Seven Abstracts Accepted at APA, Plus a Keynote on Clinical Translation of Pancreatic MRI AI

Our team will present seven accepted abstracts at APA, spanning pancreatic cyst risk prediction, automated image analysis, and pediatric pancreatitis. Prof. Ulas Bagci will also deliver the keynote, “AI in Pancreatic MRI: Addressing Clinical Translation Barriers.”

The contributions share a central objective: advancing pancreatic imaging AI toward reliable, clinically meaningful use. They examine complementary questions about how AI generalizes across patient populations and imaging cohorts, how quantitative MRI can characterize disease, and whether predicted risk probabilities can be trusted across centers.

Pancreatic cyst risk and trustworthy AI

• Good Discrimination, Untrustworthy Probabilities: Site-Aware Calibration for Multi-Center Pancreatic Cyst (IPMN) Risk AI
  Zhixiang Wang, Eminenur Sen Tasci, Raj Keswani, Ulas Bagci.
• Toward Feature-Guided Risk Stratification of Main-Duct IPMN using MRI-Based Deep Learning
  Eminenur Sen Tasci and colleagues. Abstract #263.

Automated pancreatic quantification across diverse cohorts

• Feasibility of a State-of-the-Art Automated Pancreas Segmentation Algorithm in Heterogeneous Multimodal Imaging Cohorts
  Eminenur Sen Tasci and colleagues. Abstract #267.
• Automated Pancreatic Subregion Segmentation for Region-Specific Quantitative Assessment Across Heterogeneous Cohorts
  Eminenur Sen Tasci and colleagues. Abstract #286.

Pediatric pancreatic imaging

• Automated Pancreas Segmentation in Healthy Children and Pediatric Pancreatitis: A Zero-Shot Evaluation of an Adult-Trained AI Model
  Elif Keles and colleagues. Abstract #253.
• T2 MRI Radiomics Accurately Differentiates Pediatric Pancreatitis from Healthy Controls
  Elif Keles and colleagues. Abstract #230.
• From Healthy Pancreas to Chronic Disease: MRI Volumetry Reveals Pancreatic Atrophy Across the Pediatric Pancreatitis Spectrum
  Elif Keles and colleagues. Abstract #225.

Keynote: From promising algorithms to clinical translation

Prof. Ulas Bagci’s keynote, “AI in Pancreatic MRI: Addressing Clinical Translation Barriers,” will address the challenges of translating AI research into clinical practice—a theme connecting the team’s work on generalizability, quantitative assessment, and trustworthy risk prediction.

Congratulations to Zhixiang Wang, Eminenur Sen Tasci, Elif Keles, Gorkem Durak, and all coauthors and collaborators. These contributions reflect a broad collaborative effort across AI, radiology, gastroenterology, and pancreatic disease research.

We look forward to sharing our research and engaging with the APA community.