OHDSI Home | Forums | Wiki | Github

ICDO Histology in Usagi

I’ve downloaded the ICDO3 codes and have Condition in Usagi but not Histology. I see the Histology is Non-standard. Why is this the case? I’ve got a large number of ICDO3 SEER codes/ terms to map

Well, ICDO3 histology are not standard concepts. You cannot map to source concepts because you cannot use them in the data. Conditions are combinations of histology and topology. If you don’t have those, go up the hierarchy into SNOMED territory.

@jliddil1
Do you have both an ICDO topography and an ICDO histology in your source data? If so, you can use the following SQL to map the combination to a single standard Condition concept :

SELECT c1.concept_id
, c1.concept_name
, c1.vocabulary_id
, c1.concept_code
, c1.standard_concept
, cr1.relationship_id
, c2.concept_id
, c2.concept_name
, c2.vocabulary_id
, c2.concept_code
, c2.standard_concept
, c3.concept_class_id
, c3.concept_code
, c4.concept_class_id
, c4.concept_code
FROM concept c1 JOIN concept_relationship cr1 on c1.concept_id = cr1.concept_id_1 AND cr1.relationship_id = ‘Maps to’
JOIN concept c2 on cr1.concept_id_2 = c2.concept_id AND c2.standard_concept = ‘S’
JOIN concept_relationship cr2 ON cr2.concept_id_1 = c1.concept_id AND cr2.relationship_id = ‘Has Topography ICDO’
JOIN concept c3 ON cr2.concept_id_2 = c3.concept_id AND c3.vocabulary_id = ‘ICDO3’
JOIN concept_relationship cr3 ON cr3.concept_id_1 = c1.concept_id AND cr3.relationship_id = ‘Has Histology ICDO’
JOIN concept c4 ON cr3.concept_id_2 = c4.concept_id AND c4.vocabulary_id = ‘ICDO3’
WHERE c3.concept_code = ‘C07.9’
AND c4.concept_code = ‘8140/3’

If in your source data you only have ICDO topography or only have ICDO histology, right now we can’t handle that case, but we will be adding vocabulary support to cover these cases in the future.

I’m not a coder. I’m doing mapping I’ll have to speak to our data lake implementation folks about doing the join via sparql. Yes we collect both histology as a part of out standard process. This is a standard practice in oncology.

… and what? Topology? Then you are all set, @jliddil1.

For a proper diagnosis you need both. But we realize people just say “osteosarcoma” without indicating of which bone. You can still find those in SNOMED, like Osteosarcoma of bone.

In recent version some of ICD-O-3 codes have been made standard (all Conditions and some Histologies and Topographies), but non-standard codes always have mapping relations to standard Morphologies SNOMED vocabulary, and they share a single hierarchy. It should be possible simply to use USAGI with no barriers now.

But if you have topography and histology codes ready, you should really just find needed standard concepts based on codes combination. It’s more reliable than trusting USAGI to make mappings based on text similarity.

t