Quality analysis of a breast thermal images database

Area of Science:

• Medical Imaging
• Data Science
• Oncology

Background:

• Early breast cancer detection is crucial for effective treatment.
• The infrared imaging database is widely used in mastology research.
• Data quality is paramount for reliable research outcomes.

Purpose of the Study:

• To conduct an exhaustive analysis of a widely used mastology research database containing infrared images.
• To identify and categorize anomalies within the database based on five quality dimensions: completeness, correctness, concordance, plausibility, and currency.
• To establish control queries for ensuring database quality and to review existing literature utilizing this database.

Main Methods:

• Performed an in-depth analysis of the breast cancer infrared imaging database.
• Evaluated data quality across five dimensions: completeness, correctness, concordance, plausibility, and currency.
• Developed and applied control queries to detect anomalies in personal, clinical, and thermal image data.

Main Results:

• Identified a total of 365 anomalies within the database.
• Anomalies were found in personal and clinical data, as well as in thermal images.
• A review of over 40 papers using the database revealed no mention of these identified anomalies.

Conclusions:

• The discovered errors in the database may significantly alter results and conclusions of previous studies.
• This research provides a foundation for improving the quality and integrity of the mastology database.
• Findings will assist future researchers in utilizing the database more effectively and with greater confidence.