Cyst-X: A Multicenter MRI Benchmark for Pancreatic Cyst Risk Stratification and Cross-Institutional AI
Research paper by Hongyi Pan and collaborators. arXiv record dated 2026-11-03. Research area: Computer vision.
Resource information
- Authors
- Hongyi Pan, Gorkem Durak, Halil Ertugrul Aktas, Ziliang Hong, Zhixiang Wang, Elif KelesDeniz Seyithanoglu, Zheyuan Zhang, Alpay Medetalibeyoglu, Andrea Mia Bejar, Yavuz Taktak, Gulbiz Dagoglu Kartal, Mehmet Sukru Erturk, Timurhan Cebeci, Yury Velichko, Lili Zhao, Emil Agarunov, Federica Proietto Salanitri, Concetto Spampinato, Pallavi Tiwari, Ziyue Xu, Sachin Jambawalikar, Ivo G. Schoots, Marco J. Bruno, Chenchan Huang, Candice W. Bolan, Tamas Gonda, Frank H. Miller, Rajesh N. Keswani, Michael B. Wallace, Ulas Bagci
- Research area
- Computer vision
- Application domain
- Healthcare
- AI methods
- Research topics
- eess.IV, cs.CV
- arXiv record date
- 2026-11-03
- Publication
- arXiv
- Upstream last modified
- Nov 4, 2026
- License
- CC BY 4.0
- Last record update
- Nov 4, 2026
Abstract
Intraductal papillary mucinous neoplasms (IPMNs) are pancreatic cancer precursors, and guideline-based risk stratification both over-treats indolent lesions and misses progression. We present Cyst-X, a public benchmark of 1,461 T1-weighted (T1W) and T2-weighted (T2W) MRI volumes from 764 patients at seven centers, with expert pancreas masks, risk labels, splits and reference AI risk-prediction models. PanSegNet segmented the pancreas with Dice of 86.34% (T1W) and 89.26% (T2W) on held-out centers. For the primary task, high-risk versus low-risk IPMN, a 3D DenseNet-121 achieved cross-validated AUCs of 74.59% and 73.45%; complementary broad binary and three-class tasks reached 78.67--81.28% and 73.47--75.76%. Because patient images often cannot be pooled across institutions, we also simulated federated learning across the seven centers. On T2W MRI, FedProx was similar to centralized training in center-averaged external AUC (71.00% versus 71.18%) but not in pooled AUC (62.35% versus 77.49%). Radiologist comparisons showed cohort-dependent sensitivity--specificity tradeoffs. Cyst-X supports reproducible method development and identifies within-IPMN risk discrimination and cross-center score comparability as open challenges.
