Poster
Quantitative and explainable AI-powered approaches to predict ulcerative colitis disease activity from hematoxylin and eosin (H&E)-stained whole slide images (WSI)
Crohn’s and Colitis Congress
Study Background
Microscopic inflammation has been shown to be an important indicator of disease activity in ulcerative colitis (UC)1. However, manual histologic scoring is semi-quantitative and subject to interobserver variation, and AI-based solutions often lack interpretability2.
Here we report two distinct quantitative approaches to predict disease activity scores and histological remission using AI-powered digital pathology. Both the random forest classifier (RFC) and graph neural network (GNN) further provide explainability and biological insight by identifying histological features informing model predictions.
Here we report two distinct quantitative approaches to predict disease activity scores and histological remission using AI-powered digital pathology. Both the random forest classifier (RFC) and graph neural network (GNN) further provide explainability and biological insight by identifying histological features informing model predictions.
1 Gonzalez-Partida, I. et al. Eur J Gastroenterol Hepatol. 2021; 33:e796-e802.
2 Romkens, TH. et al. J Crohns Colitis. 2018; 12:425-431
Authors
Kathleen Sucipto1, Archit Khosla1, Michael G. Drage1, Yilan Wang1, Darren Fahy1, Mary Lin1, Murray B. Resnick1, Michael Montalto1, Andrew Beck1, Ilan Wapinski1, Stephanie Hennek1, Christina K.B. Jayson1, & Fedaa Najdawi1
1 PathAI, Boston, Massachusetts