Deep Learning Analysis of Histologic Images from Intestinal Specimen Reveals Adipocyte Shrinkage and Mast Cell Infiltration to Predict Postoperative Crohn Disease

Artificial intelligence: A new tool in the pathologist's armamentarium for the diagnosis of IBD

Inflammatory bowel diseases (IBD) are classified into two entities, namely Crohn's disease (CD) and ulcerative colitis (UC), which differ in disease trajectories, genetics, epidemiological, clinical, endoscopic, and histopathological aspects. As no single golden standard modality for diagnosing IBD exists, the differential diagnosis among UC, CD, and non-IBD involves a multidisciplinary approach, considering professional groups that include gastroenterologists, endoscopists, radiologists, and pathologists. In this context, histological examination of endoscopic or surgical specimens plays a fundamental role. Nevertheless, in differentiating IBD from non-IBD colitis, the histopathological evaluation of the morphological lesions is limited by sampling and subjective human judgment, leading to potential diagnostic discrepancies. To overcome these limitations, artificial intelligence (AI) techniques are emerging to enable automated analysis of medical images with advantages in accuracy, precision, and speed of investigation, increasing interest in the histological analysis of gastrointestinal inflammation. This review aims to provide an overview of the most recent knowledge and advances in AI methods, summarizing its applications in the histopathological analysis of endoscopic biopsies from IBD patients, and discussing its strengths and limitations in daily clinical practice.

Artificial intelligence and endo-histo-omics: new dimensions of precision endoscopy and histology in inflammatory bowel disease

Integrating artificial intelligence into inflammatory bowel disease (IBD) has the potential to revolutionise clinical practice and research. Artificial intelligence harnesses advanced algorithms to deliver accurate assessments of IBD endoscopy and histology, offering precise evaluations of disease activity, standardised scoring, and outcome prediction. Furthermore, artificial intelligence offers the potential for a holistic endo-histo-omics approach by interlacing and harmonising endoscopy, histology, and omics data towards precision medicine. The emerging applications of artificial intelligence could pave the way for personalised medicine in IBD, offering patient stratification for the most beneficial therapy with minimal risk. Although artificial intelligence holds promise, challenges remain, including data quality, standardisation, reproducibility, scarcity of randomised controlled trials, clinical implementation, ethical concerns, legal liability, and regulatory issues. The development of standardised guidelines and interdisciplinary collaboration, including policy makers and regulatory agencies, is crucial for addressing these challenges and advancing artificial intelligence in IBD clinical practice and trials.

Histological Image-based Ensemble Model to Identify Myenteric Plexitis and Predict Endoscopic Postoperative Recurrence in Crohn's Disease: A Multicentre, Retrospective Study

BACKGROUND AND AIMS

Myenteric plexitis is correlated with postoperative recurrence of Crohn's disease when relying on traditional statistical methods. However, comprehensive assessment of myenteric plexus remains challenging. This study aimed to develop and validate a deep learning system to predict postoperative recurrence through automatic screening and identification of features of the muscular layer and myenteric plexus.

METHODS

We retrospectively reviewed 205 patients who underwent bowel resection surgery from two hospitals. Patients were divided into a training cohort [n = 108], an internal validation cohort [n = 47], and an external validation cohort [n = 50]. A total of 190 960 patches from 278 whole-slide images of surgical specimens were analysed using the ResNet50 encoder, and 6144 features were extracted after transfer learning. We used five robust algorithms to construct classification models. The performances of the models were evaluated based on the area under the receiver operating characteristic curve [AUC] in three cohorts.

RESULTS

The stacking model achieved satisfactory accuracy in predicting postoperative recurrence of CD in the training cohort (AUC: 0.980; 95% confidence interval [CI] 0.960-0.999), internal validation cohort [AUC: 0.908; 95% CI 0.823-0.992], and external validation cohort [AUC: 0.868; 95% CI 0.761-0.975]. The accuracy for identifying the severity of myenteric plexitis was 0.833, 0.745, and 0.694 in the training, internal validation and external validation cohorts, respectively.

CONCLUSIONS

Our work initially established an interpretable stacking model based on features of the muscular layer and myenteric plexus extracted from histological images to identify the severity of myenteric plexitis and predict postoperative recurrence of CD.

Challenges in IBD Research 2024: Precision Medicine

Precision medicine is part of 5 focus areas of the Challenges in IBD Research 2024 research document, which also includes preclinical human IBD mechanisms, environmental triggers, novel technologies, and pragmatic clinical research. Building on Challenges in IBD Research 2019, the current Challenges aims to provide a comprehensive overview of current gaps in inflammatory bowel diseases (IBDs) research and deliver actionable approaches to address them with a focus on how these gaps can lead to advancements in interception, remission, and restoration for these diseases. The document is the result of multidisciplinary input from scientists, clinicians, patients, and funders, and represents a valuable resource for patient-centric research prioritization. In particular, the precision medicine section is focused on the main research gaps in elucidating how to bring the best care to the individual patient in IBD. Research gaps were identified in biomarker discovery and validation for predicting disease progression and choosing the most appropriate treatment for each patient. Other gaps were identified in making the best use of existing patient biosamples and clinical data, developing new technologies to analyze large datasets, and overcoming regulatory and payer hurdles to enable clinical use of biomarkers. To address these gaps, the Workgroup suggests focusing on thoroughly validating existing candidate biomarkers, using best-in-class data generation and analysis tools, and establishing cross-disciplinary teams to tackle regulatory hurdles as early as possible. Altogether, the precision medicine group recognizes the importance of bringing basic scientific biomarker discovery and translating it into the clinic to help improve the lives of IBD patients.

Cross-scale multi-instance learning for pathological image diagnosis

Analyzing high resolution whole slide images (WSIs) with regard to information across multiple scales poses a significant challenge in digital pathology. Multi-instance learning (MIL) is a common solution for working with high resolution images by classifying bags of objects (i.e. sets of smaller image patches). However, such processing is typically performed at a single scale (e.g., 20× magnification) of WSIs, disregarding the vital inter-scale information that is key to diagnoses by human pathologists. In this study, we propose a novel cross-scale MIL algorithm to explicitly aggregate inter-scale relationships into a single MIL network for pathological image diagnosis. The contribution of this paper is three-fold: (1) A novel cross-scale MIL (CS-MIL) algorithm that integrates the multi-scale information and the inter-scale relationships is proposed; (2) A toy dataset with scale-specific morphological features is created and released to examine and visualize differential cross-scale attention; (3) Superior performance on both in-house and public datasets is demonstrated by our simple cross-scale MIL strategy. The official implementation is publicly available at https://github.com/hrlblab/CS-MIL.

Clinical Update on the Prevention and Management of Postoperative Crohn's Disease Recurrence

PURPOSE OF REVIEW

Despite advances in therapeutics, a significant portion of patients with Crohn's disease still require surgical management. In this article, we present updates to the natural history, prognostication and postoperative monitoring, and novel therapeutics in the prevention and treatment of postoperative Crohn's disease recurrence.

RECENT FINDINGS

Clinical risk factors have been associated with higher rates of postoperative recurrence (POR), and in recent studies demonstrate an increased cumulative risk with presence of additional risk factors. Additional novel clinical, histologic, and "-omic" risk factors for recurrence have recently been elucidated, including the role of the mesentery on recurrence and perioperative intraabdominal septic complications. High-risk patients benefit most from medical prophylaxis, including anti-TNF with or without immunomodulator therapy to prevent recurrence. New biologics such as vedolizumab and ustekinumab have emerging evidence in the use of prophylaxis, especially with recent REPREVIO trial data. Non-invasive disease monitoring, such as cross-sectional enterography, intestinal ultrasound, and fecal calprotectin, have been validated against ileocolonoscopy. Recent advances in the prediction, prevention, and monitoring algorithms of postoperative Crohn's disease may be leading to a reduction in postoperative recurrence. Ongoing trials will help determine optimal monitoring and management strategies for this at-risk population.

Machine and deep learning in inflammatory bowel disease

PURPOSE OF REVIEW

The Management of inflammatory bowel disease (IBD) has evolved with the introduction and widespread adoption of biologic agents; however, the advent of artificial intelligence technologies like machine learning and deep learning presents another watershed moment in IBD treatment. Interest in these methods in IBD research has increased over the past 10 years, and they offer a promising path to better clinical outcomes for IBD patients.

RECENT FINDINGS

Developing new tools to evaluate IBD and inform clinical management is challenging because of the expansive volume of data and requisite manual interpretation of data. Recently, machine and deep learning models have been used to streamline diagnosis and evaluation of IBD by automating review of data from several diagnostic modalities with high accuracy. These methods decrease the amount of time that clinicians spend manually reviewing data to formulate an assessment.

SUMMARY

Interest in machine and deep learning is increasing in medicine, and these methods are poised to revolutionize the way that we treat IBD. Here, we highlight the recent advances in using these technologies to evaluate IBD and discuss the ways that they can be leveraged to improve clinical outcomes.

Creeping Fat in Crohn's Disease-Surgical, Histological, and Radiological Approaches

During the course of Crohn's disease, the response of mesenteric adipose tissue to the production of inflammatory mediators and bacterial invasion through the intestinal mucosa results in the formation of creeping fat. Creeping fat describes the arresting finger-like projections that surround the inflamed bowel. In this review, the microscopic and macroscopic features of creeping fat and histological evidence for the importance of this tissue are discussed. Moreover, the most recent insights into the radiological assessment of creeping fat in patients with Crohn's disease are reported. Advances in imaging techniques have revolutionized the possibility of visualization and quantification of adipose tissue depots with excellent accuracy. Visceral fat has been significantly correlated with various Crohn's-disease-related outcomes. Despite the difficulties in distinguishing physiologic perienteric fat from creeping fat, the growing interest in fat-wrapping in Crohn's disease has rejuvenated radiologic research. With regard to the noninvasive fat-wrapping assessment, a novel CT enterography-based mesenteric creeping fat index has been developed for the mitigation of the confounding effect of normal retroperitoneal and perienteric adipose tissue. Research on machine learning algorithms and computational radiomics in conjunction with mechanistic studies may be the key for the elucidation of the complex role of creeping fat in Crohn's disease.

Mise au point de l’utilisation de l’intelligence artificielle dans la prise en charge des maladies inflammatoires chroniques de l’intestin

Complexity of inflammatory bowel diseases (IBD) lies on their management and their biology. Clinics, blood and fecal samples tests, endoscopy and histology are the main tools guiding IBD treatment, but they generate a large amount of data, difficult to analyze by clinicians. Because of its capacity to analyze large number of data, artificial intelligence is currently generating enthusiasm in medicine, and this technology could be used to improve IBD management. In this review, after a short summary on IBD management and artificial intelligence, we will report pragmatic examples of artificial intelligence utilisation in IBD. Lastly, we will discuss the limitations of this technology.

Cross-scale Attention Guided Multi-instance Learning for Crohn's Disease Diagnosis with Pathological Images

Multi-instance learning (MIL) is widely used in the computer-aided interpretation of pathological Whole Slide Images (WSIs) to solve the lack of pixel-wise or patch-wise annotations. Often, this approach directly applies "natural image driven" MIL algorithms which overlook the multi-scale (i.e. pyramidal) nature of WSIs. Off-the-shelf MIL algorithms are typically deployed on a single-scale of WSIs (e.g., 20× magnification), while human pathologists usually aggregate the global and local patterns in a multi-scale manner (e.g., by zooming in and out between different magnifications). In this study, we propose a novel cross-scale attention mechanism to explicitly aggregate inter-scale interactions into a single MIL network for Crohn's Disease (CD), which is a form of inflammatory bowel disease. The contribution of this paper is two-fold: (1) a cross-scale attention mechanism is proposed to aggregate features from different resolutions with multi-scale interaction; and (2) differential multi-scale attention visualizations are generated to localize explainable lesion patterns. By training ~250,000 H&E-stained Ascending Colon (AC) patches from 20 CD patient and 30 healthy control samples at different scales, our approach achieved a superior Area under the Curve (AUC) score of 0.8924 compared with baseline models. The official implementation is publicly available at https://github.com/hrlblab/CS-MIL.