MMR Workshop

MINUTES - InSiGHT MMR Variant Interpretation Workshop

By Maija Kohonen-Corish
Venue: Hyatt Regency Hill Country, San Antonio, Texas
Date: 30 March 2011

Participants:


This Workshop was held prior to the 4th Biennial Meeting of the International Society for Gastrointestinal Hereditary Tumours (InSiGHT) and was convened by Maurizio Genuardi and Finlay Macrae. The participants included members of the InSiGHT Committee for DNA Mismatch Repair (MMR) Variant Interpretation and other interested individuals.

Finlay Macrae welcomed the participants and explained that the objectives of this Workshop were to reach a consensus on a set of qualitative criteria for the classification of MMR gene sequence variants (1) as a basis for this InSiGHT committee to classify variants of uncertain significance (VUS) on the InSiGHT Colon Cancer Gene Variant Databases, (2) to inform the development of quantitative estimates and (3) to provide gene-specific criteria for diagnostic laboratories to assist in their internal classifications.

With the appointment of John-Paul Plazzer as the new full-time Curator of the InSiGHT Databases, these criteria can be incorporated into the gene variant data increasing their clinical utility. John-Paul is based in Melbourne, Australia, and liaises closely with Johan den Dunnen, the developer of the Leiden Open Variation Database (LOVD) platform, which hosts the InSiGHT Databases. He has travelled widely to meet the key data submitters, including the curators of two of the original MMR gene variant databases (Rolf Sijmons, and Michael Woods) and to encourage variant submissions from diagnostic and research laboratories. John-Paul Plazzer is currently funded by the Melbourne-based philanthropic Hicks Foundation until September 2012. Partial funding to support John-Paul's position has been included in the International Mismatch Repair Consortium RO1 Penetrance grant (CIs Haile, Jenkins and Macrae) submitted in February 2011.

The first part of the meeting discussed the rules to classify variants qualitatively, based largely on the work of Amanda Spurdle and Bryony Thompson, with improvements suggested by other participants of the Workshop. As a starting point to the discussion, Maurizio Genuardi presented the American College of Medical Genetics (ACMG) recommendations, which aim to standardize the reporting and interpretation of genetic tests and cover sequence variations for all disease susceptibility genes (Genet Med 2008; 10:294-300). These are based on a 6-tier classification ranging from Class 1, that represents previously reported and recognized causes of the disorder, to Class 5 for previously reported and recognized neutral variants. Class 6 includes the variants, which are not known or expected to be causative of disease, but are found to be associated with a clinical presentation in case-control studies. ACMG suggests a number of follow-up studies (i.e., segregation analysis, functional and in silico studies, etc.) that can be helpful for the purpose of obtaining more accurate classification.

The meeting discussed the ACMG classification and concluded that the recommended follow-up studies are already being integrated in the International Agency for Research on Cancer (IARC) recommendations for sequence variant classification. Accordingly, for the actual structure of the classification, this committee should continue to pursue the IARC classification system, which is most appropriate for unclassified variants of the MMR genes. The 5-class IARC system links classes to probabilities and is endorsed by the Human Variome Project. It has already been used in assigning pathogenicity to breast and colon cancer gene variants. For the purpose of classifying the MMR gene variants, the quantitative classification model should be supplemented with classification based on qualitative information, including functional assay results, family history and colon cancer specific tests such as microsatellite instability, MMR protein expression with immunohistochemistry, and tumor BRAF mutation status. The final consensus guidelines for the qualitative criteria to classify VUS, as modified during and after this Workshop will soon be published on the InSiGHT website. It is recognized that this classification scheme is a Work In Progress, as there are likely to be refinements, when more experience is gathered with its implementation. These will be minuted and updated regularly on the InSiGHT website, and be regarded as the official recommended approach to qualitative assessment of pathogenicity.

Amanda Spurdle and Bryony Thompson have been systematically applying their qualitative approach to the classification of MMR gene sequence variants identified amongst NIH Colon Cancer Family Register participants. This work is nearing publication. The deliberations of the current Workshop (and its teleconferencing continuation) will be published separately. This publication will include a description of the process and procedures, which the InSiGHT MMR Variant Interpretation Committee has implemented since 2007 (as documented by Finlay Macrae), including the agreed revised qualitative criteria for assessment (above). It will also include the outcome of the formal classification of VUS considered by the Committee. The Committee's deliberations will also be published on the InSIGHT LOVD database (and in referenced journals if agreed by this group) so that they become the official InSiGHT recommendation and can be attached to clinical reports.

It was also agreed that the guidelines should include an ordering of the reliability of RNA-based assays and functional tests or perhaps at least a list of endorsed tests. Also, a logical interpretation based on the functional parameters of each assay should be included. This should be accompanied by an explanation that if the unclassified variant is not expressed at the RNA level, a functional assay carried out at the protein level may not be informative. Amanda Spurdle will consult with Lene Rasmussen and other experts and will develop a supporting document that explains the processes for interpreting functional assay results. This advice will be incorporated into the qualitative classification scheme.

The second part of the meeting discussed variants that have multiple interpretations in the InSiGHT database. Four sets of 12 variants, selected by John-Paul Plazzer, were sent to experts in the field, Rolf Sijmons, Brigitte Royer-Pokora, Amanda Spurdle, Bryony Thompson and Bharati Bapat, who were asked to classify them using qualitative methods. In the meeting, one set of 12 variants was discussed and tentative classification was assigned. It was decided that classification attempts are still premature. In order to have reliable classification, shared qualitative criteria need to be set out. In addition, a minimum number of reports and of qualitative datasets (i.e., segregation studies, in silico analysis, in vitro functional assays, RNA studies) must be available for each variant under scrutiny. Therefore, the whole series of 48 variants will have to be reassessed to check whether they comply with minimal criteria for evaluation.

One of the variants was found to have a co-occurring pathogenic mutation in a just-published dataset from Brigitte Royer-Pokora. It was agreed that in order to ensure that enough information is available for classification of the selected variants in future, they should be published on the InSiGHT/LOVD website with a call for further data submission on these variants before the Committee is convened (link to John-Paul Plazzer's email address john-paul@variome.org). A further link from the InSiGHT website could be implemented. This method could also be used for submission of new VUS or other mutations whose pathogenicity interpretation is problematic so that these can be assessed by the InSiGHT MMR Variant Interpretation Committee.

In summary, the Workshop - and the subsequent post-meeting email discussion - recommended that the InSiGHT MMR Variant Interpretation Committee should function as follows: