Simultaneous Detection and Conversion Among Endoscopic Enhancement Modalities

Chaudhari U, Kolawole BB, Santacroce G, Zanmarchi I, Del Amor R, Meseguer P, Buda A, Bisschop R, Naranjo V, Ghosh S, Iacucci M, Grisan E, Bhandari P, De Hertogh G, Ferraz JG, Goetz M, Gui X, Hayee B, Kiesslich R, Metelli C, Lazarev M, Panaccione R, Parra-Blanco A, Pastorelli L, Rath T, Røyset ES, Vieth M, Villanacci V, Zardo D (2024)

Publication Type: Conference contribution

Publication year: 2024

Journal

Proceedings - International Symposium on Biomedical Imaging

Publisher: IEEE Computer Society

Conference Proceedings Title: Proceedings - International Symposium on Biomedical Imaging

Event location: Athens, GRC

ISBN: 9798350313338

DOI: 10.1109/ISBI56570.2024.10635724

Abstract

White Light Endoscopy (WLE) is the basic endoscopic imaging modality but has several limitations in enabling clinicians to detect and assess gastrointestinal diseases. More recently optical or digital imaging tools have been developed to improve the detection of subtle changes in different aspects of the mucosa: vessel, pit pattern, superficial abnormalities. These virtual chromoendoscopy (VCE) methods have proved to be superior to WLE in detecting gastric diseases and cancer precursors linked to inflammatory bowel disease (IBD). However, the different types of optical/digital chromoendoscopy are device-specific; hence, guidelines and scoring developed using one enhancement cannot be easily transferred to be used with others, cannot be applied to review past studies or endoscopic data recorded by devices from different manufacturers. The objective of this work is to develop a tool that enables endoscopists to view images that mimic the different enhancement independently of the acquisition device, possibly simultaneously on screen with baseline WLE.We propose to achieve this by first building a classifier that can detect the type of acquisition (no enhancement/WLE,VCE-IScan2, VCE-IScan3) and then apply a Cycle Consistent Adversarial Network (CycleGAN) to convert the acquired images to the missing modalities.The classifier can detect the correct acquisition type with 92% accuracy on the test set, thus using the wrong downstream CycleGANs on less than one frame every ten.

Authors with CRIS profile

Timo Rath Juniorprofessur für Molekulare Endoskopie Michael Vieth Professur für Allgemeine Pathologie am Medizincampus Oberfranken

Involved external institutions

London South Bank University (LSBU) GB United Kingdom (GB) University College Cork (UCC) IE Ireland (IE) Polytechnic University of Valencia / Universidad Politécnica de Valencia ES Spain (ES) European Society for Medical Oncology CH Switzerland (CH) University Hospital Leuven (UZ) / Universitaire ziekenhuizen Leuven BE Belgium (BE) Queen Alexandra Hospital GB United Kingdom (GB) University of Calgary CA Canada (CA) University of Washington US United States (USA) (US) King’s College London GB United Kingdom (GB) HELIOS Kliniken DE Germany (DE) Azienda Socio Sanitaria Territoriale degli Spedali Civili di Brescia IT Italy (IT) Johns Hopkins Hospital US United States (USA) (US) University of Nottingham GB United Kingdom (GB) NTNU Trondheim - Norwegian University of Science and Technology NO Norway (NO) Ospedale San Bortolo IT Italy (IT)

How to cite

APA:

Chaudhari, U., Kolawole, B.B., Santacroce, G., Zanmarchi, I., Del Amor, R., Meseguer, P.,... Zardo, D. (2024). Simultaneous Detection and Conversion Among Endoscopic Enhancement Modalities. In Proceedings - International Symposium on Biomedical Imaging. Athens, GRC: IEEE Computer Society.

MLA:

Chaudhari, Ujwala, et al. "Simultaneous Detection and Conversion Among Endoscopic Enhancement Modalities." Proceedings of the 21st IEEE International Symposium on Biomedical Imaging, ISBI 2024, Athens, GRC IEEE Computer Society, 2024.

BibTeX: Download