Intrapapillary capillary loop classification in magnification endoscopy: open dataset and baseline methodology

Luis C García-Peraza-Herrera,Tom Vercauteren,Martin Everson,Wen Lun Wang,Hsiu-Po Wang,Danail Stoyanov,Laurence Lovat,Rehan Haidry,Sébastien Ourselin

doi:10.1007/s11548-020-02127-w

Abstract

PurposeEarly squamous cell neoplasia (ESCN) in the oesophagus is a highly treatable condition. Lesions confined to the mucosal layer can be curatively treated endoscopically. We build a computer-assisted detection system that can classify still images or video frames as normal or abnormal with high diagnostic accuracy.MethodsWe present a new benchmark dataset containing 68K binary labelled frames extracted from 114 patient videos whose imaged areas have been resected and correlated to histopathology. Our novel convolutional network architecture solves the binary classification task and explains what features of the input domain drive the decision-making process of the network.ResultsThe proposed method achieved an average accuracy of 91.7% compared to the 94.7% achieved by a group of 12 senior clinicians. Our novel network architecture produces deeply supervised activation heatmaps that suggest the network is looking at intrapapillary capillary loop patterns when predicting abnormality.ConclusionWe believe that this dataset and baseline method may serve as a reference for future benchmarks on both video frame classification and explainability in the context of ESCN detection. A future work path of high clinical relevance is the extension of the classification to ESCN types.

Highlights

This work was supported through an Innovative Engineering for Health award by Wellcome Trust (WT101957); Engineering and Physical Sciences Research Council (EPSRC) (NS/A00027/1) and a Wellcome/EPSRC Centre award (203145Z/16/Z and NS/A000050/1).Electronic supplementary material The online version of this article contains supplementary material, which is available to authorized users.Oesophageal cancer is the sixth most common cause of cancer deaths worldwide [16] and a burgeoning health issue in developing nations from Africa along a ‘cancer belt’ to China
Squamous cell neoplasia (ESCN) is a highly treatable type of oesophageal cancer, with recent advances in endoscopic therapy meaning that lesions confined to the mucosal layer can be curatively resected endoscopically with a < 2% incidence of local lymph node metastasis [1]
No conclusive evidence has been found that it is paying attention to large deep submucosal vessels to detect normal tissue. We believe that this baseline method may serve as a reference for future benchmarks on both video frame classification and explainability in the context of Early squamous cell neoplasia (ESCN) detection

Summary

Introduction

Oesophageal cancer is the sixth most common cause of cancer deaths worldwide [16] and a burgeoning health issue in developing nations from Africa along a ‘cancer belt’ to China. The current gold standard to investigate oesophageal cancer is gastroscopy with biopsies for histological analysis. Squamous cell neoplasia (ESCN) is a highly treatable type of oesophageal cancer, with recent advances in endoscopic therapy meaning that lesions confined to the mucosal layer can be curatively resected endoscopically with a < 2% incidence of local lymph node metastasis [1]. The endoscopic appearances of ESCN lesions are subtle and missed, with significant miss rates on endoscopy within the 3 years preceding diagnosis [10]. Cancers invading into the submucosa are likely to have local lymph node metastasis

Objectives

Methods

Results

Conclusion