Mapping to ICD9 or ICD10 procedures

[frdougal]

Hello,
Thank you for the work on MAxO. This is an extremely useful ontology. Are there any mappings connecting MAxO classes to ICD9 or ICD10 procedures?

[pnrobinson]

@frdougal currently, no, but this would be a good idea.
Currently not able to access the codes, the FDA website says: "The NHSN website is currently undergoing maintenance as we implement updates to ensure compliance with Executive Order 14168, titled "Defending Women from Gender Ideology Extremism and Restoring Biological Truth to the Federal Government.""
@matentzn - Nico, maybe we can find some time to do a mapping operation for ICD10?

[frdougal]

@pnrobinson thanks for the reply! It looks like the links here work (as of Feb 3): https://www.cms.gov/medicare/coding-billing/icd-10-codes

[matentzn]

@frdougal thank your for your request; many apologies for the disappointing answer here.

We will discuss your request internally, but right now this is not high up on the list of our priorities. Would you be able to make a very strong case for such a mapping? How and in the context of what project would you leverage it?

In the meantime, I believe there is a path to go from MAXO to NCIT, and there must be some NCIT to ICD10 mappings out there as well.

[frdougal]

@matentzn thank you for the reply. The task of mapping clinical data from EHR systems into molecular ontologies is very difficult. This is an important task to aid Learning Health Systems (LHS). Without these mappings, the data stays siloed within the EHR system (ICD, SNOMED, etc.) and cannot be connected to genomic data and systems biology ontologies (GO, Reactome, UBERON). These connections allow researchers to improve standard of care and further biological understanding of diseases.

There are some mapping files connecting ICD9/ICD10 diagnoses to MONDO/HPO and some mappings for drugs. The MAXO resource represents the only method (AFAIK) that allows a user to connect a concept of a procedure to the anatomic areas the procedure affects. Most procedures are coded in EHR using either ICD procedure codes or CPT. I do not believe EHR systems would code procedures using NCIT ids.

As you point out, MAXO contains mappings to NCIT. However, there are only around 1800 NCIT dbxrefs out of 14,000 MAXO classes (in the maxo-full.owl file). Many of these mappings (I do not have a method for accurately counting these) are not procedures but are cancer related terms imported from other ontologies (ex: HP, UBERON, CL). Furthermore, the cancer-centric nature of these mappings excludes common procedures like appendectomy.

Adding mappings to ICD10 makes MAXO more useful to a larger audience. It permits users to connect MAXO to EHR data. I think this mapping would greatly improve the utility of this project.

[pnrobinson]

@frdougal we would welcome contributions in this area. I am afraid we do not have the resources right now to embark upon a major mapping effort.