Related Experiment Video
Updated: Nov 20, 2025
05:33
Introduction of an Integrated Pathology Image Management, Artificial Intelligence, and Reporting System
Published on: July 11, 2025
552
Current Available Computer-Aided Detection Catches Cancer but Requires a Human Operator
Florentino Saenz Rios1, Giri Movva1, Hari Movva2
1Department of Radiology, University of Texas Medical Branch, Galveston, USA.
Cureus
|January 25, 2021
Summary
Computer-aided detection (CAD) for mammograms is helpful but not a replacement for radiologists. While CAD can identify potential cancers, it frequently misidentifies benign lesions, underscoring the need for expert human interpretation.
Area of Science:
- Radiology
- Medical Imaging
- Artificial Intelligence in Healthcare
Background:
- Computer-aided detection (CAD) systems are widely used in mammography interpretation.
- Current CAD systems have limitations in accurately differentiating malignant from benign breast lesions.
- The integration of advanced AI algorithms is being explored to enhance radiologist capabilities.
Purpose of the Study:
- To evaluate the diagnostic performance of CAD systems in mammogram interpretation.
- To compare the accuracy of CAD with radiologist interpretation for breast cancer detection.
- To determine the role of CAD as an adjunct tool for radiologists.
Main Methods:
- Retrospective analysis of mammograms from patients with BI-RADS 6 findings (confirmed malignancies).
- Utilized CAD read images from Hologic and General Electric systems (2019-2020).
- Correlated CAD findings with pathology reports and institutional medical records.
Main Results:
- CAD systems demonstrated statistically significant misidentification of breast cancer in the study population.
- Radiologist interpretation proved to be the most effective tool for accurate diagnosis.
- A small sample size (24 patients) was due to COVID-19 restrictions.
Conclusions:
- CAD systems struggle to differentiate benign from malignant lesions accurately.
- CAD should be utilized as a supplementary tool, not a standalone diagnostic solution.
- Human expertise remains crucial for precise mammogram interpretation and cancer diagnosis.
