Exploring SNOMED CT for Breast Cancer Research in a Real-World Context

In clinical data management and research, accurately coding diagnoses and treatments is vital. SNOMED CT offers precise standardized codes for disease classification and therapeutic interventions.

We will explore identifying codes to find patients with breast cancer diagnoses in electronic health records using SNOMED CT coding.

Women with hormone receptor (estrogen receptor, ER) positive breast cancer commonly receive endocrine therapy using drugs such as tamoxifen or aromatase inhibitors (AIs) as part of their treatment. However these treatments are generally given to different groups of women;

• Tamoxifen is a selective estrogen receptor modulator (SERM) primarily used in pre-menopausal women, and occasionally indicated for post-menopausal women; it works by blocking ER on cancer cells, stopping estrogen-driven growth.

• Aromatase inhibitors (AI, e.g., anastrozole, letrozole, exemestane) prevent estrogen production by inhibiting the aromatase enzyme and are mostly used in post-menopausal women.

• Tamoxifen remains the standard adjuvant therapy for pre-menopausal women, while AIs are preferred for post-menopausal women due to better efficacy.

• Menopausal status is often approximated using age 55 as a surrogate marker in clinical and research settings.-

Exercise

Query SNOMED-CT for breast cancer SNOMED-CT codes Refine to find codes for hormone positive breast cancer

• Submit your all code steps with comments.
• Provide the data table containing breast cancer-related ICD-10 codes and descriptions derived from SNOMED CT mappings.

R

Enclosing R code (quatro markdown) using the R package Rdiagnosislist to query SNOMED-CT. Other R packages that might be useful are snomedizer (R interface to the SNOMED CT Terminology Server REST API) or https://github.com/ropensci/trud. (Inferface to NHS TRUD API to programmatically download SNOMED and other NHS reference data directly )

Python

If you do not wish to use, R. You can develop similar code in python Libraries and resources include

A Python equivalent for working with SNOMED CT similar to the R repository you have (https://github.com/BM6063/SNOMEDCT) can be built using libraries such as PyMedTermino or Pathling:

1. PyMedTermino Offers Python bindings to access SNOMED CT concepts, relationships, and ontological queries. Allows you to load SNOMED CT ontologies and search concepts by code or description. Supports accessing synonyms, parents, children, and associated clinical findings.

Example snippet:

python from pymedtermino.snomedct import * SNOMEDCT.load()

concept = SNOMEDCT[254837009] print(concept) print(concept.definition) print(concept.parents) Documentation: https://pythonhosted.org/PyMedTermino/

2. Pathling (Python API) Library to analyze and group SNOMED CT concepts with a terminology server backend. Works with Spark for scalable data pipelines. Allows you to query value sets, explore hierarchies, and display SNOMED terms.

Example snippet:

python from pathling import PathlingContext

pc = PathlingContext.create() df = pc.spark.read.csv("snomed_data.csv", header=True) df = df.withColumn("Snomed Term", pc.snomed.display("concept_column")) df.show() Tutorial: https://pathling.csiro.au/docs/libraries/examples/grouping-snomed

3. Using REST APIs like Snowstorm Snowstorm API offers RESTful services to query SNOMED concepts, relationships, and maps. You can use Python libraries like requests to query and retrieve SNOMED data.

Example snippet:

python import requests

url = "https://snowstorm.example.com/fhir/CodeSystem/$lookup?code=254837009" response = requests.get(url) print(response.json())

GitHub project: https://github.com/AberystwythSystemsBiology/SCTTSRApy