# Predicting ustekinumab treatment response in Crohn’s disease using pre-treatment biopsy images

## TL;DR

This paper introduces a new AI method to predict how Crohn’s disease patients will respond to ustekinumab treatment using biopsy images.

## Contribution

A clustering-enhanced weakly supervised learning framework is proposed to improve treatment response prediction from histopathological images.

## Key findings

- The model achieved an AUC of 0.938 in predicting treatment response from whole-slide images.
- The method outperformed baseline patch-level models with high sensitivity and specificity.
- The framework provides interpretable predictions using Grad-CAM and multi-instance learning.

## Abstract

Crohn’s disease (CD) exhibits substantial variability in response to biological therapies such as ustekinumab (UST), a monoclonal antibody targeting interleukin-12/23. However, predicting individual treatment responses remains difficult due to the lack of reliable histopathological biomarkers and the morphological complexity of tissue. While recent deep learning methods have leveraged whole-slide images (WSIs), most lack effective mechanisms for selecting relevant regions and integrating patch-level evidence into robust patient-level predictions. Therefore, a framework that captures local histological cues and global tissue context is needed to improve prediction performance.

We propose a novel clustering-enhanced weakly supervised learning framework to predict UST treatment response from pre-treatment WSIs of CD patients. First, patches from WSIs were encoded using a pre-trained vision foundation model, and k-means clustering was applied to identify representative morphological patterns. Discriminative patches associated with treatment outcomes were selected via a DenseNet-based classifier, with Grad-CAM used to enhance interpretability. To aggregate patch-level predictions, we adopted a multi-instance learning approach, from which whole-slide features were extracted using both patch likelihood histograms and bag-of-words representations. These features were subsequently used to train a classifier for final response prediction. Experimental results on an independent test set demonstrated that our WSI-level model achieved superior predictive performance with an AUC of 0.938 (95% CI: 0.879–0.996), sensitivity of 0.951, and specificity of 0.825, outperforming baseline patch-level models. These findings suggest that our method enables accurate, interpretable, and scalable prediction of biological therapy response in CD, potentially supporting personalized treatment strategies in clinical settings.

https://github.com/caicai2526/USTAIM.