Barrett esophagus: What to expect from Artificial Intelligence?

https://doi.org/10.1016/j.bpg.2021.101726Get rights and content

Abstract

The evaluation and assessment of Barrett’s esophagus is challenging for both expert and nonexpert endoscopists. However, the early diagnosis of cancer in Barrett’s esophagus is crucial for its prognosis, and could save costs. Pre-clinical and clinical studies on the application of Artificial Intelligence (AI) in Barrett’s esophagus have shown promising results. In this review, we focus on the current challenges and future perspectives of implementing AI systems in the management of patients with Barrett’s esophagus.

Introduction

In the past few years, the application of Artificial Intelligence (AI) in gastrointestinal (GI) endoscopy has been the subject of intensive research; significant progress has been made, especially in the domain of image and pattern recognition [1,2]. Ultimately, the goal of AI in GI endoscopy, as in all other aspects of medicine, is to improve the quality of patient care, by aiding physicians during diagnosis and treatment.

Section snippets

Artificial Intelligence in gastrointestinal endoscopy

AI and Computer-aided Diagnosis (CAD) have been studied in both benign as well as malignant disorders of the gastrointestinal tract (GIT), including Helicobacter pylori infection, peptic ulcer evaluation, detection and characterization of cancers in the esophagus and stomach, as well as in pancreaticobiliary disease [[3], [4], [5], [6], [7], [8], [9], [10], [11], [12]]. The greatest impact of AI in GI endoscopy, up till now, has been made in the detection and characterization of colonic

Technical aspects of Artificial Intelligence

The generic term “Artificial Intelligence” is established for all procedures where input data and a training procedure are used to solve a task, such as the prediction of an object class. In the training procedure, AI uses the input data to “learn”, with the main goal of improving the ability to process new data samples which were not initially part of the training data. This process of “machine learning” (ML) can be supervised, unsupervised or reinforced, depending on whether an algorithm is

Barrett’s esophagus and its challenges in daily practice

The incidence of BE and Barrett’s cancer (BC) has risen significantly in the past decade, and because of its close association with the metabolic syndrome, this trend is expected to continue [26,27]. The early diagnosis of BC is critical for its prognosis, and justifies the need for efficient surveillance, detection and characterization strategies. However, the detection of focal regions of dysplasia or early cancer, and the characterization of abnormalities within BE, can be challenging, even

Current data on the use of artificial intelligence in Barrett’s esophagus

As described above, the detection of focal regions of dysplasia or early adenocarcinoma in BE can be challenging, particularly for non-experts. When a focal lesion has been detected, the further differentiation between non-dysplasia or inflammation, low-grade dysplasia, high-grade dysplasia, and adenocarcinoma can be extremely difficult [34]. Mendel et al. and Ebigbo et al. were able to demonstrate the feasibility of using a deep learning approach in the detection and characterization of focal

What to expect from Artificial Intelligence in Barrett’s esophagus – future perspectives

The applications described above show the immense potential of AI in GI endoscopy, including BE. It seems obvious that more efficient tools are necessary to aid nonexpert endoscopists during the assessment of patients with BE. Computer-aided analysis and AI may be one such instrument. Because of its enormous potential, AI may not only influence the performance of endoscopists during endoscopic assessment of BE but may also improve the quality of post-procedure parameters.

In summary, and based

Practice points

  • Endoscopic assessment of BE is a challenge especially for nonexpert endoscopists.

  • Artificial intelligence has made significant progress in diagnostic endoscopy, especially in the identification of pathology (detection task).

  • Pre-clinical and clinical studies on the application of AI in BE assessment have shown excellent results and enormous future potential.

  • AI-assisted BE assessment could improve the quality of performance of endoscopists and also improve patient experience.

  • AI could have a

Research agenda

  • Large scale clinical trials with randomized and controlled or even tandem design.

  • Virtual reconstruction, intra- or postprocedural, of entire BE surface.

  • Staging of early neoplasia in BE, and better differentiation between T1a and T1b.

  • Cost-effectiveness of AI in BE management.

  • Effect of AI on training of nonexpert endoscopists.

Funding

None.

Declaration of competing interest

None.

References (41)

  • M.R. Struyvenberg

    Advanced imaging and sampling in barrett’s esophagus: artificial intelligence to the rescue?

    Gastrointest Endosc Clin N Am

    (2021)
  • P. Sinonquel

    Artificial Intelligence and its impact on quality improvement in upper and lower gastrointestinal endoscopy

    Dig Endosc

    (2021)
  • M. Alagappan

    Artificial intelligence in gastrointestinal endoscopy: the future is almost here

    World J Gastrointest Endosc

    (2018)
  • T. Hirasawa

    Application of artificial intelligence using a convolutional neural network for detecting gastric cancer in endoscopic images

    Gastric Cancer

    (2018)
  • T. Itoh

    Deep learning analyzes Helicobacter pylori infection by upper gastrointestinal endoscopy images

    Endosc Int Open

    (2018)
  • H. Nakashima

    Artificial intelligence diagnosis of Helicobacter pylori infection using blue laser imaging-bright and linked color imaging: a single-center prospective study

    Ann Gastroenterol

    (2018)
  • G.L. Wong

    Machine learning model to predict recurrent ulcer bleeding in patients with history of idiopathic gastroduodenal ulcer bleeding

    Aliment Pharmacol Ther

    (2019)
  • Y.Y. Zhao

    Computer-assisted diagnosis of early esophageal squamous cell carcinoma using narrow-band imaging magnifying endoscopy

    Endoscopy

    (2019)
  • T. Kuwahara

    Current status of artificial intelligence analysis for endoscopic ultrasonography

    Dig Endosc

    (2020)
  • T. Kuwahara

    Usefulness of deep learning analysis for the diagnosis of malignancy in intraductal papillary mucinous neoplasms of the pancreas

    Clin Transl Gastroenterol

    (2019)
  • View full text