Stream: implementers
Topic: Cancer genomics
Halina Labikova (May 27 2021 at 19:25):
We're currently rebuilding our genomics IG to align with mCODE. I'm trying to wrap my head around how the CancerGeneticVariant profile can be constrained further to capture specific cancer-related mutations.
Let's say I want to create a profile that captures METex14 skipping mutation (common in lung cancer). If I understand correctly:
1) valueCodeableConcept is set as 'positive'
2) component:geneStudied is set to 'HGNC:7029' (MET)
3) ... what about the exon 14 skipping part? There are hundreds of distinct genetic alterations (by ClinVar) that would lead to exon14 skipping, figuring out which ones exactly to bind a valueset to the geneStudied or the variationCode components doesn't seem like the right approach. Similarly, using GenomicRegionStudied in the report doesn't seem like the way to go either, since the critical mutation can happen in the adjacent introns as well.
I guess to generalize this question - how do I use the profile to specify that the mutation has an oncogenic effect (either by describing the mutation type or the effect on the produced protein)?
@Patrick Werner @May Terry @Alex Goel
Patrick Werner (May 27 2021 at 19:32):
We were discussing the lack of support/limited support for structural variants, translocations and fusions during the WGM
Patrick Werner (May 27 2021 at 19:33):
The genomics WG will create a sub-group to improve the support and modify/create existing profiles.
Patrick Werner (May 27 2021 at 19:33):
1) is correc
Patrick Werner (May 27 2021 at 19:33):
2) correct
Patrick Werner (May 27 2021 at 19:34):
3) exon4 would go into: component:CytogenicLocation
Patrick Werner (May 27 2021 at 19:35):
deletion would mean: 1 copies in: component:copy-number
Patrick Werner (May 27 2021 at 19:36):
The oncogenic effect would be a diagnostic Implication derivedFrom the variant
Halina Labikova (May 27 2021 at 19:40):
The cytogenic location is not necessarily exon 14 - certain types of mutations in surrounding intronic regions would have the same effect
And the skipping mutation is not necessarily a deletion, some point mutations also mimic the loss of this region
Halina Labikova (May 27 2021 at 19:52):
Using the Implication profile seems like the most suitable approach so far. Any plans on bringing this one over to mCODE?
Halina Labikova (May 27 2021 at 19:59):
Thank you btw! Where I can follow up on the sub-group's progress in this topic?
May Terry (May 27 2021 at 20:58):
@Halina Labikova - for mCODE, the CancerGeneticVariant.valueCodeableConcept would be 'present' and "positive" is in the interpretation field.
We are planning on adding DiagnosticImplication into mCODE primarily for the support of molecular consequence as well as clinical significance and have a JIRA ticket opened by Dr. Jim Chen, who is further asking for representation of fusion events among other additions - https://jira.hl7.org/browse/FHIR-32008. I have a "post-STU2" mCODE development branch and am starting to add some of these additions for presentation and approval during the mCODE ballot reconciliation process.
May Terry (May 27 2021 at 21:02):
@Patrick Werner - Sorry I missed this session during the WGM. I'd like to be part of the discussions for how CGWG handles fusion events since we have this need for mCODE. I recall we originally touched on how to represent fusion events and tumor normal tests sometime last year but it never gained traction. I'd really like to revisit this again.
Patrick Werner said:
We were discussing the lack of support/limited support for structural variants, translocations and fusions during the WGM
Last updated: Apr 12 2022 at 19:14 UTC
