PT Journal
AU Cristina Sanchez Montes
   F. Javier Sanchez
   Jorge Bernal
   Henry Cordova
   Maria Lopez Ceron
   Miriam Cuatrecasas
   Cristina Rodriguez de Miguel
   Ana Garcia Rodriguez
   Rodrigo Garces Duran
   Maria Pellise
   Josep Llach
   Gloria Fernandez Esparrach
TI Computer-aided Prediction of Polyp Histology on White-Light Colonoscopy using Surface Pattern Analysis
SO Endoscopy
JI END
PY 2019
BP 261
EP 265
VL 51
IS 3
DI 10.1055/a-0732-5250
AB Background and study aims: To evaluate a new computational histology prediction system based on colorectal polyp textural surface patterns using high definition white light images.Patients and methods: Textural elements (textons) were characterized according to their contrast with respect to the surface, shape and number of bifurcations, assuming that dysplastic polyps are associated with highly contrasted, large tubular patterns with some degree of bifurcation. Computer-aided diagnosis (CAD) was compared with pathological diagnosis and the diagnosis by the endoscopists using Kudo and NICE classification.Results: Images of 225 polyps were evaluated (142 dysplastic and 83 non-dysplastic). CAD system correctly classified 205 (91.1%) polyps, 131/142 (92.3%) dysplastic and 74/83 (89.2%) non-dysplastic. For the subgroup of 100 diminutive (<5 mm) polyps, CAD correctly classified 87 (87%) polyps, 43/50 (86%) dysplastic and 44/50 (88%) non-dysplastic. There were not statistically significant differences in polyp histology prediction based on CAD system and on endoscopist assessment.Conclusion: A computer vision system based on the characterization of the polyp surface in the white light accurately predicts colorectal polyp histology.
ER