Inflamm Bowel Dis. 2024 Sep 16:izae204. doi: 10.1093/ibd/izae204. Online ahead of print.

David T Rubin ¹, Olga Kubassova ², Christopher R Weber ³, Shashi Adsul ⁴, Marcelo Freire ⁴, Luc Biedermann ⁵, Viktor H Koelzer ^6 7, Brian Bressler ⁸, Wei Xiong ⁸, Jan H Niess ⁹, Matthias S Matter ⁷, Uri Kopylov ¹⁰, Iris Barshack ¹⁰, Chen Mayer ¹⁰, Fernando Magro ¹¹, Fatima Carneiro ¹², Nitsan Maharshak ¹³, Ariel Greenberg ¹³, Simon Hart ², Jamshid Dehmeshki ^2 14, Laurent Peyrin-Biroulet ¹⁵

Author information

¹Department of Pathology, University of Chicago, Chicago, IL, USA.

²Image Analysis Group, London, UK.

³Inflammatory Bowel Disease Center, University of Chicago Medicine, Chicago, IL, USA.

⁴Takeda, Cambridge, MA, USA.

⁵University Hospital of Zurich, University of Zurich, Zurich, Switzerland.

⁶Department of Pathology and Molecular Pathology, University Hospital and University of Zurich, Zurich, Switzerland.

⁷Institute of Medical Genetics and Pathology, University Hospital Basel, Basel, Switzerland.

⁸Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada.

⁹Department of Biomedicine and University Digestive Healthcare Center, University of Basel, Clarunis, Basel, Switzerland.

¹⁰Chaim Sheba Medical Center Ramat Gan Israel, Ramat Gan, Israel.

¹¹Center for Health Technology and Services Research (CINTESIS@RISE), Faculty of Medicine, University of Porto, Porto, Portugal.

¹²Faculty of Medicine, University of Porto and ULS São João, Porto, Portugal.

¹³Tel Aviv Sourasky Medical Center affiliated with the Faculty of Medical and Health Sciences, Tel Aviv University, Tel Aviv, Israel.

¹⁴Faculty of Engineering, Computing and the Environment, Kingston University London, London, UK.

¹⁵Department of Gastroenterology, INFINY Institute, INSERM NGERE, CHRU Nancy, F-54500 Vandoeuvre-lès-Nancy, France.

Abstract

Background: Ulcerative colitis (UC) is a chronic inflammatory bowel disease characterized by increased stool frequency, rectal bleeding, and urgency. To streamline the quantitative assessment of histopathology using the Nancy Index in UC patients, we developed a novel artificial intelligence (AI) tool based on deep learning and tested it in a proof-of-concept trial. In this study, we report the performance of a modified version of the AI tool.

Methods: Nine sites from 6 countries were included. Patients were aged ≥18 years and had UC. Slides were prepared with hematoxylin and eosin staining. A total of 791 images were divided into 2 groups: 630 for training the tool and 161 for testing vs expert histopathologist assessment. The refined AI histology tool utilized a 4-neural network structure to characterize images into a series of cell and tissue type combinations and locations, and then 1 classifier module assigned a Nancy Index score.

Results: In comparison with the proof-of-concept tool, each feature demonstrated an improvement in accuracy. Confusion matrix analysis demonstrated an 80% correlation between predicted and true labels for Nancy scores of 0 or 4; a 96% correlation for a true score of 0 being predicted as 0 or 1; and a 100% correlation for a true score of 2 being predicted as 2 or 3. The Nancy metric (which evaluated Nancy Index prediction) was 74.9% compared with 72.3% for the proof-of-concept model.

Conclusions: We have developed a modified AI histology tool in UC that correlates highly with histopathologists' assessments and suggests promising potential for its clinical application.