Computer-aided classification of colorectal polyps using blue-light and linked-color imaging

T. Scheeve, Ramon-Michel Schreuder, F. van der Sommen, Joep E.G. IJspeert, Evelien Dekker, Erik J. Schoon, P.H.N. de With

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

Abstract

Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths. Since most CRCs develop from colorectal polyps (CRPs), accurate endoscopic differentiation facilitates decision making on resection of CRPs, thereby increasing cost-efficiency and reducing patient risk. Current classification systems based on whitelight imaging (WLI) or narrow-band imaging (NBI) have limited predictive power, or they do not consider sessile serrated adenomas/polyps (SSA/Ps), although these cause up to 30% of all CRCs. To better differentiate adenomas, hyperplastic polyps, and SSA/Ps, this paper explores the feasibility of two approaches: (1) an accurate computer-aided diagnosis (CADx) system for automated diagnosis of CRPs, and (2) novel endoscopic imaging techniques like blue-light imaging (BLI) and linked-color imaging (LCI). Two methods are explored to predict histology: (1) direct classification using a support vector machine (SVM) classifier, and (2) classification via a clinical classification model (WASP classification) combined with an SVM. The use of probabilistic features of SVM facilitates objective quantification of the detailed classification process. Automated differentiation of colonic polyp subtypes reaches accuracies of 78−96%, thereby improving medical expert results by 4−20%. Diagnostic accuracy for directly predicting adenomatous from hyperplastic histology reaches 93% and 87−90% using NBI and the novel BLI and LCI techniques, respectively, thus improving medical expert results by 26% and 20−23%, respectively. Predicting adenomatous histology in diminutive polyps with high confidence yields NPVs of 100%, clearly satisfying the PIVI guideline recommendation on endoscopic innovations (≥90% NPV). Our CADx system outperforms clinicians, while the novel BLI technique adds performance value.
LanguageEnglish
Title of host publicationMedical Imaging 2019: Computer-Aided Diagnosis
Subtitle of host publicationComputer-Aided Diagnosis
EditorsKensaku Mori, Horst K. Hahn
Place of PublicationBellingham
PublisherSPIE
Number of pages8
ISBN (Electronic)9781510625471
DOIs
StatePublished - 13 Mar 2019
EventMedical Imaging 2019: Computer-Aided Diagnosis - San Diego, United States
Duration: 17 Feb 201920 Feb 2019

Publication series

NameProceedings of SPIE
Volume10950
ISSN (Print)0277-786X

Conference

ConferenceMedical Imaging 2019: Computer-Aided Diagnosis
CountryUnited States
CitySan Diego
Period17/02/1920/02/19

Fingerprint

Polyps
Color
Light
Narrow Band Imaging
Adenoma
Histology
Colonic Polyps
Colorectal Neoplasms
Decision Making
Guidelines
Costs and Cost Analysis
Support Vector Machine
Neoplasms

Keywords

  • Biomedical optical imaging
  • Classification
  • Colorectal polyps
  • Gastroenterology
  • Machine learning

Cite this

Scheeve, T., Schreuder, R-M., van der Sommen, F., IJspeert, J. E. G., Dekker, E., Schoon, E. J., & de With, P. H. N. (2019). Computer-aided classification of colorectal polyps using blue-light and linked-color imaging. In K. Mori, & H. K. Hahn (Eds.), Medical Imaging 2019: Computer-Aided Diagnosis: Computer-Aided Diagnosis [1095012] (Proceedings of SPIE; Vol. 10950). Bellingham: SPIE. DOI: 10.1117/12.2508223
Scheeve, T. ; Schreuder, Ramon-Michel ; van der Sommen, F. ; IJspeert, Joep E.G. ; Dekker, Evelien ; Schoon, Erik J. ; de With, P.H.N./ Computer-aided classification of colorectal polyps using blue-light and linked-color imaging. Medical Imaging 2019: Computer-Aided Diagnosis: Computer-Aided Diagnosis. editor / Kensaku Mori ; Horst K. Hahn. Bellingham : SPIE, 2019. (Proceedings of SPIE).
@inproceedings{b0d54f3f0d6741e09a635c5d00e0ad46,
title = "Computer-aided classification of colorectal polyps using blue-light and linked-color imaging",
abstract = "Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths. Since most CRCs develop from colorectal polyps (CRPs), accurate endoscopic differentiation facilitates decision making on resection of CRPs, thereby increasing cost-efficiency and reducing patient risk. Current classification systems based on whitelight imaging (WLI) or narrow-band imaging (NBI) have limited predictive power, or they do not consider sessile serrated adenomas/polyps (SSA/Ps), although these cause up to 30{\%} of all CRCs. To better differentiate adenomas, hyperplastic polyps, and SSA/Ps, this paper explores the feasibility of two approaches: (1) an accurate computer-aided diagnosis (CADx) system for automated diagnosis of CRPs, and (2) novel endoscopic imaging techniques like blue-light imaging (BLI) and linked-color imaging (LCI). Two methods are explored to predict histology: (1) direct classification using a support vector machine (SVM) classifier, and (2) classification via a clinical classification model (WASP classification) combined with an SVM. The use of probabilistic features of SVM facilitates objective quantification of the detailed classification process. Automated differentiation of colonic polyp subtypes reaches accuracies of 78−96{\%}, thereby improving medical expert results by 4−20{\%}. Diagnostic accuracy for directly predicting adenomatous from hyperplastic histology reaches 93{\%} and 87−90{\%} using NBI and the novel BLI and LCI techniques, respectively, thus improving medical expert results by 26{\%} and 20−23{\%}, respectively. Predicting adenomatous histology in diminutive polyps with high confidence yields NPVs of 100{\%}, clearly satisfying the PIVI guideline recommendation on endoscopic innovations (≥90{\%} NPV). Our CADx system outperforms clinicians, while the novel BLI technique adds performance value.",
keywords = "Biomedical optical imaging, Classification, Colorectal polyps, Gastroenterology, Machine learning",
author = "T. Scheeve and Ramon-Michel Schreuder and {van der Sommen}, F. and IJspeert, {Joep E.G.} and Evelien Dekker and Schoon, {Erik J.} and {de With}, P.H.N.",
year = "2019",
month = "3",
day = "13",
doi = "10.1117/12.2508223",
language = "English",
series = "Proceedings of SPIE",
publisher = "SPIE",
editor = "Kensaku Mori and Hahn, {Horst K.}",
booktitle = "Medical Imaging 2019: Computer-Aided Diagnosis",
address = "United States",

}

Scheeve, T, Schreuder, R-M, van der Sommen, F, IJspeert, JEG, Dekker, E, Schoon, EJ & de With, PHN 2019, Computer-aided classification of colorectal polyps using blue-light and linked-color imaging. in K Mori & HK Hahn (eds), Medical Imaging 2019: Computer-Aided Diagnosis: Computer-Aided Diagnosis., 1095012, Proceedings of SPIE, vol. 10950, SPIE, Bellingham, Medical Imaging 2019: Computer-Aided Diagnosis, San Diego, United States, 17/02/19. DOI: 10.1117/12.2508223

Computer-aided classification of colorectal polyps using blue-light and linked-color imaging. / Scheeve, T.; Schreuder, Ramon-Michel; van der Sommen, F.; IJspeert, Joep E.G.; Dekker, Evelien; Schoon, Erik J.; de With, P.H.N.

Medical Imaging 2019: Computer-Aided Diagnosis: Computer-Aided Diagnosis. ed. / Kensaku Mori; Horst K. Hahn. Bellingham : SPIE, 2019. 1095012 (Proceedings of SPIE; Vol. 10950).

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Computer-aided classification of colorectal polyps using blue-light and linked-color imaging

AU - Scheeve,T.

AU - Schreuder,Ramon-Michel

AU - van der Sommen,F.

AU - IJspeert,Joep E.G.

AU - Dekker,Evelien

AU - Schoon,Erik J.

AU - de With,P.H.N.

PY - 2019/3/13

Y1 - 2019/3/13

N2 - Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths. Since most CRCs develop from colorectal polyps (CRPs), accurate endoscopic differentiation facilitates decision making on resection of CRPs, thereby increasing cost-efficiency and reducing patient risk. Current classification systems based on whitelight imaging (WLI) or narrow-band imaging (NBI) have limited predictive power, or they do not consider sessile serrated adenomas/polyps (SSA/Ps), although these cause up to 30% of all CRCs. To better differentiate adenomas, hyperplastic polyps, and SSA/Ps, this paper explores the feasibility of two approaches: (1) an accurate computer-aided diagnosis (CADx) system for automated diagnosis of CRPs, and (2) novel endoscopic imaging techniques like blue-light imaging (BLI) and linked-color imaging (LCI). Two methods are explored to predict histology: (1) direct classification using a support vector machine (SVM) classifier, and (2) classification via a clinical classification model (WASP classification) combined with an SVM. The use of probabilistic features of SVM facilitates objective quantification of the detailed classification process. Automated differentiation of colonic polyp subtypes reaches accuracies of 78−96%, thereby improving medical expert results by 4−20%. Diagnostic accuracy for directly predicting adenomatous from hyperplastic histology reaches 93% and 87−90% using NBI and the novel BLI and LCI techniques, respectively, thus improving medical expert results by 26% and 20−23%, respectively. Predicting adenomatous histology in diminutive polyps with high confidence yields NPVs of 100%, clearly satisfying the PIVI guideline recommendation on endoscopic innovations (≥90% NPV). Our CADx system outperforms clinicians, while the novel BLI technique adds performance value.

AB - Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths. Since most CRCs develop from colorectal polyps (CRPs), accurate endoscopic differentiation facilitates decision making on resection of CRPs, thereby increasing cost-efficiency and reducing patient risk. Current classification systems based on whitelight imaging (WLI) or narrow-band imaging (NBI) have limited predictive power, or they do not consider sessile serrated adenomas/polyps (SSA/Ps), although these cause up to 30% of all CRCs. To better differentiate adenomas, hyperplastic polyps, and SSA/Ps, this paper explores the feasibility of two approaches: (1) an accurate computer-aided diagnosis (CADx) system for automated diagnosis of CRPs, and (2) novel endoscopic imaging techniques like blue-light imaging (BLI) and linked-color imaging (LCI). Two methods are explored to predict histology: (1) direct classification using a support vector machine (SVM) classifier, and (2) classification via a clinical classification model (WASP classification) combined with an SVM. The use of probabilistic features of SVM facilitates objective quantification of the detailed classification process. Automated differentiation of colonic polyp subtypes reaches accuracies of 78−96%, thereby improving medical expert results by 4−20%. Diagnostic accuracy for directly predicting adenomatous from hyperplastic histology reaches 93% and 87−90% using NBI and the novel BLI and LCI techniques, respectively, thus improving medical expert results by 26% and 20−23%, respectively. Predicting adenomatous histology in diminutive polyps with high confidence yields NPVs of 100%, clearly satisfying the PIVI guideline recommendation on endoscopic innovations (≥90% NPV). Our CADx system outperforms clinicians, while the novel BLI technique adds performance value.

KW - Biomedical optical imaging

KW - Classification

KW - Colorectal polyps

KW - Gastroenterology

KW - Machine learning

UR - http://www.scopus.com/inward/record.url?scp=85068159793&partnerID=8YFLogxK

U2 - 10.1117/12.2508223

DO - 10.1117/12.2508223

M3 - Conference contribution

T3 - Proceedings of SPIE

BT - Medical Imaging 2019: Computer-Aided Diagnosis

PB - SPIE

CY - Bellingham

ER -

Scheeve T, Schreuder R-M, van der Sommen F, IJspeert JEG, Dekker E, Schoon EJ et al. Computer-aided classification of colorectal polyps using blue-light and linked-color imaging. In Mori K, Hahn HK, editors, Medical Imaging 2019: Computer-Aided Diagnosis: Computer-Aided Diagnosis. Bellingham: SPIE. 2019. 1095012. (Proceedings of SPIE). Available from, DOI: 10.1117/12.2508223