Bowel Polyp Detection in Capsule Endoscopy Images with Color and Shape Features

Following incomplete colonoscopy (IC) patients often undergo computed tomography colonography (CTC), but colon capsule endoscopy (CCE) may be an alternative. We compared the completion rate, sensitivity and diagnostic yield for polyp detection from CCE and CTC following IC. A systematic literature search resulted in twenty-six studies. Extracted data included inter alia, complete/incomplete investigations and polyp findings. Pooled estimates of completion rates of CCE and CTC and complete colonic view rates (CCE reaching the most proximal point of IC) of CCE were calculated. Per patient diagnostic yields of CCE and CTC were calculated stratified by polyp sizes. CCE completion rate and complete colonic view rate were 76% (CI 95% 68–84%) and 90% (CI 95% 83–95%). CTC completion rate was 98% (CI 95% 96–100%). Diagnostic yields of CTC and CCE were 10% (CI 95% 7–15%) and 37% (CI 95% 30–43%) for any size, 13% (CI 95% 9–18%) and 21% (CI 95% 12–32%) for >5-mm and 4% (CI 95% 2–7%) and 9% (CI 95% 3–17%) for >9-mm polyps. No study performed a reference standard follow-up after CCE/CTC in individuals without findings, rendering sensitivity calculations unfeasible. The increased diagnostic yield of CCE could outweigh its slightly lower complete colonic view rate compared to the superior CTC completion rate. Hence, CCE following IC appears feasible for an introduction to clinical practice. Therefore, randomized studies investigating CCE and/or CTC following incomplete colonoscopy with a golden standard reference for the entire population enabling estimates for sensitivity and specificity are needed.

Download Full-text

Intestinal polyp recognition in capsule endoscopy images using color and shape features

2009 IEEE International Conference on Robotics and Biomimetics (ROBIO) ◽

10.1109/robio.2009.5420969 ◽

2009 ◽

Cited By ~ 10

Author(s):

Baopu Li ◽

Yichen Fan ◽

Max Q.-H. Meng ◽

Lin Qi

Keyword(s):

Capsule Endoscopy ◽

Shape Features ◽

Intestinal Polyp

Download Full-text

Region-Based Automated Localization of Colonoscopy and Wireless Capsule Endoscopy Polyps

Applied Sciences ◽

10.3390/app9122404 ◽

2019 ◽

Vol 9 (12) ◽

pp. 2404 ◽

Cited By ~ 13

Author(s):

Sudhir Sornapudi ◽

Frank Meng ◽

Steven Yi

Keyword(s):

Capsule Endoscopy ◽

Wireless Capsule Endoscopy ◽

Automated Detection ◽

Fine Tuning ◽

Screening Tests ◽

Polyp Detection ◽

False Negatives ◽

Polyp Detection Rate ◽

Manual Examination ◽

Wireless Capsule

The early detection of polyps could help prevent colorectal cancer. The automated detection of polyps on the colon walls could reduce the number of false negatives that occur due to manual examination errors or polyps being hidden behind folds, and could also help doctors locate polyps from screening tests such as colonoscopy and wireless capsule endoscopy. Losing polyps may result in lesions evolving badly. In this paper, we propose a modified region-based convolutional neural network (R-CNN) by generating masks around polyps detected from still frames. The locations of the polyps in the image are marked, which assists the doctors examining the polyps. The features from the polyp images are extracted using pre-trained Resnet-50 and Resnet-101 models through feature extraction and fine-tuning techniques. Various publicly available polyp datasets are analyzed with various pertained weights. It is interesting to notice that fine-tuning with balloon data (polyp-like natural images) improved the polyp detection rate. The optimum CNN models on colonoscopy datasets including CVC-ColonDB, CVC-PolypHD, and ETIS-Larib produced values (F1 score, F2 score) of (90.73, 91.27), (80.65, 79.11), and (76.43, 78.70) respectively. The best model on the wireless capsule endoscopy dataset gave a performance of (96.67, 96.10). The experimental results indicate the better localization of polyps compared to recent traditional and deep learning methods.

Download Full-text