Basic Study
Copyright ©The Author(s) 2022. Published by Baishideng Publishing Group Inc. All rights reserved.
Artif Intell Cancer. Apr 28, 2022; 3(2): 27-41
Published online Apr 28, 2022. doi: 10.35713/aic.v3.i2.27
Learning models for colorectal cancer signature reconstruction and classification in patients with chronic inflammatory bowel disease
Mariem Abaach, Ian Morilla
Mariem Abaach, Mathématiques Appliquées à Paris 5, Unité mixte de Recherche, Centre National de la Recherche Scientifique, Université de Paris, Paris 75006, France
Ian Morilla, Laboratoire Analyse, Géométrie et Applications, Centre National de la Recherche Scientifique (Unité mixte de Recherche), Université Sorbonne Paris Nord, Villetaneuse, Paris 93430, France
Author contributions: Morilla I conceived and designed the computational experiments; Abaach M and Morilla I performed computational experiments, analyzed the miRNomic data, performed formal analysis; Morilla I wrote the original manuscript Abaach M and Morilla I reviewed and edited the manuscript.
Institutional review board statement: The protocols involving human participants conformed to the local Ethics Committee (CPP-Île de France IV No. 2009/17) and to the principles set out in the WMA Declaration of Helsinki, and the Belmont Report from the Department of Health and Human Services. Human ileal biopsies were obtained from the IBD Gastroenterology Unit, Beaujon Hospital and a written informed consent was obtained from all the patients before inclusion in the study.
Institutional animal care and use committee statement: The protocols involving human participants conformed to the local Ethics Committee (CPP-Île de France IV No. 2009/17) and to the principles set out in the WMA Declaration of Helsinki, and the Belmont Report from the Department of Health and Human Services. Human ileal biopsies were obtained from the IBD Gastroenterology Unit, Beaujon Hospital and a written informed consent was obtained from all the patients before inclusion in the study.
Conflict-of-interest statement: All authors declare no conflicts of interest in this paper.
Data sharing statement: The R code for implementing the inference procedures is available at https://figshare.com/account/home#/projects/36290. The results of the in-ference, along with instructions on how to use these files to recreate the figures in this paper, are available at https://figshare.com/account/home#/projects/36290/.
ARRIVE guidelines statement: The authors have read the ARRIVE guidelines, and the manuscript was prepared and revised according to the ARRIVE guidelines.
Open-Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Ian Morilla, PhD, Assistant Professor, Research Associate, Laboratoire Analyse, Géométrie et Applications, Centre National de la Recherche Scientifique (Unité mixte de Recherche), Université Sorbonne Paris Nord, 99 avenue Jean Baptiste clément, Villetaneuse, Paris 93430, France. morilla@math.univ-paris13.fr
Received: December 9, 2021
Peer-review started: December 9, 2021
First decision: January 26, 2022
Revised: February 16, 2022
Accepted: April 28, 2022
Article in press: April 28, 2022
Published online: April 28, 2022
Processing time: 139 Days and 21.8 Hours
Abstract
BACKGROUND

In their everyday life, clinicians face an overabundance of biological indicators potentially helpful during a disease therapy. In this context, to be able to reliably identify a reduced number of those markers showing the ability of optimising the classification of treatment outcomes becomes a factor of vital importance to medical prognosis. In this work, we focus our interest in inflammatory bowel disease (IBD), a long-life threaten with a continuous increasing prevalence worldwide. In particular, IBD can be described as a set of autoimmune conditions affecting the gastrointestinal tract whose two main types are Crohn’s disease and ulcerative colitis.

AIM

To identify the minimal signature of microRNA (miRNA) associated with colorectal cancer (CRC) in patients with one chronic IBD.

METHODS

We provide a framework of well-established statistical and computational learning methods wisely adapted to reconstructing a CRC network leveraged to stratify these patients.

RESULTS

Our strategy resulted in an adjusted signature of 5 miRNAs out of approximately 2600 in Crohn’s Disease (resp. 8 in Ulcerative Colitis) with a percentage of success in patient classification of 82% (resp. 81%).

CONCLUSION

Importantly, these two signatures optimally balance the proportion between the number of significant miRNAs and their percentage of success in patients’ stratification.

Keywords: Inflammatory bowel disease; microRNA; Muti-group comparison; Machine learning; Colorectal cancer; Sparse partial least squares-discriminant analysis

Core Tip: This study provides an optimised strategy based on classic learning methods and multi-group variable selection combination from 2600 microRNAs of 225 patients with one chronic inflammatory bowel disease to identify the minimal signature of microRNAs associated with the development of colorectal cancer in these patients.