TY - JOUR
T1 - Quantitative assessment of endoscopic images for degree of villous atrophy in celiac disease
AU - Ciaccio, Edward J.
AU - Bhagat, Govind
AU - Tennyson, Christina A.
AU - Lewis, Suzanne K.
AU - Hernandez, Lincoln
AU - Green, Peter H.R.
N1 - Funding Information:
Acknowledgments The implementation of this study was made possible in part with a grant from the Celiac Sprue Association (CSA) Peer Review Research Grant Program.
PY - 2011/3
Y1 - 2011/3
N2 - Background: The degree of villous atrophy in celiac disease is difficult to assess at endoscopy. We sought to develop a quantitative technique for the evaluation of villous atrophy in endoscopic images. Method: In ten celiac patients as identified by standard endoscopy with biopsy, and ten control patients, standard and videocapsule endoscopic images of the duodenum were digitized. Subimages 7.5 × 7.5 mm2 in area from random locations within each image were assessed by measuring the length of mucosal fissures per unit area (L), and correlating L with the histologic grade of villous atrophy as determined by modified Marsh criteria. Results: Mean L values for standard endoscopic images were 37.8, 43.3, 64.1, and 83.5 mm for Marsh grades II, IIIa, IIIb, and IIIc, respectively. Mean L values for videocapsule images were 49.1, 50.0, 64.7, and 72.4 mm for Marsh grades II, IIIa, IIIb, and IIIc, respectively. Significant differences in the means existed between celiac images (Marsh scores II-IIIc) versus controls (p < 0.001) for both endoscopic and videocapsule images. There were no significant differences between measurements obtained from endoscopic versus videocapsule images. Conclusions: Quantified image analysis correlates with the histologic grade of villous atrophy, is automated, and lacks observer bias, thus lending itself to standardization.
AB - Background: The degree of villous atrophy in celiac disease is difficult to assess at endoscopy. We sought to develop a quantitative technique for the evaluation of villous atrophy in endoscopic images. Method: In ten celiac patients as identified by standard endoscopy with biopsy, and ten control patients, standard and videocapsule endoscopic images of the duodenum were digitized. Subimages 7.5 × 7.5 mm2 in area from random locations within each image were assessed by measuring the length of mucosal fissures per unit area (L), and correlating L with the histologic grade of villous atrophy as determined by modified Marsh criteria. Results: Mean L values for standard endoscopic images were 37.8, 43.3, 64.1, and 83.5 mm for Marsh grades II, IIIa, IIIb, and IIIc, respectively. Mean L values for videocapsule images were 49.1, 50.0, 64.7, and 72.4 mm for Marsh grades II, IIIa, IIIb, and IIIc, respectively. Significant differences in the means existed between celiac images (Marsh scores II-IIIc) versus controls (p < 0.001) for both endoscopic and videocapsule images. There were no significant differences between measurements obtained from endoscopic versus videocapsule images. Conclusions: Quantified image analysis correlates with the histologic grade of villous atrophy, is automated, and lacks observer bias, thus lending itself to standardization.
KW - Celiac disease
KW - Endoscopy
KW - Videocapsule
KW - Villous atrophy
UR - https://www.scopus.com/pages/publications/79952440268
U2 - 10.1007/s10620-010-1371-6
DO - 10.1007/s10620-010-1371-6
M3 - Article
C2 - 20844959
AN - SCOPUS:79952440268
SN - 0163-2116
VL - 56
SP - 805
EP - 811
JO - Digestive Diseases and Sciences
JF - Digestive Diseases and Sciences
IS - 3
ER -