Article Text
PDF
Abstract
Background Constitutional mismatch repair deficiency syndrome (CMMRD) is the most aggressive cancer predisposition syndrome associated with multiorgan cancers, often presenting in childhood. There is variability in age and presentation of cancers and benign manifestations mimicking neurofibromatosis type 1. Genetic testing may not be informative and is complicated by pseudogenes associated with the most commonly associated gene, PMS2. To date, no diagnostic criteria exist. Since surveillance and immune-based therapies are available, establishing a CMMRD diagnosis is key to improve survival.
Methods In order to establish a robust diagnostic path, a multidisciplinary international working group, with representation from the two largest consortia (International Replication Repair Deficiency (IRRD) consortium and European Consortium Care for CMMRD (C4CMMRD)), was formed to establish diagnostic criteria based on expertise, literature review and consensus.
Results The working group established seven diagnostic criteria for the diagnosis of CMMRD, including four definitive criteria (strong evidence) and three likely diagnostic criteria (moderate evidence). All criteria warrant CMMRD surveillance. The criteria incorporate germline mismatch repair results, ancillary tests and clinical manifestation to determine a diagnosis. Hallmark cancers for CMMRD were defined by the working group after extensive literature review and consultation with the IRRD and C4CMMRD consortia.
Conclusions This position paper summarises the evidence and rationale to provide specific guidelines for CMMRD diagnosis, which necessitates appropriate surveillance and treatment.
- diagnosis
- heredity
- gastrointestinal diseases
- genetics
- medical
- medical oncology
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Statistics from Altmetric.com
Introduction
Constitutional mismatch repair deficiency syndrome (CMMRD), caused by biallelic pathogenic variants (PVs) in mismatch repair (MMR) genes, is the most aggressive and complex hereditary cancer syndrome with a very high incidence of cancers, resulting in patient death in early life. CMMRD can also present with benign manifestations resembling neurofibromatosis type 1 (NF1).1 While most malignancies present in childhood, there is variability in age and presentation of cancers and benign manifestations.
Of the four MMR genes, PMS2 is responsible for over 60% of CMMRD families, followed by MSH6 (20%–30%), MLH1 and MSH2 (10%–20%).2–4 Consanguinity has been reported in 39%–45% of CMMRD families, varying slightly depending on country of origin.3–6 Homozygous variants have been reported in 10%–28.5% of families that deny consanguinity suggesting common founder mutations, with higher frequency in isolated populations or those that may share a common ancestor.3 4 6 7
In contrast, Lynch syndrome (LS), an autosomal dominant condition, has an inverse distribution of gene mutations, with monoallelic PVs in MSH2 and MLH1 accounting for 80% of families, MSH6 in 13% and PMS2 in 6%.8 In addition, deletions of the terminal exons of the EPCAM gene that include 3’UTR sequences upstream from MSH2 cause methylation of MSH2 leading to an LS phenotype.9–12 Of note, the first homozygous EPCAM patient is described later in this paper. LS is traditionally characterised by a high incidence of endometrial and colorectal cancer (CRC) through generations of the family; however, PMS2 and MSH6 have a lower penetrance of CRC and endometrial cancer (10% and 12% in PMS2 and 20% and 40% in MSH6, respectively),8 resulting in lack of cancer family history.1 4 6
While LS is a fairly well-described condition, CMMRD is a newer and more complex condition with molecular testing complicated by the fact that the most commonly impacted gene, PMS2, can pose testing challenges due to pseudogenes and frequent gene conversion events that can be circumvented only by specialised assays.13 14 Variants of uncertain significance (VUS) can be difficult to interpret, and PMS2 has the highest incidence of VUS among the MMR genes, comprising 49% of reported variants in this gene.15
The clinical presentation of CMMRD is variable, typically beginning in childhood with brain, lymphoma or gastrointestinal (GI) adenocarcinomas; however, there is a broad spectrum of malignancies that can present at any age, including individuals who present without cancer and/or NF1 manifestations.
A scoring system to identify individuals who should undergo CMMRD genetic testing was developed by the European Consortium Care for CMMRD (C4CMMRD) in 2014, based on a calculated clinical score of malignancy and benign features (table 1).1 The C4CMMRD consortium also defined CMMRD counselling and testing criteria for children suspected to have sporadic NF1 but without a detectable mutation in NF1 or SPRED1 and without a malignancy.16
Scoring system to determine germline testing eligibility for CMMRD (Wimmer et al 1)
These scoring systems are an entry point for evaluation; however, no diagnostic criteria currently exist. Given the variable presentation of CMMRD and the lack of one definitive test for some cases, an international committee was convened to develop a robust pathway for genetic diagnosis combining molecular diagnosis, ancillary testing and clinical manifestation. The first established CMMRD diagnostic criteria are outlined in this paper. An accurate diagnosis has important implications for genetic counselling, tumour surveillance2 17 18 (table 2) and access to immunotherapy cancer treatment.19 20
Surveillance protocols for patients with CMMRD
Methods
In order to establish diagnostic criteria, an international working group was formed, including 13 experts, with representatives from the International Replication Repair Deficiency (IRRD) consortium, C4CMMRD, as well as other clinical and molecular geneticists, genetic counsellors, paediatric haematologist and oncologists, from Canada, USA, Israel and France.
To validate the recommendations, the clinical working group used data from the IRRD consortium, based in Toronto Canada at The Hospital for Sick Children, and from the C4CMMRD database, based in Villejuif, France at Gustave Roussy Cancer Campus. The IRRD consortium was established in 2007 to assess individuals suspected of having syndromes with replication repair deficiencies, including a large cohort of patients with CMMRD (n=110), originating from 45 countries. Updates on surveillance and health status are obtained at regular intervals. The C4CMMRD consortium was established in 2013, with one of its objectives to create a database dedicated to patients with CMMRD. This database contains retrospective and prospective data, from 15 countries, with approximately 100 patients with confirmed CMMRD.
Based on literature review and expertise from the panel, draft clinical criteria were established by the working group at an international IRRD workshop held in Toronto on 15–16 October 2017. The draft clinical criteria from the working group were presented to all workshop attendees, made up of 60 experts from 18 countries for input and consensus vote. The agreed upon criteria were further evaluated through an indepth survey sent to the 13 members of the working group and completed by 8 members of the group by 23 April 2018. If consensus of 80% was reached, the criterion was adopted. A follow-up conference call with all members was arranged to discuss criteria that did not reach 80% agreement for review to eliminate or revote. Hypothetical cases, as well as families with typical and atypical presentations of CMMRD from the IRRD consortium, were reviewed to test the draft criteria. A second survey to finalise data was completed by 28 February 2019 by all members of the working group. The focus of this working group was the development of CMMRD diagnostic criteria, thereby necessitating surveillance.
Results
CMMRD diagnostic criteria
The expert panel established seven diagnostic criteria: four criteria with strong evidence of CMMRD (ie, definitive diagnosis) and three criteria with moderate evidence (ie, likely diagnosis), as outlined in table 3. All criteria outlined warrant CMMRD surveillance. Three components were used to establish criteria, namely (1) MMR germline results, (2) ancillary testing and (3) clinical manifestations. Multigene panel testing is recommended to investigate overlapping conditions which can mimic CMMRD. Ancillary testing and clinical manifestation, including a new definition of CMMRD hallmark tumours, were explored in detail.
CMMRD diagnostic criteria
Table 3 outlines the criteria beginning with germline MMR results with diminishing evidence of molecular confirmation. Based on the strength of the molecular results for a diagnosis, additional evidence may be required using ancillary testing and/or clinical phenotype as defined in the other columns of the table. Even confirmation of biallelic PVs may not be diagnostic of CMMRD (eg, PV that cannot be confirmed in trans, or if PVs in unaffected adults). Using molecular results as the sole diagnostic criterion can be complicated by the interpretation of the PMS2 gene, which has a highly homologous pseudogene. As an example, several unaffected adults have been reported to have homozygous PMS2 c.2182_2184delACTinsG variants, which is classified as a PV in ClinVar, but is now thought to have been incorrectly assigned and actually occurring in the pseudogene, PMS2CL.21 The diagnostic criteria concludes with three ‘likely diagnostic’ criteria (criteria 5–7) that were established to capture individuals who may present with atypical cancers or at older ages.
The working group recommends that the diagnostic criteria be reassessed as data emerge and as the criteria are applied to newly identified families suspected of CMMRD. The committee did not establish criteria to eliminate the suspicion of CMMRD and recognises there may be families that meet a level of suspicion for this condition, but do not meet the strict evidence to confirm diagnosis. Clinical judgement should be used in all cases.
Rationale for criteria
MMR genes germline testing
The American College of Medical Genetics outlined five variant classifications for germline analysis of Mendelian disorders: pathogenic (P), likely pathogenic (LP), variants of unknown significance (VUS), likely benign (LB) and benign (B) variants.22 Actionable variants include P/LP, while a VUS should not be used for clinical decision-making and LB/B variants can be assumed not to cause phenotype.
Non-definitive MMR germline results have been described in multiple patients with CMMRD, including cases with biallelic VUS or monoallelic VUS in combination with a P/LP variant.4 23 24 This suggests that while a monoallelic VUS may not be clinically actionable or may lack evidence of pathogenicity, it may sufficiently impact protein function when in combination with a second VUS or P/LP variant. One example of this complexity was reported by Taeubner et al,25 who found both homozygous MSH2 and MSH6 variants in a 13-month-old with desmoplastic medulloblastoma and striking skin pigmentation using ancillary assays.26 27 It was concluded that CMMRD was caused by the MSH6 homozygous VUS.25
A review of the literature outlined suspected CMMRD families in the absence of definitive biallelic pathogenic MMR variants, or based solely on clinical phenotype and ancillary testing.3 4 24 28 A study by Bakry et al 28 reported MMR germline results for 12 CMMRD families, identifying 67% (n=8) with biallelic P/LP variants, 17% (n=2) with biallelic VUS, 8% (n=1) with monoallelic PV and 8% (n=1) with no variants identified. Similar distribution was seen in 38 patients with CMMRD with brain tumours from the C4CMMRD consortium, with 74% (n=28) biallelic P/LP variants, 18% (n=7) biallelic VUS, 8% (n=13) monoallelic PV and no monoallelic VUS.29
This highlights the need to use ancillary testing in the absence of confirmatory MMR germline testing mutation analysis.
Ancillary testing
Ancillary testing, summarised in table 4 and outlined in this section, can be used to support a diagnosis of CMMRD.30
Examples of ancillary tests available to assist in CMMRD diagnosis
Immunohistochemistry (IHC) of non-neoplastic tissue (ie, skin or normal colon biopsy or adjacent normal tissue to a cancer) is the most widely available and least expensive test to assess for deficiency of MMR protein (dMMR); however, interpretation is subjective and obtaining normal tissue can be invasive.28 Pathologists must be aware of the suspected CMMRD diagnosis as the positive stain which serves as an internal control in patients with LS is often absent in all tissues of individuals with CMMRD. In addition, pathogenic missense variants, frequent in CMMRD, can result in a non-functional protein staining positive, thereby leading to a misleading IHC positive result.31 Despite limitations, IHC of non-neoplastic tissue has been reported to have nearly 100% specificity and >90% sensitivity in experienced pathology laboratories and can be a useful tool in the diagnosis of CMMRD.28 29
Functional assays have been developed, although widespread access at commercial laboratories may be limited. While conventional Microsatellite instability (MSI) testing using an National Cancer Institute (NCI) panel of mononucleotide and dinucleotide repeat markers was validated for adult-onset LS cancers,32 it is not sensitive in detecting MSI in CMMRD. A method to detect MSI in non-neoplastic tissue, termed germline MSI, has been developed, and this assay relies on the analysis of ‘stutter’ peaks typically associated with microsatellite PCR products.26 Its main limitation is that it uses dinucleotide microsatellites and therefore is insensitive to an MSH6 deficiency. It is routinely used for diagnosis purposes in some European countries and allows for rapid results.
Another functional assay includes a combination of tolerance to methylating agents (a characteristic of MMR-deficient cells) and MSI from lymphoblastoid cell lines called ex vivo MSI (evMSI).27 A presumptive diagnosis of CMMRD using this assay requires both tolerance to methylating agents and MSI to be concordant. A diagnosis of CMMRD can be ruled out if both tests show normal activity. The method is 100% sensitive and 100% specific. A limitation of this test is the timing necessary for immortalisation and culture cells for evMSI (at least 120 days). This assay is available in routine diagnosis in France in an accredited laboratory, and since 2015, 77 patients have been tested prospectively with the two functional assays and all 15 patients with proven or highly probable CMMRD had both abnormal results consistent with a diagnosis of CMMRD (M Muleris and C Colas, personal communication in 2020). These assays are not easily accessible clinically in North America.
An in vitro repair assay was developed to quantify MMR activity from patient-derived lymphoblastoid cell lines.33 Testing a series of patients with CMMRD, LS, NF1, Li-Fraumeni syndrome (LFS) and polymerase proofreading-associated polyposis syndrome, the assay demonstrated high specificity and sensitivity. The assay requires live cell cultures as MMR proteins are maximally produced during cell division, which limits its scalability. In addition, the assay is a complementation assay which allows determination of the defective protein complex (MSH2-MSH6 or MLH1-PMS2). This assay is not yet established as a clinically approved test.
More recently next-generation sequencing (NGS)-based methods have proven to be very sensitive and specific in diagnosing CMMRD as they can detect low levels of MSI from constitutional tissue.34 35 NGS-based methods for MSI detection have the advantage of being easily scalable, cost-effective and without the need to grow lymphoblastoid cell lines.
Paediatric tumours showing high tumour mutation burden (TMB) (>10 mutations/Mb) and MMR-deficient signatures may raise the suspicion of CMMRD.36 CMMRD tumours with associated somatic mutations in POLE/POLD1 can also lead to ultrahypermutation phenotypes (>100 mutations/Mb).37 Unfortunately, TMB and tumour signature cannot clearly differentiate between LS, somatic biallelic MMR and CMMRD, and are insufficient as a sole diagnostic tool for CMMRD. It is also noted dMMR in a malignancy is not sufficient to confirm a diagnosis of CMMRD as this feature is characteristic of LS or sporadic biallelic MMR mutations.37 38 The working group does not support using TMB, tumour signature or dMMR in malignancy alone as ancillary tests to confirm CMMRD; however, they can be used to support a diagnosis once outlined diagnostic criteria are met.
The working group recommends using ancillary testing in challenging cases and should be used to interpret atypical presentation of CMMRD or inconclusive germline molecular results within the clinical context. If a discrepancy occurs among tests, additional ancillary tests, preferably by orthogonal methods, should be performed to reach a more conclusive decision. Consideration may be given to implement tests that are already published with high sensitivity and specificity in accredited (eg, College of American Pathologist (CAP)-inspected) laboratories authorised to give a clinically usable report. The working group gives a framework of current ancillary tests; however, a definitive list was not incorporated as new functional assays may be developed over time. A summary of the current tests outlining pros and cons is provided in table 4.
Clinical presentation
Cancer type and age of presentation
In order to determine the hallmark cancers as well as atypical presentations, the working group reviewed the literature and consulted the IRRD and C4CMMRD databases. Results on cancer site and age of onset are summarised in table 5. It reveals that 86%–100% of individuals with CMMRD have reported a cancer diagnosis.3 4 39 40 This may be a result of ascertainment bias because cancer is the most common criterion leading to the diagnosis of CMMRD in index cases.
Age and site of first malignancy as reported in large CMMRD series
Wimmer et al 1 outlined the most common CMMRD tumours to include WHO grade III/IV glioma <25 years, non-Hodgkin’s lymphoma of T cell lineage <18 years, GI adenocarcinoma <25 years or GI adenomatous polyposis <18 years. Our review determined that 60%–84% of CMMRD present with these cancers (table 5).3 4 39 40 By expanding the criteria to include primary cancers from all subtypes of brain, cerebral and haematological malignancies, an additional 18.75% (18 of 96) and 34% (31 of 90) of patients from the IRRD and C4CMMRD databases would be included. The contribution of GI cancers as the primary diagnosis accounted for 15% (15 of 96) of IRRD cases and 18% (16 of 90) of C4CMMRD cases, most often reported as colorectum or small bowel cancer. Adenomatous polyposis is also reported as the first diagnosis in patients with CMMRD, often diagnosed under the age of 18.4 28 41 Based on expertise and literature review, this working group defined new ‘hallmark tumours’ in CMMRD as (1) glioma or Central nervous system (CNS) embryonal tumours <25 years, (2) haematological cancer (excluding Hodgkin’s lymphoma) <18 years, (3) GI adenocarcinoma <25 years and (4) GI adenomatous polyposis (>10 adenomas) <18 years (after ruling out polyposis conditions, defined in the rational for criteria: differential diagnosis section). Using this definition of hallmark tumours, 76%–96% of CMMRD had one of these primary diagnoses (table 5). It is noted that these hallmark tumours are not exclusively associated with CMMRD, and the diagnostic criteria outlined in table 3 require molecular test confirmation and ancillary testing to establish a diagnosis.
While the majority of CMMRD cases present with a hallmark tumour, there is a broader spectrum of less common primary and metachronous malignancies. These include embryonal tumours (eg, neuroblastoma, retinoblastoma, Wilms’ tumour), germ cell tumours (eg, yolk sac tumour), sarcomas (eg, osteosarcoma, dermatofibrosarcoma protuberans, rhabdomyosarcoma), ganglioneuroma, melanoma, urinary tract (bladder, kidney, ureter), prostate, breast, gynaecological (endometrial, ovarian) and other GI (ampullary, gastric) cancers.1 3 4 24 28 40–44
Onset of first cancer during adulthood has been noted in CMMRD (table 5). Using unpublished data from the two largest consortia (IRRD and C4CMMRD, respectively), 11%–15.6% of cases presented at age 18 or older and 1%–7% at age 25 or older. The oldest age at first malignancy reported in CMMRD was rectal cancer at age 40 in an individual with in trans P/LP PMS2 variants with dMMR in tumour and non-neoplastic tissue.44 Cases of atypical cancer at older age of onset have been reported as the primary cancer in individuals with breast cancer over the age of 25.42 45 Older ages of onset of first cancer appear to be more commonly associated with epithelial cancers, although older-onset lymphoma (diagnosed at age 33) has also been reported.46 Genotype–phenotype correlations may explain some of these older-onset cases as described by Li et al,7 who identified a founder homozygous PV of PMS2 in Inuit families displaying an attenuated phenotype due to residual PMS2 expression. These clinically unusual cases highlight the spectrum of CMMRD presentation and the challenges with establishing a diagnosis based on clinical presentation alone.
Benign manifestations
NF1 features described in CMMRD include café-au-lait macules (CALM), neurofibromas and axillary freckling. Only a minority of individuals with CMMRD meet the NIH diagnostic criteria for NF1, although 18.5% (n=27/146) have more than one NF1 feature and between 62% and 95% have CALM, although not all reach the threshold of ≥6 CALM.1 4 47 CALM associated with CMMRD tends to have irregular, jagged borders and varies in pigmentation, which may be differentiated from CALM in NF1 which is more likely to have smooth borders and uniform pigmentation.47 48
CALM is common in the general population. While largely underestimated, one study reports the incidence to be 0.3%–1.8% of white individuals between birth and childhood years and in 18%–27% within that age range in individuals who identify as black.48 In addition to CMMRD and NF1, CALM can be associated with other hereditary conditions including Legius syndrome (SPRED1 gene), Noonan syndrome (occurring with multiple lentigines), other RASopathies and McCune-Albright syndrome. Hypochromic spots, not present in NF1, have been described from 16% (Lavoine et al 4) to 29% (unpublished C4CMMRD data) of patients with CMMRD. If suspicious of CMMRD, a complete and careful skin examination must be performed and may provide elements in favour of the diagnosis.
Other clinical features reported in patients with CMMRD include agenesis of the corpus callosum, grey matter heterotopia,49 venous anomalies,50 multiple pilomatrixoma,51 paediatric systemic lupus erythematosus,52 intracranial tuber-like lesions and renal angiomyolipoma.53 Decreased IgA and/or IgG2/4 levels1 have also been reported; however, this was not substantiated in a recent study.54 While these manifestations have been observed, they are not routinely investigated at the time of CMMRD diagnosis. Examinations of these benign manifestations can be of value, especially when diagnosis is doubtful, as they may provide arguments in favour of the diagnosis.
CMMRD is a legitimate differential diagnosis in children suspected to have sporadic NF1 in whom no NF1 or SPRED1 mutation is identified. Suerink et al 16 and the C4CMMRD (2018) estimated the prevalence of CMMRD in these children at 0.39%, and developed criteria to preselect those children where CMMRD is more likely, including (1) the presence of at least one diagnostic NF1 feature including at least two CALMs; (2) absence of NF1 and SPRED1 germline mutations; and (3) absence of NF1 signs in both parents plus at least one or more feature suggestive of CMMRD (ie, LS diagnosis or cancer in family, NF1 feature or childhood cancer in sibling, atypical skin lesions outlined by the C4CMMRD).
Differential diagnoses
The differential diagnosis of CMMRD includes a wide range of other hereditary tumour conditions. LS and CMMRD are both caused by MMR mutations, and while they phenotypically differ, atypical cases can mimic each other. There have been rare cases of LS with glioblastoma or GI cancers diagnosed in childhood or young adulthood, and CMMRD cases presenting with adult-onset cancer.55–57 Polyposis conditions can also mimic CMMRD, including individuals with specific POLE variants (ie, p.V411L or p.A456P) who present with childhood hypermutant brain and CRC.58 59 Conversely, individuals with CMMRD have also presented with adenomatous polyposis (ie, 10–100 adenomas) mimicking conditions such as familial adenomatous polyposis (FAP or attenuated FAP) and MUTYH-associated polyposis. There is also a rare case presenting with ≥30 juvenile polyps typically associated with juvenile polyposis syndrome.60 Early-onset malignancies can also overlap with LFS, caused by germline TP53 mutations, which increases the risk of haematological, sarcoma, brain and GI cancers presenting at early ages.61
There has also been an overlap with conditions that have CALM, skin manifestation such as NF1 and Legius syndrome,47 62 and tuberous sclerosis.53 Multigene panel testing is recommended to assess for hereditary CRC/polyps conditions, childhood malignancy conditions (ie, TP53) or syndromes based on skin findings (ie, NF1, Legius) according to clinical presentation.
Other considerations
Full siblings
Full siblings of kin who meet one of the seven diagnostic criteria outlined in table 3 should undergo CMMRD surveillance unless CMMRD is ruled out by one of the following (note: LS diagnosis may still be possible):
Criteria 1, 2 or 5: if siblings did not inherit one of the familial variants as identified in the index case.
If sibling(s) reaches the age of 30 years with no history of colorectal polyps or CMMRD-related cancer.
If sibling has negative ancillary testing (when index case has proven positive ancillary test).
EPCAM
To date, there is one known case of homozygous EPCAM 3’UTR deletions, unpublished and followed by the IRRD consortium. The 5-year-old has colorectal polyposis, small bowel adenomas and one CALM. There has also been a 9-year-old child reported to have compound heterozygous MSH2 PV and EPCAM 3’UTR deletion (in trans), with 70–80 colorectal adenomas and 2 synchronous CRCs.63 At this time, there is not enough clinical information to determine the impact of biallelic EPCAM 3’ UTR deletions on cancer risks and whether its phenotype is limited to the GI tract or if it extends to the other hallmark cancers of CMMRD. This may be also dependent on the involvement of the MSH2 gene, as homozygous deletions that extend into MSH2 would be suspected to cause CMMRD. The working group recommends CMMRD surveillance for biallelic EPCAM 3’ UTR deletions or individuals with EPCAM/MSH2, until further information becomes available.
MSH3
There have been few reports of individuals with biallelic MSH3 described as having adult-onset adenomatous polyposis of the colon.64 Thyroid and small bowel adenomas and astrocytoma have also been reported, all occurring in young adulthood.64 It is unclear whether individuals with biallelic MSH3 require the same intensive screening as patients with CMMRD, and our panel concluded that insufficient evidence exists at this time to determine if cancer risk warrants CMMRD surveillance.
Conclusions
This is the first consensus on the diagnostic criteria for CMMRD which was established by a working group of international experts, including members of the IRRD and the C4CMMRD consortia who are involved in the management of most individuals with CMMRD. The seven diagnostic criteria described allow for clear surveillance recommendations and familial genetic counselling.
CMMRD is a condition with a variable clinical presentation encompassing a broad spectrum of cancers that may mimic other hereditary conditions. This working group defined a set of hallmark cancers, while recognising there is variability in presentation of cancer site and age. Benign manifestations are broad, may be non-specific and can mimic other hereditary conditions. Molecular testing is unable to conclusively identify biallelic pathogenic MMR gene variants in approximately 40% of patients. The diagnostic criteria outlined are essential given the complexity of the syndrome. As information continues to evolve, the diagnostic criteria and surveillance recommendations should be periodically reassessed.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Acknowledgments
The authors would like to acknowledge the support of Dr Melissa Edward and Dr Vanessa Bianchi for facilitating the initial workshop and overseeing the IRRD consortium.
References
Footnotes
Twitter @kamiws
Contributors All authors contributed to planning, conduct and reporting of this work.
Funding The International Replication Repair Deficiency (IRRD) consortium is partially supported by grants from Stand Up to Cancer (SU2C-AACR-CT07-17), Canadian Institutes of Health Research (CIHR) (#PJT-156006 and 108188-001 as part of Joint Canada-Israel Health Research Program) and Meagan’s Walk (MW-2014-10). The C4CMMRD database is supported by La Fondation Gustave Roussy, Guérir le cancer de l’Enfant au 21ème siècle. These agencies had no involvement in the study design, analysis or interpretation of data, writing or decision to submit for publication.
Competing interests HH reports the following financial relationships: Scientific Advisory Board at Invitae Genetics, Medical Advisory Board at Promega, Scientific Advisory Board at Genome Medical, and consultant at 23andMe. None of the relationships presents a conflict of interest in this study. KJ is a full-time employee at Ambry Genetics. This does not present a conflict of interest in this study. LB reports grants from Fondation Gustave Roussy, during the conduct of the study. UT reports grants from SU2C Catalyst, grants from Meagan’s Walk, grants from IDRC - Joint Canada-Israel Health Program and grants from Canadian Institutes of Health, during the conduct of the study.
Provenance and peer review Not commissioned; externally peer reviewed.