Text Extraction and Standardization System Development for Pathological Records in the Korea Biobank Network

Area of Science:

• Biomedical Informatics
• Cancer Research
• Biobanking

Background:

• The Korea BioBank Network (KBN) manages valuable pathological records crucial for research.
• Efficient and accurate data extraction from these records is essential for advancing scientific discovery.
• Current data extraction methods can be time-consuming and prone to errors.

Purpose of the Study:

• To establish a time-efficient and error-reduced system for extracting data from KBN pathological records.
• To validate the system's performance using real-world cancer cohort data.
• To facilitate broader data utilization across multiple institutions.

Main Methods:

• Development of a systematic, step-by-step data extraction protocol.
• Implementation of quality control measures to minimize errors.
• Testing the system on large cohorts of lung cancer (769) and breast cancer (1292) patients.

Main Results:

• The developed system demonstrated significant improvements in time efficiency and error reduction.
• An overall accuracy of 91% was achieved during the extraction process.
• Successful validation on diverse cancer cohorts indicates system robustness.

Conclusions:

• The established data extraction system offers a reliable and efficient solution for processing KBN pathological records.
• This advancement can accelerate research by improving access to high-quality biobank data.
• The system holds potential for application in other institutions managing similar datasets.