pancreatic adenocarcinoma located in some pancreas?

[jonathanbona]

Pancreatic adenocarcinoma (http://purl.obolibrary.org/obo/DOID_4074) is a subclass of (cancer and ('located in' some pancreas).

My concern about using this is that when someone's pancreatic adenocarcinoma metastasizes to his lungs then there's an instance of pancreatic adenocarcinoma that is not located in the pancreas, and that may persist even if the patient has no pancreas (for instance if the pancreas has to be removed), leading to false inferences like: this person has a pancreas.

This also seems to be an instance of the located_in domain violation discussed in issue https://github.com/DiseaseOntology/HumanDiseaseOntology/issues/649.

[melodyswen]

Thanks Jonathan - we'll discuss this and get back to you!

[lschriml]

Hello Jonathan, Thank you for connecting with us.

As DO was developed we grappled with this exact problem. In representing diseases as distinct classes, we represent each disease distinctly, that is, the primary cancer in this example, pancreatic adenocarcinoma. And the latter disease as lung adenocarcinoma. In our data model, we do not represent metastasis, or disease
progression/disease course.

NCI thesaurus represents metastatis: https://www.ebi.ac.uk/ols/ontologies/ncit/terms?iri=http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FNCIT_C19151

Cheers, Lynn

[andredekker]

I would request to reopen this issue. It is a given reality that a e.g. a prostate cancer can be located in the spine. It could even be that the prostate is no longer there (e.g. because it has been removed). By including the "located_in" axiom you assume that a prostate must be there which is not true. In my opinion you should change it to a property such as the NCI thesaurus which uses has_primary_site.

[jonathanbona]

Could this be resolved by replacing located_in with disease_has_location, as in https://github.com/DiseaseOntology/HumanDiseaseOntology/issues/649#issuecomment-461236504 ?