Celiac.com 06/23/2010 - A team of researchers evaluated the possibility of diagnosing celiac disease using quantitative analysis of videocapsule endoscopy images.
The team included Edward J. Ciaccioa, Christina A. Tennysonb, Suzanne K. Lewisb, Suneet, Krishnareddy, Govind Bhagat, and Peter H.R. Green. They are variously associated with the Department of Pharmacology, Department of Medicine, Department of Pathology, Columbia University College of Physicians and Surgeons in New York.
The team gathered capsule endoscopy images from a group of 11 celiac patients with small bowel pathology, and from a group of 10 control patients.
Images had a resolution of 576Ã—576 pixels, with 256 grayscale tones, and a frame-rate of 2 sâˆ’1. The team measured over 10Ã—10 pixel sub-images for pixel brightness and image texture. They then averaged the results for for 56Ã—56 sub-images per frame.
For each patient, the team took measurements at from five locations in the proximal to distal small intestine. At each location, they figured measurements using 200 consecutive image frames (100 s).
For classification with a nonlinear discriminant function, they computed mean frame-to-frame pixel brightness, image texture, periodicity in brightness, and estimated wall motion or intestinal motility.
By pooling the data, the team found that images from the celiac group showed greater texture than did images from control group (p < 0.001).
Images from the celiac disease group exhibited more frame-to-frame brightness variation as well (p = 0.032). Celiac patients showed longer dominant period of brightness in celiacs (p = 0.001), which may indicate reduced motility.
Markers for three-dimensional nonlinear classification of celiacs versus controls showed sensitivity of 92.7% and specificity of 93.5%. Both celiac patients and control subjects showed an approximately linear association between dominant period and small intestinal transit time (r2 = 0.42 and r2 =0 .55, respectively).
The results show that videocapsule images can be used to reveal villous atrophy throughout the small intestine, and to distinguish individuals with celiac disease from individuals without mucosal atrophy.