobacter pylori.51 The NCCN recommends upper gastrointestinal surveillance with esophagogastroduodenoscopy and colonoscopy, and biopsy to assess for H. pylori.51 Transvaginal ultrasound and endometrial aspirate are also recommended by the SEOM in ovarian and endometrial surveillance,47 in contrast to the NCCN guidelines, which recommend endometrial biopsy.48 Risk-reducing hysterectomy and salpingo-oophorectomy should be contemplated according to the SEOM, ESMO and NCCN recommendations.27 40 47
In subjects harbouring a PV/LPV in TP53, clinical breast examination (NCCN) and breast MRI (ESMO and NCCN) and screening for additional cancers should be performed.27 51 According to the ESMO and NCCN, risk-reducing mastectomy counselling should consider the degree of protection, reconstruction options, family history and residual BC risk with age and life expectancy.47 49 The European Reference Network-Genetic Tumour Risk Syndromes (GENTURIS) published a surveillance protocol specific for carriers of germline disease-causing TP53 variants, as detailed in online supplemental table S4.52
The Danish Coordination Group for VHL published guidelines for the diagnosis and surveillance of VHL carriers, which include annual clinical examination by a paediatrician of those aged 0–14 years, among others.53
In conclusion, our study showed that 1.22% of our cohort harboured PVs/LPVs in genes beyond those specifically related to the diagnosed cancer. Despite few of these IFs being present in low or moderate penetrance, we found that the majority are considered clinically actionable, providing these families with an opportunity to join cancer early detection programmes, as well as secondary cancer prevention. Moreover, these findings could facilitate the diagnosis of asymptomatic individuals and the early management of cancer once it develops. The actionability of these genes and their implications for the subjects’ and relatives’ care reinforce opportunistic genetic screening in all subjects undergoing genetic testing within the framework of genetic counselling.
Data availability statement
Data are available on reasonable request.
Patient consent for publication
This study was approved by the Clinical Research Ethics Committee of Sant Joan University Hospital, Reus, Spain (reference 134/2018). Written informed consent was obtained from all participants.
We would like to acknowledge the technicians from the Biobanc-IISPV in Reus (http://www.iispv.cat) for sample management. We are grateful to all the patients and families participating in this study.
SF-C and BR contributed equally.
Contributors BR, SF-C and MR-B formulated the research goals and aims of the present work and curated the data included in the work. JBo, MR-B and JG managed and coordinated the research activity planning and execution. JBo and JG were in charge of the acquisition of financial support for the project and was the supervisor of all of the research activities performed. BR, SF-C, MM, SS, MS, MQ, JB, MP, DP, JBa, JBo, MR-B and JG were involved in conducting the investigation process by both collecting and analysing data. BR and MR-B validated and reproduced the outcomes presented here. AC was in charge of the fibroblast experiments. BR, SF-C and MR-B were in charge of the preparation and creation of the paper. SF-C wrote the initial draft, and BR and MR-B comprehensively reviewed it. A second and final revision was performed by MM, SS, MS, MQ, JBa, MP, AC, DP, JBa, JBo and JG.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.