AI-Enabled Histologic Assessment of Bile Duct Hyperplasia in Mice

Time: 12:45 pm
day: Conference Day Two

Details:

• A segmentation model a U-Net architecture and EfficientNet backbone was developed to interrogate hyperplastic biliary epithelium in H&E stained mouse liver sections
• The model revealed high sensitivity (95%) and specificity (97%), indicating robust model accuracy. The model-predicted grades showed strong correlation (R² = 0.86) with pathologists’ scores
• This innovative AI-driven approach not only enhances accuracy and efficiency of bile duct hyperplasia assessment in rodent studies and paves the way for more standardized, objective, and high-throughput histopathological evaluations in preclinical research and toxicology studies

Speakers:

Dinesh Bangari Dinesh Bangari is the Head of Global Discovery Pathology & Multimodal Imaging department within Sanofi’s research platform. Dinesh has over 20 years of research experience Read more