Quantifying fibrosis in inflammatory bowel diseases - contribution of digital pathology

Abstract

Background: Fibrosis is a common characteristic of inflammatory bowel diseases (IBD), especially Crohn's disease (CD). Despite its significant contribution to disease morbidity, current histological methods for evaluating fibrosis in IBD lack standardization and suffer from poor reproducibility. Adding to the complexity, fibrosis varies across the different layers of the bowel wall.

Methods: We applied a quantitative and phenotypic digital image analysis platform (FibroNest™) to whole slide images of IBD resection specimens stained with Masson trichrome to assess fibrosis and generate reproducible continuous scores for fibrosis severity. We also calculated four subscores based on different phenotypes of histological fibrosis to provide insights in the underlying differences of fibrosis in different bowel wall layers and disease states (CD with or without fibrostenosis and ulcerative colitis (UC)).

Results and conclusions: The generated scores showed a clear and significant separation of CD with fibrostenosis from CD without fibrostenosis and UC, while CD without fibrostenosis showed scores intermediate between UC and CD with fibrostenosis. We also devised a specific protocol to assess fibrosis in the mucosal layer, which has very little fibrosis but is accessible to endoscopic biopsies in contrast to the deeper layers. The mucosal protocol detected significant differences in fibrosis characteristics of the mucosa in CD with fibrosis and CD without fibrosis, which could possibly detect patients with underlying fibrostenosis before bowel resection.

Keywords: Crohn’s disease; Digital pathology; Fibrosis; Image analysis.