Colorectal cancer is the second leading cause of cancer-related deaths in the US, claiming more than 50,000 lives in 2006. Colonoscopy is currently the preferred screening modality for colorectal cancer. However, recent data suggest that there is a significant (4-12 percent) miss-rate associated with colonoscopy for the detection of even large polyps and cancers. In 2006, the American College of Gastroenterology (ACG) and the American Society of Gastroenterology (ASGE) published consensus guidelines defining a good quality colonoscopy but several recent reports show that adherence to these guidelines varies among endoscopists. To complicate matters, detailed documentation of quality-related aspects of colonoscopy is complex, labor-intensive and expensive and until now has to be done manually. We hypothesize that computer algorithms can generate quality-related metrics from video files obtained during colonoscopy, that metrics are different for beginning versus experienced endoscopists, and that awareness of monitoring alters endoscopist behavior. Using novel computer algorithms developed specifically for generation of quality-related metrics, we propose to address three Specific Aims. First, we will test whether computer-derived metrics reflect quality of colonoscopy as defined by, but not limited to, the ACG and ASGE guidelines. Second, we will test whether computer-derived metrics reflecting quality differ among beginning and experienced endoscopists. And third, we will test whether awareness of automated quality monitoring alters endoscopic behavior towards best possible adherence to recommended ACG and ASGE guidelines. Successful evaluation and implementation of the proposed, automated system has the potential to improve the quality of care of over 14 million US citizens - the approximate number of people undergoing colonoscopy - on an annual basis. In addition, the technology lends itself for rapid adaptation to other endoscopic medical procedures such as upper gastrointestinal endoscopy, cystoscopy, arthroscopy and bronchoscopy.

Public Health Relevance

Our long-term objective is to ensure that all endoscopic examinations of the colon, so called colonoscopies, are of excellent quality. Unfortunately, at present this is not the case and """"""""missed lesions"""""""", i.e. large polyps or colorectal cancer, sometimes are detected shortly after a """"""""negative"""""""" colonoscopy. Several patient-, equipment- and endoscopist-related factors may be responsible for this. We and others believe that the endoscopist-related factors are most important as the endoscopist can mitigate unfavorable conditions related to patient- or equipment-related conditions. In this application we show, that we have developed software tools able to derive metrics about endoscopist-related factors from video files obtained during colonoscopy. We intend to study whether the metrics derived by these software tools indeed measure colonoscopy quality. This year, approximately 14 million US citizens will undergo colonoscopy for a total cost of around $24 billion and 55,000 patients will die of colorectal cancer. We think that effective methods to ensure an excellent quality colonoscopy will significantly decrease """"""""missed lesions"""""""" and expect that insights gained from our proposal will ultimately lead to improved detection and prevention of colorectal cancer.

Agency
National Institute of Health (NIH)
Institute
National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)
Type
Research Project (R01)
Project #
1R01DK083745-01A1
Application #
7649076
Study Section
Nursing Science: Adults and Older Adults Study Section (NSAA)
Program Officer
Hamilton, Frank A
Project Start
2009-09-01
Project End
2012-05-31
Budget Start
2009-09-01
Budget End
2010-05-31
Support Year
1
Fiscal Year
2009
Total Cost
$238,105
Indirect Cost
Name
Mayo Clinic, Rochester
Department
Type
DUNS #
006471700
City
Rochester
State
MN
Country
United States
Zip Code
55905
Hong, DongHo; Tavanapong, Wallapak; Wong, Johnny et al. (2014) 3D Reconstruction of virtual colon structures from colonoscopy images. Comput Med Imaging Graph 38:22-33
Nawarathna, Ruwan; Oh, JungHwan; Muthukudage, Jayantha et al. (2014) Abnormal Image Detection in Endoscopy Videos Using a Filter Bank and Local Binary Patterns. Neurocomputing 144:70-91
Samuel, Sunil; Bruining, David H; Loftus Jr, Edward V et al. (2013) Validation of the ulcerative colitis colonoscopic index of severity and its correlation with disease activity measures. Clin Gastroenterol Hepatol 11:49-54.e1
Stanek, Sean R; Tavanapong, Wallapak; Wong, Johnny et al. (2012) Automatic real-time detection of endoscopic procedures using temporal features. Comput Methods Programs Biomed 108:524-35
Edakkanambeth Varayil, J; Smyrk, T C; de Groen, P C (2011) Five interval colorectal cancers. Endoscopy 43 Suppl 2 UCTN:E203-4
Wang, Yi; Tavanapong, Wallapak; Wong, Johnny S et al. (2010) Detection of quality visualization of appendiceal orifices using local edge cross-section profile features and near pause detection. IEEE Trans Biomed Eng 57:685-95
de Groen, Piet C (2010) Advanced systems to assess colonoscopy. Gastrointest Endosc Clin N Am 20:699-716