OHDSI / OncologyWG

Oncology Working Group Repository
https://ohdsi.github.io/OncologyWG
Apache License 2.0
53 stars 24 forks source link

Delta Review of Oncology-Related Terms in OMOP Vocabulary #643

Open p-talapova opened 2 months ago

p-talapova commented 2 months ago

Hello everyone,

I have completed a review of the recent updates to the oncology-related terminology within the OMOP vocabulary (DeltaVocab). This message is not to highlight a problem per se, but rather to enhance our understanding of the variations between different versions of the vocabulary. I recommend this issue be closed after it has been evaluated by the working group.

Findings: 1) Ambiguous Source Concept Names: The concept table shows 89 instances where different concept_ids share identical names despite having different semantic meanings. This overlap can lead to confusion and mismatches when these names are used to link source data with non-standard terms. A possible solution is to revise the concept names to include more descriptive, distinguishing details and move the current identical names to the concept_synonym table.

Snippet 1:

Source Concept ID Source Name Relationship ID Target Concept ID Target Concept Name Concept Label Category Type
35919047 0 Maps to 1634502 c-AJCC/UICC-Stage-0 AJCC/UICC clinical Stage 0 Staging/Grading Measurement
35919331 0 Maps to 1633542 p-AJCC/UICC-Stage-0 AJCC/UICC pathological Stage 0 Staging/Grading Measurement
35919511 0a Maps to 1635865 p-AJCC/UICC-Stage-0A AJCC/UICC pathological Stage 0a Staging/Grading Measurement
35919890 0a Maps to 1635434 c-AJCC/UICC-Stage-0A AJCC/UICC clinical Stage 0a Staging/Grading Measurement

2) Potentially Misaligned Stage Mappings: I have identified 21 mappings that potentially incorrectly align pathological stages with clinical stages. To rectify this, a review and realignment of these mappings in OMOP Vocabulary are suggested, ensuring each term accurately reflects its designated stage category, either clinical or pathological.

Snippet 2:

Source Concept ID Source Name Relationship ID Target Concept ID Target Concept Name Additional Relationship ID Concept ID Concept Name
45876314 Stage IVC Maps to 1635567 AJCC/UICC pathological Stage 4C Answer of 3022698 Stage group.clinical Cancer
45876315 Stage IVB Maps to 1633577 AJCC/UICC pathological Stage 4B Answer of 3022698 Stage group.clinical Cancer
45876316 Stage IIA Maps to 1634960 AJCC/UICC pathological Stage 2A Answer of 3022698 Stage group.clinical Cancer
45878382 Stage IIC Maps to 1633859 AJCC/UICC pathological Stage 2C Answer of 3022698 Stage group.clinical Cancer

Please find the attached XLSX file for a detailed review. I have used the SSSOM approach for the "IdenticalSourceNames_DiffMeaning" spreadsheet and modified the typical OMOP mapping representation in the "StageConflict_PathVsClinical" spreadsheet (just for your reference).

Looking forward to further discussion.

OncologyWG-VocabDeltaReview.xlsx .