Exclusive enteral nutrition in adult Crohn’s disease, machine learning in capsule endoscopy and dosing of Vedolizumab based on T cell function

In this month’s blog we focus on two recently published articles in BMJOG and discuss a highly impactful basic science manuscript published in Gut which is likely to have implications of dosing of Vedolizumab. As we move forward from COVID-19 and face a potentially difficult winter it remains vital to practice evidence-based speciality medicine. BMJOG, alongside it’s sister journals continue to publish high impact and clinically relevant guidelines, articles and opinion pieces. Read them for free at the BMJOG website.

In our first article, Mitrev et al report a narrative review on the use, efficacy and clinical implications of exclusive enteral nutrition (EEN) in adult Crohn’s disease. EEN is an established induction regimen for paediatric onset Crohn’s disease, demonstrating superior efficacy to corticosteroids in mild-moderate disease, whilst providing nutritional replenishment for those patients with malnutrition. In this study the authors found 79 articles with reference to EEN use in adult disease, with substantial evidence of improvement in clinical, biomarker, endoscopic and radiologic measures of disease activity. There was additional evidence suggesting efficacy in reducing post-operative complications for those undergoing surgery. Whilst the authors point to good evidence of EEN utility they acknowledge a significant drawback related to compliance, with multiple studies detailing lack of adherence to EEN regimens. The authors conclude that large scale studies are needed to confirm these beneficial effects and ability of patients to comply with EEN treatment. It may be that selected patients will benefit from EEN, with additional individuals benefiting from the novel Crohn’s disease diets (CD-TREAT, CDED) that are under development and assessment.

Secondly, Mascarenhas Saraiva et al report on a hugely interesting topic with fantastic future potential, the use of machine learning for endoscopy interpretation. In this study the authors focused on capsule endoscopy (CE) for evaluation of small bowel disease, specifically concerning the bleeding potential of lesions in the small bowel. They employed a convolutional neural network, a type of supervised machine learning algorithm, which the authors trained and tested on CE images containing a number of lesions (lymphangiectasia, xanthomas, ulcers, erosions, vascular lesions, protruding lesions, and blood). In total over 53,000 CE images were used, with the authors reporting a PPV of 87% and NPV of 99% for classification of the bleeding potential of lesions. Future clinical application of these data is of great interest to reduce time consuming analysis of CE images, however replication in an independent cohort with distinct images is vital.

Finally, we take a quick look at an article published in Gut, detailing the molecular reasoning behind reports of higher doses of Vedolizumab leading to poorer response (GEMINI II + III trials). Becker et al utilised 500 samples from 300 patients and studied Vedolizumab binding to T-cells. Vedolizumab is an anti-α4β7 integrin monoclonal antibody, with the primary mechanism of action thought to be through blockade of gut-homing of T-cells and B-cells in chronic inflammatory states. The authors report single-cell sequencing data on populations of T-cells resistant to Vedolizumab, identifying a β1+PI16+ TReg cell subset appearing to drive this resistance. At ‘therapeutic’ doses, Vedolizumab inhibits effector T-cells, typically associated with a proinflammatory response, but does not inhibit the TReg cell subset. The resistant regulatory T-cell type was identified as having a powerful anti-inflammatory function. Inhibition of these cells at high concentrations of Vedolizumab results in loss of effect of the drug, which correlated with dosage data from GEMINI II + III trials. These data provide an eloquent demonstration of the potential for a therapeutic window for drug dosing, with sufficient levels required for inhibition of effector T-cells, but low enough doses so as to not effect regulatory cells.

(Visited 57 times, 1 visits today)

Leave a Reply

Your email address will not be published. Required fields are marked *

Name *