Colorectal cancer

Modelling dysplasia detection in ulcerative colitis: clinical implications of surveillance intensity

D Awais¹, C A Siegel¹, P D R Higgins²

¹ Dartmouth–Hitchcock Medical Center, Lebanon, New Hampshire, USA
² SPC 5682, 1150 West Medical Center Drive, Ann Arbor, Michigan, USA

^{Correspondence to} Dr P D R Higgins, SPC 5682, 1150 West Medical Center Drive, Ann Arbor, MI 48109, USA; phiggins{at}umich.edu

Background: Endoscopic surveillance of chronic colitis uses random biopsies to find dysplastic fields. Enhanced endoscopic methods are more sensitive for dysplasia detection, but their specificity for colorectal cancer risk is unknown.

Aims: To develop a mathematical model of the sensitivity of random biopsy surveillance, and determine the implications of negative, a single positive, or multiple positive biopsies for dysplasia, and compare the detection threshold to that detectable by enhanced endoscopy.

Methods: Using mathematical modelling, we calculated the confidence level with which dysplasia can be excluded, the dysplastic field size detection threshold, the predicted area of a dysplastic field, and the number of biopsies needed for a given dysplasia detection threshold and confidence level.

Results: 32 random biopsies provide only 80% confidence that dysplasia involving 5% of the colon can be detected. When a single biopsy of 18 is dysplastic, this predicts a dysplastic area (89 cm²) several orders of magnitude greater than dysplastic fields that are readily detectable by enhanced endoscopy (1 cm diameter), and the predicted field size increases rapidly with multiple positive biopsies.

Conclusions: Random biopsy surveillance is sufficiently sensitive to detect large dysplastic fields with significant colorectal cancer risk. Enhanced endoscopy can detect much smaller dysplastic fields, but these have unknown (perhaps much lower) colorectal cancer risk. Small dysplastic fields should not be assumed to indicate a high colorectal cancer risk that warrants colectomy. Prospective studies are needed to define the colorectal cancer risk and optimal management of small dysplastic lesions.

CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    What's this?