Machine learning identifies novel blood protein predictors of penetrating and stricturing complications in newly diagnosed paediatric Crohn's disease

Biomarkers predicting the effect of anti-TNF treatment in paediatric and adult inflammatory bowel disease

OBJECTIVES

Paediatric and adult inflammatory bowel disease (pIBD, aIBD) patients may lose response to anti-TNF treatment within the first year. Adult-extrapolated weight-based dosing is incorrect in children, due to age-related pharmacokinetic differences. We investigated biomarkers for initial and maintenance of response to infliximab (IFX) or adalimumab (ADA), comparing pIBD and aIBD patients.

METHODS

In this prospective, observational study, pIBD (n = 24) and aIBD (n = 21) patients were included when initiating anti-TNF. Escalation from standard dosing and continued anti-TNF at 12 and 18 months were assessed. Biomarkers included clinical laboratory parameters, faecal calprotectin (FCP) and IFX trough levels (TLs). Plasma proteomics was performed in pIBD.

RESULTS

During our study, treatment escalation (in clinical loss of response) occurred more common in pIBD versus aIBD (p = 0.02). We established that IFX therapy escalation in pIBD patients was not due to low infliximab levels. We identified 9 pro-inflammatory proteins that were elevated in patients losing response.

CONCLUSION

Anti-TNF exposure-response relationship may be different in pIBD versus aIBD. No biomarkers for maintained response were identified, but 9 inflammatory proteins were of interest as potential predictors for loss of response in pIBD.

Pediatric stricturing Crohn's disease

Crohn's disease (CD) is a chronic inflammatory disease of the digestive tract. The incidence of pediatric CD is increasing and is currently 2.5-11.4 per 100000 world-wide. Notably, approximately 25% of children with CD develop stricturing CD (SCD) that requires intervention. Symptomatic stricturing diseases refractory to pharmacological management frequently require non-pharmacological interventions. Non-pharmacological therapeutic strategies include endoscopic balloon dilatation, stricturoplasty, and surgical resection of the strictured segment. However, strictures tend to recur postoperatively regardless of treatment modality. The lifetime risk of surgery in patients with childhood SCD remains at 50%-90%. Thus, new and emerging strategies, advanced diagnostic tools, and minimally invasive approaches are under investigation to improve the outcomes and overall quality of life of pediatric patients with SCD.

Linking gene expression to clinical outcomes in pediatric Crohn's disease using machine learning

Pediatric Crohn's disease (CD) is characterized by a severe disease course with frequent complications. We sought to apply machine learning-based models to predict risk of developing future complications in pediatric CD using ileal and colonic gene expression. Gene expression data was generated from 101 formalin-fixed, paraffin-embedded (FFPE) ileal and colonic biopsies obtained from treatment-naïve CD patients and controls. Clinical outcomes including development of strictures or fistulas and progression to surgery were analyzed using differential expression and modeled using machine learning. Differential expression analysis revealed downregulation of pathways related to inflammation and extra-cellular matrix production in patients with strictures. Machine learning-based models were able to incorporate colonic gene expression and clinical characteristics to predict outcomes with high accuracy. Models showed an area under the receiver operating characteristic curve (AUROC) of 0.84 for strictures, 0.83 for remission, and 0.75 for surgery. Genes with potential prognostic importance for strictures (REG1A, MMP3, and DUOX2) were not identified in single gene differential analysis but were found to have strong contributions to predictive models. Our findings in FFPE tissue support the importance of colonic gene expression and the potential for machine learning-based models in predicting outcomes for pediatric CD.

Prognostic modelling in IBD

In the ideal world prognostication or predicting disease course in any chronic condition would allow the clinician to anticipate disease behaviour, providing crucial information for the patient and data regarding best use of resources. Prognostication also allows an understanding of likely response to treatment and the risk of adverse effects of a treatment leading to withdrawal in any individual patient. Therefore, the ability to predict outcomes from the onset of disease is the key step to developing precision personalised medicine, which is the design of medical care to optimise efficiency or therapeutic benefit based on careful profiling of patients. An important corollary is to prevent unnecessary healthcare costs. This paper outlines currently available predictors of disease outcome in IBD and looks to the future which will involve the use of artificial intelligence to interrogate big data derived from various important 'omes' to tease out a more holistic approach to IBD.

A Novel 8-Predictors Signature to Predict Complicated Disease Course in Pediatric-onset Crohn's Disease: A Population-based Study

BACKGROUND

The identification of patients at high risk of a disabling disease course would be invaluable in guiding initial therapy in Crohn's disease (CD). Our objective was to evaluate a combination of clinical, serological, and genetic factors to predict complicated disease course in pediatric-onset CD.

METHODS

Data for pediatric-onset CD patients, diagnosed before 17 years of age between 1988 and 2004 and followed more than 5 years, were extracted from the population-based EPIMAD registry. The main outcome was defined by the occurrence of complicated behavior (stricturing or penetrating) and/or intestinal resection within the 5 years following diagnosis. Lasso logistic regression models were used to build a predictive model based on clinical data at diagnosis, serological data (ASCA, pANCA, anti-OmpC, anti-Cbir1, anti-Fla2, anti-Flax), and 369 candidate single nucleotide polymorphisms.

RESULTS

In total, 156 children with an inflammatory (B1) disease at diagnosis were included. Among them, 35% (n = 54) progressed to a complicated behavior or an intestinal resection within the 5 years following diagnosis. The best predictive model (PREDICT-EPIMAD) included the location at diagnosis, pANCA, and 6 single nucleotide polymorphisms. This model showed good discrimination and good calibration, with an area under the curve of 0.80 after correction for optimism bias (sensitivity, 79%, specificity, 74%, positive predictive value, 61%, negative predictive value, 87%). Decision curve analysis confirmed the clinical utility of the model.

CONCLUSIONS

A combination of clinical, serotypic, and genotypic variables can predict disease progression in this population-based pediatric-onset CD cohort. Independent validation is needed before it can be used in clinical practice.

Multi-omics in Crohn's disease: New insights from inside

Crohn's disease (CD) is an inflammatory bowel disease (IBD) with complex clinical manifestations such as chronic diarrhea, weight loss and hematochezia. Despite the increasing incidence worldwide, cure of CD remains extremely difficult. The rapid development of high-throughput sequencing technology with integrated-omics analyses in recent years has provided a new means for exploring the pathogenesis, mining the biomarkers and designing targeted personalized therapeutics of CD. Host genomics and epigenomics unveil heredity-related mechanisms of susceptible individuals, while microbiome and metabolomics map host-microbe interactions in CD patients. Proteomics shows great potential in searching for promising biomarkers. Nonetheless, single omics technology cannot holistically connect the mechanisms with heterogeneity of pathological behavior in CD. The rise of multi-omics analysis integrates genetic/epigenetic profiles with protein/microbial metabolite functionality, providing new hope for comprehensive and in-depth exploration of CD. Herein, we emphasized the different omics features and applications of CD and discussed the current research and limitations of multi-omics in CD. This review will update and deepen our understanding of CD from integration of broad omics spectra and will provide new evidence for targeted individualized therapeutics.

Using machine learning to impact on long-term clinical care: principles, challenges, and practicalities

The rise of machine learning in healthcare has significant implications for paediatrics. Long-term conditions with significant disease heterogeneity comprise large portions of the routine work performed by paediatricians. Improving outcomes through discovery of disease and treatment prediction models, alongside novel subgroup clustering of patients, are some of the areas in which machine learning holds significant promise. While artificial intelligence has percolated into routine use in our day to day lives through advertising algorithms, song or movie selections and sifting of spam emails, the ability of machine learning to utilise highly complex and dimensional data has not yet reached its full potential in healthcare. In this review article, we discuss some of the foundations of machine learning, including some of the basic algorithms. We emphasise the importance of correct utilisation of machine learning, including adequate data preparation and external validation. Using nutrition in preterm infants and paediatric inflammatory bowel disease as examples, we discuss the evidence and potential utility of machine learning in paediatrics. Finally, we review some of the future applications, alongside challenges and ethical considerations related to application of artificial intelligence. IMPACT: Machine learning is a widely used term; however, understanding of the process and application to healthcare is lacking. This article uses clinical examples to explore complex machine learning terms and algorithms. We discuss limitations and potential future applications within paediatrics and neonatal medicine.

Over-the-Counter Breast Cancer Classification Using Machine Learning and Patient Registration Records

This study aims to determine the feasibility of machine learning (ML) and patient registration record to be utilised to develop an over-the-counter (OTC) screening model for breast cancer risk estimation. Data were retrospectively collected from women who came to the Hospital Universiti Sains Malaysia, Malaysia for breast-related problems. Eight ML models were used: k-nearest neighbour (kNN), elastic-net logistic regression, multivariate adaptive regression splines, artificial neural network, partial least square, random forest, support vector machine (SVM), and extreme gradient boosting. Features utilised for the development of the screening models were limited to information in the patient registration form. The final model was evaluated in terms of performance across a mammographic density. Additionally, the feature importance of the final model was assessed using the model agnostic approach. kNN had the highest Youden J index, precision, and PR-AUC, while SVM had the highest F2 score. The kNN model was selected as the final model. The model had a balanced performance in terms of sensitivity, specificity, and PR-AUC across the mammographic density groups. The most important feature was the age at examination. In conclusion, this study showed that ML and patient registration information are feasible to be used as the OTC screening model for breast cancer.

Age-dependent antibody profiles to plasmodium antigens are differentially associated with two artemisinin combination therapy outcomes in high transmission setting

The impact of pre-existing immunity on the efficacy of artemisinin combination therapy is largely unknown. We performed in-depth profiling of serological responses in a therapeutic efficacy study [comparing artesunate-mefloquine (ASMQ) and artemether-lumefantrine (AL)] using a proteomic microarray. Responses to over 200 Plasmodium antigens were significantly associated with ASMQ treatment outcome but not AL. We used machine learning to develop predictive models of treatment outcome based on the immunoprofile data. The models predict treatment outcome for ASMQ with high (72–85%) accuracy, but could not predict treatment outcome for AL. This divergent treatment outcome suggests that humoral immunity may synergize with the longer mefloquine half-life to provide a prophylactic effect at 28–42 days post-treatment, which was further supported by simulated pharmacokinetic profiling. Our computational approach and modeling revealed the synergistic effect of pre-existing immunity in patients with drug combination that has an extended efficacy on providing long term treatment efficacy of ASMQ.

Early Diagnosis, Early Stratification, and Early Intervention to Deliver Precision Medicine in IBD

Despite huge advances in understanding the molecular basis of IBD, clinical management has continued to rely on a "trial and error" approach. In addition, a therapeutic ceiling has emerged whereby even the most effective interventions are only beneficial for approximately 30% of patients. Consequently, several tools have been developed to aid stratification and guide treatment-decisions. We review the potential application for many of these precision medicine approaches, which are now almost within reach. We highlight the importance of early action (and avoiding inaction) to ensure the best outcomes for patients and how combining early action with precision tools will likely ensure the right treatment is delivered at the right time and place for each individual person living with IBD. The lack of clinical impact to date from precision medicine, despite much hype and investment, should be tempered with the knowledge that clinical translation can take a long time, and many promising breakthroughs might be ready for clinical implementation in the near future. We discuss some of the remaining challenges and barriers to overcome for clinical adoption. We also highlight that early recognition, early diagnosis, early stratification, and early intervention go hand in hand with precision medicine tools. It is the combination of these approaches that offer the greatest opportunity to finally deliver on the promise of precision medicine in IBD.

A Systematic Review of Artificial Intelligence and Machine Learning Applications to Inflammatory Bowel Disease, with Practical Guidelines for Interpretation

BACKGROUND

Inflammatory bowel disease (IBD) is a gastrointestinal chronic disease with an unpredictable disease course. Computational methods such as machine learning (ML) have the potential to stratify IBD patients for the provision of individualized care. The use of ML methods for IBD was surveyed, with an additional focus on how the field has changed over time.

METHODS

On May 6, 2021, a systematic review was conducted through a search of MEDLINE and Embase databases, with the search structure ("machine learning" OR "artificial intelligence") AND ("Crohn* Disease" OR "Ulcerative Colitis" OR "Inflammatory Bowel Disease"). Exclusion criteria included studies not written in English, no human patient data, publication before 2001, studies that were not peer reviewed, nonautoimmune disease comorbidity research, and record types that were not primary research.

RESULTS

Seventy-eight (of 409) records met the inclusion criteria. Random forest methods were most prevalent, and there was an increase in neural networks, mainly applied to imaging data sets. The main applications of ML to clinical tasks were diagnosis (18 of 78), disease course (22 of 78), and disease severity (16 of 78). The median sample size was 263. Clinical and microbiome-related data sets were most popular. Five percent of studies used an external data set after training and testing for additional model validation.

DISCUSSION

Availability of longitudinal and deep phenotyping data could lead to better modeling. Machine learning pipelines that consider imbalanced data and that feature selection only on training data will generate more generalizable models. Machine learning models are increasingly being applied to more complex clinical tasks for specific phenotypes, indicating progress towards personalized medicine for IBD.

A flexible approach for variable selection in large-scale healthcare database studies with missing covariate and outcome data

Background

Prior work has shown that combining bootstrap imputation with tree-based machine learning variable selection methods can provide good performances achievable on fully observed data when covariate and outcome data are missing at random (MAR). This approach however is computationally expensive, especially on large-scale datasets.

Methods

We propose an inference-based method, called RR-BART, which leverages the likelihood-based Bayesian machine learning technique, Bayesian additive regression trees, and uses Rubin’s rule to combine the estimates and variances of the variable importance measures on multiply imputed datasets for variable selection in the presence of MAR data. We conduct a representative simulation study to investigate the practical operating characteristics of RR-BART, and compare it with the bootstrap imputation based methods. We further demonstrate the methods via a case study of risk factors for 3-year incidence of metabolic syndrome among middle-aged women using data from the Study of Women’s Health Across the Nation (SWAN).

Results

The simulation study suggests that even in complex conditions of nonlinearity and nonadditivity with a large percentage of missingness, RR-BART can reasonably recover both prediction and variable selection performances, achievable on the fully observed data. RR-BART provides the best performance that the bootstrap imputation based methods can achieve with the optimal selection threshold value. In addition, RR-BART demonstrates a substantially stronger ability of detecting discrete predictors. Furthermore, RR-BART offers substantial computational savings. When implemented on the SWAN data, RR-BART adds to the literature by selecting a set of predictors that had been less commonly identified as risk factors but had substantial biological justifications.

Conclusion

The proposed variable selection method for MAR data, RR-BART, offers both computational efficiency and good operating characteristics and is utilitarian in large-scale healthcare database studies.

Supplementary Information

The online version contains supplementary material available at (10.1186/s12874-022-01608-7).

Biomarkers for the Prediction and Diagnosis of Fibrostenosing Crohn's Disease: A Systematic Review

BACKGROUND AND AIMS

Intestinal strictures are a common complication of Crohn's disease (CD). Biomarkers of intestinal strictures would assist in their prediction, diagnosis, and monitoring. Herein we provide a comprehensive systematic review of studies assessing biomarkers that may predict or diagnose CD-associated strictures.

METHODS

We performed a systematic review of PubMed, EMBASE, ISI Web of Science, Cochrane Library, and Scopus to identify citations pertaining to biomarkers of intestinal fibrosis through July 6, 2020, that used a reference standard of full-thickness histopathology or cross-sectional imaging or endoscopy. Studies were categorized based on the type of biomarker they evaluated (serum, genetic, histopathologic, or fecal).

RESULTS

Thirty-five distinct biomarkers from 3 major groups were identified: serum (20 markers), genetic (9 markers), and histopathology (6 markers). Promising markers include cartilage oligomeric matrix protein, hepatocyte growth factor activator, and lower levels of microRNA-19-3p (area under the curves were 0.805, 0.738, and 0.67, respectively), and multiple anti-flagellin antibodies (A4-Fla2 [odds ratio, 3.41], anti Fla-X [odds ratio, 2.95], and anti-CBir1 [multiple]). Substantial heterogeneity was observed and none of the markers had undergone formal validation. Specific limitations to acceptance of these markers included failure to use a standardized definition of stricturing disease, lack of specificity, and insufficient relevance to the pathogenesis of intestinal strictures or incomplete knowledge regarding their operating properties.

CONCLUSIONS

There is a lack of well-defined studies on biomarkers of intestinal stricture. Development of reliable and accurate biomarkers of stricture is a research priority. Biomarkers can support the clinical management of CD patients and aid in the stratification and monitoring of patients during clinical trials of future antifibrotic drug candidates.

Artificial Intelligence for Disease Assessment in Inflammatory Bowel Disease: How Will it Change Our Practice?

Artificial intelligence (AI) has arrived and it will directly impact how we assess, monitor, and manage inflammatory bowel disease (IBD). Advances in the machine learning methodologies that power AI have produced astounding results for replicating expert judgment and predicting clinical outcomes, particularly in the analysis of imaging. This review will cover general concepts for AI in IBD, with descriptions of common machine learning methods, including decision trees and neural networks. Applications of AI in IBD will cover recent achievements in endoscopic image interpretation and scoring, new capabilities for cross-sectional image analysis, natural language processing for automated understanding of clinical text, and progress in AI-powered clinical decision support tools. In addition to detailing current evidence supporting the capabilities of AI for replicating expert clinical judgment, speculative commentary on how AI may advance concepts of disease activity assessment, care pathways, and pathophysiologic mechanisms of IBD will be addressed.

Machine Learning Modeling from Omics Data as Prospective Tool for Improvement of Inflammatory Bowel Disease Diagnosis and Clinical Classifications

Research of inflammatory bowel disease (IBD) has identified numerous molecular players involved in the disease development. Even so, the understanding of IBD is incomplete, while disease treatment is still far from the precision medicine. Reliable diagnostic and prognostic biomarkers in IBD are limited which may reduce efficient therapeutic outcomes. High-throughput technologies and artificial intelligence emerged as powerful tools in search of unrevealed molecular patterns that could give important insights into IBD pathogenesis and help to address unmet clinical needs. Machine learning, a subtype of artificial intelligence, uses complex mathematical algorithms to learn from existing data in order to predict future outcomes. The scientific community has been increasingly employing machine learning for the prediction of IBD outcomes from comprehensive patient data-clinical records, genomic, transcriptomic, proteomic, metagenomic, and other IBD relevant omics data. This review aims to present fundamental principles behind machine learning modeling and its current application in IBD research with the focus on studies that explored genomic and transcriptomic data. We described different strategies used for dealing with omics data and outlined the best-performing methods. Before being translated into clinical settings, the developed machine learning models should be tested in independent prospective studies as well as randomized controlled trials.