Cyst-X: AI-Powered Pancreatic Cancer Risk Prediction from Multicenter MRI in Centralized and Federated Learning

2025-07-29

AI Summary

Pancreatic cancer is projected to become the second leading cause of cancer-related death in Western countries by 2030, underscoring the urgent need to improve malignant risk stratification of its precursor lesion—intraductal papillary mucinous neoplasm (IPMN). Current clinical guidelines (e.g., Kyoto Guidelines) and expert radiological interpretation suffer from suboptimal sensitivity, leading to either unnecessary surgical resection or missed malignancies. To address this, we propose Cyst-X: a novel AI framework featuring (1) the first large-scale, multicenter MRI dataset of pancreatic cysts; (2) a dual-mode deep learning model supporting both centralized and federated learning—enabling collaborative model training across institutions without sharing raw patient data; and (3) end-to-end learning from fused T1- and T2-weighted MRI sequences. Our model achieves an AUC of 0.82, significantly outperforming both the Kyoto Guidelines (AUC = 0.75) and expert radiologists. Critically, the learned imaging biomarkers demonstrate biologically interpretable patterns correlated with known histopathological features.

Technology Category

• Humans and AI: Explainable AI (XAI) for Human Understanding
• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Pancreatic cancer is projected to become the second-deadliest malignancy in Western countries by 2030, highlighting the urgent need for better early detection. Intraductal papillary mucinous neoplasms (IPMNs), key precursors to pancreatic cancer, are challenging to assess with current guidelines, often leading to unnecessary surgeries or missed malignancies. We present Cyst-X, an AI framework that predicts IPMN malignancy using multicenter MRI data, leveraging MRI's superior soft tissue contrast over CT. Trained on 723 T1- and 738 T2-weighted scans from 764 patients across seven institutions, our models (AUC=0.82) significantly outperform both Kyoto guidelines (AUC=0.75) and expert radiologists. The AI-derived imaging features align with known clinical markers and offer biologically meaningful insights. We also demonstrate strong performance in a federated learning setting, enabling collaborative training without sharing patient data. To promote privacy-preserving AI development and improve IPMN risk stratification, the Cyst-X dataset is released as the first large-scale, multi-center pancreatic cysts MRI dataset.

Problem

• Predicts IPMN malignancy risk using AI and MRI data
• Improves early detection of pancreatic cancer precursors
• Enables privacy-preserving federated learning for multicenter data

Innovation

• AI framework predicts IPMN malignancy using MRI
• Federated learning enables collaborative training without data sharing
• Large-scale multi-center MRI dataset released for privacy-preserving AI