AI More Than Doubles Detection of Incidental Breast Lesions on CT
June 5, 2025
AI systems detect twice as many incidental breast lesions on routine CT scans as radiologists.
Key Details
- AI algorithms identified twice the number of incidental breast lesions compared to radiologists.
- Incidental breast lesions are detected on up to 8% of CT scans including the chest, potentially higher.
- Up to 70% of these incidental lesions turn out to be cancerous, highlighting the importance of early detection.
- Challenges for radiologists include lower spatial and contrast resolution of CT compared to mammography and MRI, and lesion location or incidental nature.
- Findings are from a new paper published in Current Problems in Diagnostic Radiology.
Why It Matters
Early and more frequent detection of incidental breast cancers on routine CT has significant potential to improve patient outcomes. Increasing AI support in imaging workflows could address current blind spots and enhance secondary cancer detection, reducing missed diagnoses.