Development of Medical Imaging Data Standardization for Imaging-Based Observational Research: OMOP Common Data Model...

Area of Science:

• Medical informatics
• Artificial intelligence in healthcare
• Biomedical imaging analysis

Background:

• Advancements in artificial intelligence (AI) and deep learning necessitate large, harmonized datasets for developing robust medical models.
• Existing data models like the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM) require extensions to fully support complex imaging research.

Purpose of the Study:

• To propose and describe the Medical Imaging Common Data Model (MI-CDM) extension for the OMOP CDM.
• To address the structural and semantic requirements for integrating and analyzing medical imaging data within a standardized framework.

Main Methods:

• Introduction of two new tables within the OMOP CDM to manage imaging data.
• Addition of two new vocabularies to standardize imaging-related terminologies.
• Development of capabilities to link DICOM data sources and track the provenance of imaging features.

Main Results:

• The MI-CDM extension facilitates the linking of DICOM data and the tracking of imaging feature origins.
• Enables the definition of phenotypes using imaging features, expanding computable imaging biomarkers.
• Provides a unified approach for imaging research and outcome studies.

Conclusions:

• The MI-CDM extension enhances the OMOP CDM's utility for AI-driven medical imaging research.
• It supports the development of standardized, computable imaging biomarkers and facilitates phenotype definition.
• Offers a comprehensive solution for leveraging imaging data in healthcare research and outcome analysis.