Time-efficient CT colonography interpretation using an advanced image-gallery-based, computer-aided "first-reader"...

Area of Science:

• Radiology
• Gastroenterology
• Medical Imaging

Background:

• Computed tomographic colonography (CTC) is a valuable tool for colorectal cancer screening.
• Computer-aided detection (CAD) systems are increasingly employed to enhance the interpretation of medical images.
• Accurate detection of colorectal adenomas, particularly those ≥ 6 mm, is crucial for effective screening.

Purpose of the Study:

• To evaluate the performance of a novel "first-reader" computer-aided detection (CAD) workflow for identifying colorectal adenomas ≥ 6 mm using CTC.
• To assess the impact of this CAD workflow on radiologist interpretation time and detection accuracy in a low-prevalence cohort.

Main Methods:

• Retrospective review of 616 colonoscopy-validated CTC datasets by a radiologist using a CAD "first-reader" workflow.
• CAD presented detections via image galleries (2D/3D) with interactive views.
• Analysis included per-patient, per-polyp, and per-adenoma sensitivities for lesions ≥ 6 mm.

Main Results:

• The CAD workflow achieved an 84.3% per-patient sensitivity and 83.2% per-polyp sensitivity for lesions ≥ 6 mm.
• Sensitivity for adenomas ≥ 6 mm was 85.9%, with 100% detection of adenomas ≥ 10 mm and cancers.
• Overall specificity was 95.6%, and the mean interpretation time was only 3.1 minutes per patient.

Conclusions:

• An image-gallery-based "first-reader" CAD workflow is feasible and effective for detecting colorectal adenomas ≥ 6 mm via CTC.
• This CAD approach can significantly reduce interpretation times, making CTC screening more efficient.
• The CAD system facilitates accurate detection, ensuring no large adenomas were missed in this study.