Colonic polyps: complementary role of computer-aided detection in CT colonography.

Area of Science:

• Radiology
• Medical Imaging
• Computer-Aided Diagnosis

Background:

• Computed tomographic (CT) colonography is a screening tool for colorectal polyps.
• Radiologist interpretation of CT colonography can be challenging, with reported sensitivities for large polyps varying significantly.
• Computer-aided detection (CAD) algorithms aim to enhance diagnostic accuracy in medical imaging.

Purpose of the Study:

• To evaluate the added benefit of a computer-aided detection (CAD) algorithm applied to supine and prone multisection helical CT colonographic images.
• To compare the performance of CAD with standard clinical interpretation by radiologists.

Main Methods:

• CT colonography was performed on 40 asymptomatic high-risk patients using a multisection helical CT scanner.
• Patients were imaged in both supine and prone positions.
• Images were interpreted by two blinded radiologists and an automated CAD algorithm.

Main Results:

• The sensitivity of radiologists for detecting large polyps (≥1.0 cm) was 48%.
• The CAD algorithm also had a sensitivity of 48% but detected 31% of large polyps missed by radiologists, increasing potential sensitivity to 64%.
• The combined sensitivity of CAD and radiologists reached 89% for retrospectively identifiable polyps, with an average of 11 false-positive detections per patient for CAD.

Conclusions:

• The CAD algorithm demonstrated a complementary role in interpreting CT colonographic images.
• It successfully identified large polyps missed by trained observers, particularly in a cohort where radiologists faced detection challenges.
• CAD can enhance the detection of significant colorectal polyps when used alongside conventional interpretation.