Potential removal of N/A as a clinical significance for Predictive evidence types

[jsaliba10]

21 Predictive EIDs (11 approved) are currently in CIViC that have N/A listed as their clinical significance. Upon a brief review, some of these should be reclassified as other evidence types and others should have their clinical significance modified with a potential change in evidence direction.

Should N/A be removed as an option for this field?

[kkrysiak]

From our old help docs. I'm not convinced this is helpful but may provide some clarity.

Style being followed below for Predictive Evidence Type: Supports/Does Not Support (Direction)- N/A (Clinical Significance): Definition

• Example

Supports - N/A: Variant does not inform clinical action

• OSI-906, an IGF1R/INSR inhibitor, does not have a significant effect on chemotherapy resistant castration-resistant prostate cancer cell lines.

Does Not Support - N/A: Variant does not inform clinical action

• There is no statistical difference in progression free survival between lung cancer patients with or without an EGFR L858R mutation following treatment with gefitinib or erlotinib.

So I can find this in the future: I got this information from here.

[jsaliba10]

Ok. Well we'd should do one of 3 options then:

Define N/A in the docs Come up with a new term for no change outcomes and define it Find a way to use existing clinical significance combinations to express what the similar result means in the context of the paper

I think the third option may be possible based on the definitions currently outlined in the docs for supports/does not support sensitive or resistance from the table and figure 8. To test this, I can try to go through these EIDs and make suggestions. If that proves difficult, then creating a new term or definition for N/A would be useful.

[kkrysiak]

related https://github.com/griffithlab/civic-client/issues/1560