Scope of Artificial Intelligence in Screening and Diagnosis of Colorectal Cancer

Hemant Goyal,Sumant Inamdar,Rupinder Mann,Abhilash Perisetti,Aman Ali,Zainab Gandhi,Neil Sharma,Benjamin Tharian,Shreyas Saligram,Khizar Aman Ali

doi:10.3390/jcm9103313

Abstract

Globally, colorectal cancer is the third most diagnosed malignancy. It causes significant mortality and morbidity, which can be reduced by early diagnosis with an effective screening test. Integrating artificial intelligence (AI) and computer-aided detection (CAD) with screening methods has shown promising colorectal cancer screening results. AI could provide a “second look” for endoscopists to decrease the rate of missed polyps during a colonoscopy. It can also improve detection and characterization of polyps by integration with colonoscopy and various advanced endoscopic modalities such as magnifying narrow-band imaging, endocytoscopy, confocal endomicroscopy, laser-induced fluorescence spectroscopy, and magnifying chromoendoscopy. This descriptive review discusses various AI and CAD applications in colorectal cancer screening, polyp detection, and characterization.

Highlights

Artificial intelligence (AI) translates as having a computer program that mimics humans’ learning and problem-solving capability [1]
This study showed that Machine Learning (ML) algorithm tools such as MeScore could be used to identify high-risk patients who should undergo screening colonoscopy [26]
Images were classified using a conditional random field (CRF) model for colonoscopic polyp detection and showed method had 86% sensitivity and 85% specificity when evaluated on a feature training set of 11,802 images from 35 colonoscopy videos with accompanying endoscopy reports

Summary

Introduction

Artificial intelligence (AI) translates as having a computer program that mimics humans’ learning and problem-solving capability [1]. Machine Learning (ML) and Deep Learning (DL) represent the virtual branch of AI in medicine [3]. The ML comprises of unsupervised and supervised learning. In supervised ML, the machine is fed with input (individual descriptions) and output (an outcome of interest) data. Machine learning is broadly classified into supervised, unsupervised, reinforcement, and active learning tasks. Supervised learning involves input data with target labels to learn a pattern. Decision trees, linear discriminants, support vector machines, logistic regression, and artificial neural networks are different models of supervised ML [4]. Unsupervised learning involves recognizing patterns from the input data without previously defined target labels. Reinforcement learning involves training intelligent agents to enhance their performance [5]

Objectives

Results

Conclusion