Variants of Uncertain Significance in BRCA Assessment
Variants of Uncertain Significance in BRCA Assessment
Background Nearly 15% of BRCA1 and BRCA2 DNA tests lead to the identification of Variants of Uncertain Significance (VUS). VUS are classified in the Netherlands according to the Bell system and it is current practice that class III VUS are communicated to counsellees, but not class II or lower VUS. Our aims were to investigate the utility of in silico characteristics in the classification of VUS and whether initial VUS classifications justify differences in communication protocols during counselling.
Methods We classified 88 missense VUS in BRCA1 and BRCA2 on the basis of an in silico analysis and compared the classification of a subset of 60 VUS of which additional information including family, genetic and tumour data was available.
Results VUS allocated to class III more frequently showed in silico indications of a deleterious effect than class II VUS. Of the 46 VUS assigned to class II by in silico analysis alone, nearly half were eventually recategorised as class I and 10% as class III when additional information was included.
Conclusions As in silico analysis alone is not always sufficient to unambiguously assign VUS to either class II or class III, we would argue that the prospect of obtaining additional information from a family should be given more weight during the decision process preceding the communication of a VUS test result. Research initiatives such as the Evidence-based Network for the Interpretation of Germline Mutant Alleles (ENIGMA), which strive to combine diverse sources of information, will be valuable in aiding a definitive classification of a VUS.
The ongoing development of sequence-based technologies in DNA diagnostic laboratories is resulting in the detection of an increasing number of variants of unknown clinical significance. These variants, referred to as Variants of Uncertain Significance (VUS), include missense changes, small in-frame deletions or insertions, non-synonymous nucleotide substitutions, as well as alterations in non-coding sequences or in untranslated regions.
Around 15% of DNA tests of the BRCA1 and BRCA2 genes result in the identification of VUS, and almost 1800 unique VUS are currently listed in the Breast Cancer Information Core database (http://research.nhgri.nih.gov/bic/) (accessed 4 Apr 2012).
In the Netherlands, over 1800 families are now known to carry a BRCA1 and BRCA2 VUS (National working group for Breast Cancer DNA Diagnostics (LOB)). These families experience considerable psychological distress, due to the possibility that they may face a cancer risk as high as that for known pathogenic mutations, and due to the uncertainty surrounding this risk.
Interpretation of VUS with respect to predicted effect on protein function, and thus on the estimated cancer risk in the families, has become a major challenge when tailoring genetic counselling and disease prevention strategies. As genetic counsellors need to be able to communicate a meaningful VUS DNA test outcome and possible consequences in a careful and understandable way to the counsellees and their families, it is essential that specialists in DNA diagnostic laboratories give a clear and objective estimation of the probability of pathogenicity for each VUS.
A variety of methods have been developed to determine whether a given variant is pathogenic or is of little or no clinical significance. Functional studies assess the impact of genetic variants on the activity of the protein in vitro. Some methods measure a direct association of the variant with disease, and include cosegregation of the variant with disease in a family, family history, co-occurrence of the variant with pathogenic BRCA1 and BRCA2 mutations on the second allele and analysis of the tumour DNA (eg, loss of heterozygosity and array comparative genomic hybridisation analysis). In silico approaches predict the consequences of DNA sequence changes in an indirect manner based on evolutionary nucleotide and amino acid conservation, the possible effect of amino acid substitutions on protein structure or the predicted effect on mRNA (messenger RNA) splicing.
In 2007, the Dutch and British societies for clinical molecular genetics proposed 'Good Practice Guidelines for the Interpretation and Reporting of Unclassified Variants in Clinical Molecular Genetics Laboratories'. A four-class system was described, with increasing probability of pathogenicity (class I to IV). This was followed by a suggested classification into five groups (Table 1), by Plon et al in 2008.
The communication of a VUS to a counsellee often results in feelings of uncertainty, distress and a possible decision to undergo prophylactic surgery. As the prior probability that a VUS will be deleterious is less than 10%, laboratory personnel in the Netherlands show understandable reservations regarding the communication of the discovery of a VUS to the counsellor, as does the counsellor when communicating with the counsellee.
Each newly identified VUS is first categorised using in silico tools. Class II categorised VUS are communicated to the counsellors, but are not generally revealed to the counsellees. A class III VUS, which is more likely to be pathogenic, is communicated to the counsellees and if possible, additional studies are performed to obtain a more accurate assessment of pathogenicity (eg, cosegregation and RNA-analysis). Risk estimates and surveillance policies for class II and class III VUS are generally based on family cancer history, and predictive DNA testing is not offered to the family members. The distinction between class II and class III VUS is a frequent topic of debate in the Netherlands, and since allotment of a VUS to either class II or class III involves a distinct communication protocol during counselling, objective assessment of the VUS is crucial.
The aim of this study was to investigate whether VUS classified in class II and III by the LOB working group show significant differences in in silico characteristics, and thus whether current counselling protocols with respect to initial communication with the counsellees are justified.
Abstract and Introduction
Abstract
Background Nearly 15% of BRCA1 and BRCA2 DNA tests lead to the identification of Variants of Uncertain Significance (VUS). VUS are classified in the Netherlands according to the Bell system and it is current practice that class III VUS are communicated to counsellees, but not class II or lower VUS. Our aims were to investigate the utility of in silico characteristics in the classification of VUS and whether initial VUS classifications justify differences in communication protocols during counselling.
Methods We classified 88 missense VUS in BRCA1 and BRCA2 on the basis of an in silico analysis and compared the classification of a subset of 60 VUS of which additional information including family, genetic and tumour data was available.
Results VUS allocated to class III more frequently showed in silico indications of a deleterious effect than class II VUS. Of the 46 VUS assigned to class II by in silico analysis alone, nearly half were eventually recategorised as class I and 10% as class III when additional information was included.
Conclusions As in silico analysis alone is not always sufficient to unambiguously assign VUS to either class II or class III, we would argue that the prospect of obtaining additional information from a family should be given more weight during the decision process preceding the communication of a VUS test result. Research initiatives such as the Evidence-based Network for the Interpretation of Germline Mutant Alleles (ENIGMA), which strive to combine diverse sources of information, will be valuable in aiding a definitive classification of a VUS.
Introduction
The ongoing development of sequence-based technologies in DNA diagnostic laboratories is resulting in the detection of an increasing number of variants of unknown clinical significance. These variants, referred to as Variants of Uncertain Significance (VUS), include missense changes, small in-frame deletions or insertions, non-synonymous nucleotide substitutions, as well as alterations in non-coding sequences or in untranslated regions.
Around 15% of DNA tests of the BRCA1 and BRCA2 genes result in the identification of VUS, and almost 1800 unique VUS are currently listed in the Breast Cancer Information Core database (http://research.nhgri.nih.gov/bic/) (accessed 4 Apr 2012).
In the Netherlands, over 1800 families are now known to carry a BRCA1 and BRCA2 VUS (National working group for Breast Cancer DNA Diagnostics (LOB)). These families experience considerable psychological distress, due to the possibility that they may face a cancer risk as high as that for known pathogenic mutations, and due to the uncertainty surrounding this risk.
Interpretation of VUS with respect to predicted effect on protein function, and thus on the estimated cancer risk in the families, has become a major challenge when tailoring genetic counselling and disease prevention strategies. As genetic counsellors need to be able to communicate a meaningful VUS DNA test outcome and possible consequences in a careful and understandable way to the counsellees and their families, it is essential that specialists in DNA diagnostic laboratories give a clear and objective estimation of the probability of pathogenicity for each VUS.
A variety of methods have been developed to determine whether a given variant is pathogenic or is of little or no clinical significance. Functional studies assess the impact of genetic variants on the activity of the protein in vitro. Some methods measure a direct association of the variant with disease, and include cosegregation of the variant with disease in a family, family history, co-occurrence of the variant with pathogenic BRCA1 and BRCA2 mutations on the second allele and analysis of the tumour DNA (eg, loss of heterozygosity and array comparative genomic hybridisation analysis). In silico approaches predict the consequences of DNA sequence changes in an indirect manner based on evolutionary nucleotide and amino acid conservation, the possible effect of amino acid substitutions on protein structure or the predicted effect on mRNA (messenger RNA) splicing.
In 2007, the Dutch and British societies for clinical molecular genetics proposed 'Good Practice Guidelines for the Interpretation and Reporting of Unclassified Variants in Clinical Molecular Genetics Laboratories'. A four-class system was described, with increasing probability of pathogenicity (class I to IV). This was followed by a suggested classification into five groups (Table 1), by Plon et al in 2008.
The communication of a VUS to a counsellee often results in feelings of uncertainty, distress and a possible decision to undergo prophylactic surgery. As the prior probability that a VUS will be deleterious is less than 10%, laboratory personnel in the Netherlands show understandable reservations regarding the communication of the discovery of a VUS to the counsellor, as does the counsellor when communicating with the counsellee.
Each newly identified VUS is first categorised using in silico tools. Class II categorised VUS are communicated to the counsellors, but are not generally revealed to the counsellees. A class III VUS, which is more likely to be pathogenic, is communicated to the counsellees and if possible, additional studies are performed to obtain a more accurate assessment of pathogenicity (eg, cosegregation and RNA-analysis). Risk estimates and surveillance policies for class II and class III VUS are generally based on family cancer history, and predictive DNA testing is not offered to the family members. The distinction between class II and class III VUS is a frequent topic of debate in the Netherlands, and since allotment of a VUS to either class II or class III involves a distinct communication protocol during counselling, objective assessment of the VUS is crucial.
The aim of this study was to investigate whether VUS classified in class II and III by the LOB working group show significant differences in in silico characteristics, and thus whether current counselling protocols with respect to initial communication with the counsellees are justified.
