대한유방검진의학회

Review Article

Full Text(PDF)

Breast cancer remains a significant global health issue, and early detection is pivotal for improving patient outcome (1-4). Breast cancer screening using mammography has been increasing continuously over the past decade and it places a substantial workload on breast radiologists (5-6). In recent years, many retrospective studies have indicated that employing artificial intelligence (AI) considerably reduced the workload of breast radiologists while improving screening performance (8-14). However, the reliability of such AI tools demands scrutiny, particularly in real-world clinical settings. In mammography utilizing AI, diagnostic errors— false negatives (FNs) and false positives (FPs) are critical challenges that significantly impact diagnostic accuracy. Understanding these issues helps to discern both the strengths and limitations of AI tools in breast cancer detection and screening. This review will focus on cases of FNs and FPs in mammography interpretation using Lunit AI system
Key Words: