# A nomogram for predicting small bowel mucosal healing in pediatric Crohn’s disease

## TL;DR

This paper presents a nomogram to predict small bowel mucosal healing in children with Crohn’s disease, aiding in clinical decision-making.

## Contribution

The novel contribution is a predictive nomogram using clinical biomarkers for small bowel mucosal healing in pediatric Crohn’s disease patients.

## Key findings

- The nomogram achieved a high discrimination power with an AUC of 0.855 for predicting mucosal healing.
- Key predictors included erythrocyte sedimentation rate, albumin levels, and C-reactive protein levels.
- The model was developed using data from 91 pediatric Crohn’s disease patients.

## Abstract

According to the updated Selecting Therapeutic Targets in Inflammatory Bowel Disease (STRIDE-II), mucosal healing (MH) is the long-term therapeutic target for Crohn’s disease (CD). Capsule endoscopy (CE) is effective in evaluating small bowel mucosal inflammation. This research seeks to construct a simple tool for predicting small bowel MH in pediatric CD to aid clinical decision-making.

Data from the medical records of patients with CD who underwent CE at the Guangzhou Women and Children’s Medical Center between November 2017 and July 2022 were retrospectively analyzed. The least absolute shrinkage and selection operator (LASSO) logistic regression algorithm was applied to identify predictive factors for small bowel MH. A nomogram incorporating these factors was constructed to predict the probability of MH in this population.

In total, 143 CE examinations performed in 91 pediatric CD patients (median age, 11 years) were included. Based on the Lewis scores, the CD patients were divided into “MH” (42 cases) and “non-MH” groups (101 cases). LASSO regression analysis identified erythrocyte sedimentation rate, albumin levels, aspartate transaminase levels, C-reactive protein levels, platelet count, and lymphocyte percentage as the most significant predictors; and thus, these factors were incorporated into the predictive nomogram model. The area under the receiver-operating characteristic (ROC) curve of the predictive nomogram model was 0.855 (95% confidence interval, 0.783–0.926), suggesting a high discrimination power.

A nomogram was constructed to predict small bowel MH in pediatric CD patients. This nomogram model can enable accurate and simple attentive observation of small bowel inflammation in CD patients.