ICD-10-CM codes captured and not captured in charlson.sql

ICD-10-CM have three categories which are not part of the WHO ICD-10 and as such are not defined in Quan et.al (2005)

• C4A: Merkel cell carcinoma
• C7A: Malignant neuroendocrine tumors
• C7B: Secondary neuroendocrine tumors

One of the lines in the SQL for mapping ICD-10 to malignant cancer has a possible positive unexpected side effect

https://github.com/MIT-LCP/mimic-code/blob/57069783095e7770e66ea97da264c0200078ddbf/mimic-iv/concepts/comorbidity/charlson.sql#L305

This will map C4A codes to malignant cancer. That seems conceptually okay but it opens the door for what to do with C7A and C7B codes.

If C4A is included in malignant cancer then shouldn't C7A codes?

Should C4A be excluded to be consistent with Quan et.al (2005)?

I think Merkel cells are part of the skin and since Quan et.al (2005) defines "any malignancy, including lymphoma and leukemia, except malignant neoplasm of skin" that C4A should be excluded?

Cancer is not my area of expertise, I honest have no idea. This is more of question of being consistent with the publication.

Start with mimic-iv/concepts/comorbidity/charlson.sql around line 305 and compare its ICD-10-CM mappings with the Quan et al. (2005) definition cited in the issue. Resolve how C4A, C7A, and C7B should be classified, then update the mapping so the SQL is consistent with that decision and review the affected codes.