Classification of Paediatric Inflammatory Bowel Disease using Machine Learning

Abstract

Paediatric inflammatory bowel disease (PIBD), comprising Crohn’s disease (CD), ulcerative colitis (UC) and inflammatory bowel disease unclassified (IBDU) is a complex and multifactorial condition with increasing incidence. An accurate diagnosis of PIBD is necessary for a prompt and effective treatment. This study utilises machine learning (ML) to classify disease using endoscopic and histological data for 287 children diagnosed with PIBD. Data were used to develop, train, test and validate a ML model to classify disease subtype. Unsupervised models revealed overlap of CD/UC with broad clustering but no clear subtype delineation, whereas hierarchical clustering identified four novel subgroups characterised by differing colonic involvement. Three supervised ML models were developed utilising endoscopic data only, histological only and combined endoscopic/histological data yielding classification accuracy of 71.0%, 76.9% and 82.7% respectively. The optimal combined model was tested on a statistically independent cohort of 48 PIBD patients from the same clinic, accurately classifying 83.3% of patients. This study employs mathematical modelling of endoscopic and histological data to aid diagnostic accuracy. While unsupervised modelling categorises patients into four subgroups, supervised approaches confirm the need of both endoscopic and histological evidence for an accurate diagnosis. Overall, this paper provides a blueprint for ML use with clinical data.

A framework for effective application of machine learning to microbiome-based classification problems

“…ML models can be used to increase our understanding of the variation in the structure of existing data and in making predictions about new data. Researchers have used ML models to diagnose and understand the ecological basis of diseases such as liver cirrhosis, colorectal cancer, inflammatory bowel diseases, obesity, and type 2 diabetes (2–19). The task of diagnosing an individual relies on a rigorously validated model.…”

“…These problems include a lack of transparency in which methods are used and how these methods are implemented; evaluating models without separate held-out test data; unreported variation between the predictive performance on different folds of cross-validation; and unreported variation between cross-validation and testing performances. Though the microbiome field is making progress to avoid some of these pitfalls including validating their models on independent datasets (8, 19, 20) and introducing ways to better use ML tools (21–24), more work is needed to improve reproducibility further and minimize overestimating for model performance.…”

“…Among microbiome researchers, the lack of justification when selecting a modeling approach has often been due to an implicit assumption that more complex models are better. This has resulted in a trend towards using non-linear models such as random forest and deep neural networks (3, 12, 25–27) over simpler models such as logistic regression or other linear models (19, 23, 28). Although in some cases, complex models may capture important non-linear relationships and therefore yield better predictions, they can also result in black boxes that lack interpretability.…”

Predictive Metagenomic Analysis of Autoimmune Disease Identifies Robust Autoimmunity and Disease Specific Microbial Signatures

“…Therefore, to better understand the role of specific taxa in autoimmunity, we have reprocessed and reanalyzed 29 16S and metagenomic studies focused on the gut microbiome and autoimmunity. To do this, we have taken advantage of several machine learning approaches to provide an alternative to the traditional diversity analysis [15][16][17] . These methods give an advantage of learning functional relationships from the data without a need to define them beforehand.…”

GenePy – a score for estimating gene pathogenicity in individuals using next-generation sequencing data

“…This first group comprised 309 patients diagnosed in childhood with IBD. This cohort (further described in [34]) includes unrelated, Caucasian patients ascertained and recruited through Southampton Children's Hospital who were diagnosed under the age of 18 years according to the modified Porto criteria [35].…”