Colonoscopy is currently the preferred screening modality for prevention of colorectal cancer. However, the effectiveness of colonoscopy depends on the quality of the procedure, which depends on several factors. In this paper, we present new methods that derive a new quality metric for automated scoring of quality of mucosa inspection performed by the endoscopist. We conducted Pearson's Correlation analysis of the computerized metric scores against the averages of the manual scores given by four domain experts on twenty-one colonoscopy videos. Our metric shows a relatively strong positive correlation (Pearson's correlation coefficient of 0.72) between the computer-generated score and the ground truth. Hence, the proposed work is very promising to be used for quality control/assurance in routine colonoscopy screening.
Bibliographical noteFunding Information:
This work is partially supported by Agencyfor Healthcare Research and Quality (AHRQ HS17537), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK DK083745), the Mayo Clinic, and Iowa State University Research Foundation. Any opinions, findings, conclusions, or recommendations expressed in this paper are those of authors. They do not necessarily reflect the views of the funding agencies. Johnny Wong, Wallapak Tavanapong and JungHwan Oh hold positions at EndoMetric, Corporation, Ames, IA 50014, U.S.A, a for profit company that markets endoscopy-related software. Johnny Wong, Wallapak Tavanapong, JungHwan Oh, Piet de Groen, and Mayo Clinic own stocks in EndoMetric. Piet de Groen, Johnny Wong, and Wallapak Tavanapong have received royalty payments from EndoMetric.
- Image analysis
- Objective quality metrics
- Quality of colonoscopy