Identify gaps in diagnostic topography hierarchy (SNOMED, ICDO3)

[kzollove]

• What are the concrete issues experienced that we'd like to address in the hierarchies of ICD-O-3 #638

[odikia]

per comment on issue 601, Peter and I are working on the ICDO3 gaps (both hist and topography).

[rtmill]

It would be interesting to see if it's possible to condense the two topography hierarchies into one, or at least find a way to consistently query for cancers by body regions (e.g. "neoplasms of thorax") that include all descendant concepts where you don't have to specify both ICDO3 topography and SNOMED topography.

Here's an example where SNOMED falls short and there isn't a 1:1 : https://athena.ohdsi.org/search-terms/terms/36546338

Looks like the trade-off is that:

1. ICDO3 has more specific sites that SNOMED doesn't
2. and SNOMED has higher reaching, more general concepts that ICDO3 lacks.

Is it worth deduplicating the middle area so we can pick descendants from anywhere in the hierarchy? More spit-balling than suggesting.

[odikia]

@rtmill , just closing the loop here with updates from the virtual meeting in the oncology WG.

@peterprinsen-iknl and I will certainly delve into this task as a first priority of the second community contribution release. The first release is focused on updating ICDO3 to match the Vocab subgroup’s release of SNOMED, as well as fix errors found in the current mapping. I’ll set a reminder for myself to check back in early January, at which time, our contribution to the February 2024 release will be complete, and onto QA.