CT colonography with computer-aided detection: recognizing the causes of false-positive reader results

Area of Science:

• Radiology
• Medical Imaging
• Gastroenterology

Background:

• Computed tomography (CT) colonography is a key screening tool for colorectal cancer.
• Computer-aided detection (CAD) aims to improve polyp detection accuracy.
• False-positive results are a known challenge in CT colonography CAD.

Purpose of the Study:

• To analyze the frequency and causes of false-positive results in CT colonography CAD.
• To evaluate the impact of these false positives on reader performance.
• To identify strategies for managing CAD-generated polyp candidates.

Main Methods:

• A trial involving 19 readers and 100 cases of CT colonography with CAD.
• Analysis of the types and dismissibility of CAD false-positive findings.
• Assessment of reader performance in response to CAD candidates.

Main Results:

• The majority of CAD false-positive results were readily dismissed by readers.
• Common dismissible causes include artifacts, normal anatomy variations, and small lesions.
• Nondismissable soft-tissue candidates >6 mm were a significant cause of reader false positives.

Conclusions:

• While most CAD false positives are easily managed, specific types can mislead readers.
• Effective evaluation strategies are crucial to avoid unnecessary follow-up procedures.
• Optimizing CAD interpretation can enhance the efficiency of colorectal cancer screening.