top of page
Featured Poster
DDW 2026 Poster
Explainable AI for Transparent EUS-Based Pancreatic Cancer Diagnosis
Our DDW 2026 poster examines how explainable AI can improve confidence in pancreatic cancer detection from endoscopic ultrasound. The study found that EfficientNetV2 outperformed ResNet50 in both classification and lesion-localization fidelity, underscoring the importance of interpretable models that focus on real pathology rather than imaging artifacts.
At the booth
What We're Presenting at DDW
Real-time AI-assisted EUS interpretation
Multimodal integration (EUS + CT/MRI + labs)
Tumor board decision support workflows
PANDA-EUS large-scale video consortium
Multicenter Initiative
PANDA-EUS Multicenter Consortium
We are building a multicenter, de-identified repository of EUS videos and clinical data to power next-generation AI for pancreatic cancer detection and management.
bottom of page