Brief Article
Copyright ©2012 Baishideng Publishing Group Co., Limited. All rights reserved.
World J Gastroenterol. Oct 21, 2012; 18(39): 5560-5569
Published online Oct 21, 2012. doi: 10.3748/wjg.v18.i39.5560
Software for automated classification of probe-based confocal laser endomicroscopy videos of colorectal polyps
Barbara André, Tom Vercauteren, Anna M Buchner, Murli Krishna, Nicholas Ayache, Michael B Wallace
Barbara André, Tom Vercauteren, Image Computing Group, Mauna Kea Technologies, 75010 Paris, France
Barbara André, Nicholas Ayache, Asclepios Research Team, The National Institute for Research in Computer Science and Control Sophia Antipolis, 06902 Sophia Antipolis, France
Anna M Buchner, Division of Gastroenterology, Hospital of the University of Pennsylvania, Philadelphia, PA 19104, United States
Murli Krishna, Laboratory Medicine and Pathology, Mayo Clinic Hospital in Jacksonville, Jacksonville, FL 3224, United States
Michael B Wallace, Department of Gastroenterology and Hepatology, Mayo Clinic Hospital in Jacksonville, Jacksonville, FL 3224, United States
Author contributions: André B was involved in study concept and design, analysis and interpretation of data, drafting of the manuscript and statistical analysis; Vercauteren T was involved in study concept and design, analysis and interpretation of data and study supervision; Buchner AM was involved in acquisition of data and analysis and interpretation of data; Krishna M was involved in analysis and interpretation of data; Ayache N was involved in study concept and design, analysis and interpretation of data and study supervision; Wallace MB was involved in acquisition of data, analysis and interpretation of data and study supervision.
Correspondence to: Barbara André, PhD, Image Computing Group, Mauna Kea Technologies, 9 rue d’Enghien, 75010 Paris, France. barbara.andre@maunakeatech.com
Telephone: +33-1-70080961  Fax: +33-1-48241218
Received: November 15, 2011
Revised: May 9, 2012
Accepted: May 26, 2012
Published online: October 21, 2012
Abstract

AIM: To support probe-based confocal laser endomicroscopy (pCLE) diagnosis by designing software for the automated classification of colonic polyps.

METHODS: Intravenous fluorescein pCLE imaging of colorectal lesions was performed on patients undergoing screening and surveillance colonoscopies, followed by polypectomies. All resected specimens were reviewed by a reference gastrointestinal pathologist blinded to pCLE information. Histopathology was used as the criterion standard for the differentiation between neoplastic and non-neoplastic lesions. The pCLE video sequences, recorded for each polyp, were analyzed off-line by 2 expert endoscopists who were blinded to the endoscopic characteristics and histopathology. These pCLE videos, along with their histopathology diagnosis, were used to train the automated classification software which is a content-based image retrieval technique followed by k-nearest neighbor classification. The performance of the off-line diagnosis of pCLE videos established by the 2 expert endoscopists was compared with that of automated pCLE software classification. All evaluations were performed using leave-one-patient-out cross-validation to avoid bias.

RESULTS: Colorectal lesions (135) were imaged in 71 patients. Based on histopathology, 93 of these 135 lesions were neoplastic and 42 were non-neoplastic. The study found no statistical significance for the difference between the performance of automated pCLE software classification (accuracy 89.6%, sensitivity 92.5%, specificity 83.3%, using leave-one-patient-out cross-validation) and the performance of the off-line diagnosis of pCLE videos established by the 2 expert endoscopists (accuracy 89.6%, sensitivity 91.4%, specificity 85.7%). There was very low power (< 6%) to detect the observed differences. The 95% confidence intervals for equivalence testing were: -0.073 to 0.073 for accuracy, -0.068 to 0.089 for sensitivity and -0.18 to 0.13 for specificity. The classification software proposed in this study is not a “black box” but an informative tool based on the query by example model that produces, as intermediate results, visually similar annotated videos that are directly interpretable by the endoscopist.

CONCLUSION: The proposed software for automated classification of pCLE videos of colonic polyps achieves high performance, comparable to that of off-line diagnosis of pCLE videos established by expert endoscopists.

Keywords: Colorectal neoplasia; Computer-aided diagnosis; Content-based image retrieval; Nearest neighbor classification software; Probe-based confocal laser endomicroscopy