Approach to a preparation of dataset combining digital mammographic images and patient clinical data from electronic medical records

Veronika Kazarinova ¹, Yuriy Vasilev ¹, Anton Vladzymyrskyy ^1,2

⁰State Budget-Funded Health Care Institution of the City of Moscow "Research and Practical Clinical Center for Diagnostics and Telemedicine Technologies of the Moscow Healthcare Department", Moscow, Russian Federation.

Quantitative Imaging in Medicine and Surgery +

|

April 16, 2025

View abstract on PubMed

Summary

This study presents a method to enrich radiological datasets with clinical information from electronic medical records, improving data reuse and compliance with FAIR principles. The enriched dataset aids in developing AI-based software for pathology evaluation using mammographic images and clinical data.

Area Of Science

Medical Informatics
Radiology
Data Science

Background

Dataset generation is resource-intensive and complex.
Reusing and enriching existing datasets aligns with FAIR principles.
Electronic medical records (EMRs) contain valuable clinical information for dataset enrichment.

Purpose Of The Study

To develop and validate a method for enriching radiological datasets with clinical information from EMRs.
To create a FAIR-compliant dataset for research and AI development.
To investigate the correlation between clinical parameters and mammographic pathology.

Main Methods

Literature review to identify relevant clinical signs.
Selection of studies with and without specific pathologies.
Extraction and processing of clinical data from EMRs.
Generation of a dataset including mammographic images and clinical variables.

Main Results

A dataset of 200 patients with mammographic images and clinical data (age, age at menopause, number of births) was generated.
Statistical analysis revealed significant differences in age at study, age at menopause, and late menopause between patients with and without pathology.
A weak correlation was observed between studied parameters and pathology presence.

Conclusions

The developed method efficiently enriches radiological datasets with clinical information, saving resources.
The generated dataset supports research and the training/testing of AI-based software for mammography analysis.
Clinical parameters, particularly age-related factors, show statistically significant associations with mammographic pathology.

Keywords:

Dataset artificial intelligence (AI) malignant breast tumors mammography