Create list of NAACCR grade concepts to destandardize

[rtmill]

"Variable" concepts: 35918519 Grade Post Therapy ,35918542 Grade Clinical ,35918640 Grade Pathological

They seem to share the same value sets with many duplications (same name and meaning but from different NAACCR schemas). There are 121 unique "answer" concepts to destandardize but only 11 unique concept names to map and then apply to all.

Unique concept names:

Grade cannot be assessed (GX); Unknown
Grade cannot be assessed; Unknown
Grade cannot be assessed (GX); Unknown |Not applicable
GB: Borderline Tumor
G3B: > 15 centroblasts per HPF and solid sheets of centroblasts
Grade cannot be assessed (GX); Unknown |Not a follicular histology (9690/3, 9691/3, 9695/3, 9698/3)
Low grade: Grade 1-2
G1: 0-5 centroblasts per HPF
TP53 or CTNNB Mutation
G2: 6-15 centroblasts per HPF
G3: > 15 centroblasts
G3A: >15 centroblasts per HPF and centrocytes present

Suggestion: Destandardize the "unknown" answers without mapping, then map the remaining.

This boils it down even further to:

GB: Borderline Tumor
G3B: > 15 centroblasts per HPF and solid sheets of centroblasts
Low grade: Grade 1-2
G1: 0-5 centroblasts per HPF
TP53 or CTNNB Mutation
G2: 6-15 centroblasts per HPF
G3: > 15 centroblasts
G3A: >15 centroblasts per HPF and centrocytes present

[kzollove]

There are more NAACCR grade concepts that need to be destandardized:

See the answers of https://athena.ohdsi.org/search-terms/terms/35918328 (and that concept itself)

select *
from prod.concept c
where concept_id = 35918328
union all 
select c.*
from prod.concept c
inner join prod.concept_relationship cr
on c.concept_id =cr.concept_id_2 
and concept_id_1 = 35918328
and relationship_id = 'Has Answer'
and standard_concept = 'S'

[kzollove]

Borderline tumor still needs to be addressed in Cancer Modifier #568

[kzollove]

GB: Borderline Tumor

Borderline tumor still needs to be addressed in Cancer Modifier https://github.com/OHDSI/OncologyWG/issues/568

G3B: > 15 centroblasts per HPF and solid sheets of centroblasts Low grade: Grade 1-2 G1: 0-5 centroblasts per HPF G2: 6-15 centroblasts per HPF G3: > 15 centroblasts G3A: >15 centroblasts per HPF and centrocytes present

These lymphoma grades seem to be different in nature than a "grade 3 tumor", which is the closest concept i found in cancer modifier (see #574)

[kzollove]

TP53 or CTNNB Mutation

I can see that this is a NAACCR grade concept, but I don't understand why. No existing mapping

[kzollove]

@rtmill can you explain this suggestion more?

Grade cannot be assessed (GX); Unknown
Grade cannot be assessed; Unknown
Grade cannot be assessed (GX); Unknown |Not applicable

Suggestion: Destandardize the "unknown" answers without mapping, then map the remaining.

[kzollove]

List of NAACCR concepts to destandardize

Source: grade_NAACCR.csv