Automated image-based colon cleansing for laxative-free CT colonography computer-aided polyp detection

Area of Science:

• Medical Imaging
• Gastroenterology
• Artificial Intelligence in Medicine

Background:

• Computed tomography colonography (CTC) is a valuable tool for colorectal cancer screening.
• Noncathartic bowel preparation for CTC aims to improve patient compliance but can lead to residual stool artifacts.
• Computer-aided detection (CAD) systems show promise for improving polyp detection rates in CTC.

Purpose of the Study:

• To evaluate the efficacy of an automated, image-based colon cleansing algorithm in conjunction with a CAD system for detecting colonic polyps.
• To assess the performance improvement of CAD for polyp detection in noncathartic CTC settings.

Main Methods:

• Development of an automated colon cleansing algorithm using intensity, shape, and texture analysis with probabilistic optimization to detect and subtract tagged stool.
• Application of the algorithm to CTC data from patients with noncathartic bowel preparation (barium or Gastroview, no laxatives).
• Comparative analysis of CAD system performance with and without the stool subtraction algorithm on polyp-enriched datasets.

Main Results:

• The study included 38 CTC scans from 19 patients, with 44 polyps larger than 10 mm.
• Without stool subtraction, the CAD system achieved 70.5% sensitivity with 5.75 false positives per scan.
• With the automated colon cleansing algorithm, sensitivity significantly improved to 86.4% true positive rate (p=0.009) at the same false positive rate.

Conclusions:

• Automated image-based colon cleansing substantially enhances the sensitivity of colon CAD systems for polyp detection in noncathartic CTC.
• The algorithm effectively addresses stool-related artifacts, leading to an approximate 15% increase in CAD sensitivity.
• This approach offers a promising solution for improving the accuracy of CTC-based polyp detection without requiring laxative bowel preparation.