Related Experiment Video
Updated: Jul 8, 2026
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Pediatric pancreas segmentation from MRI scans with deep learning
Elif Keles1, Merve Yazol2, Gorkem Durak1
1Department of Radiology, Northwestern University, IL, USA.
Insights
PanSegNet, a deep learning algorithm, accurately segments pediatric pancreas MRI scans in children with pancreatitis and healthy controls. This validated tool offers expert-level performance, advancing accessible pediatric pancreatic imaging.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Medicine
- Pediatric Radiology
Background:
- Pediatric pancreatic diseases require accurate imaging for diagnosis and management.
- Manual segmentation of pancreas on MRI is time-consuming and subject to inter-observer variability.
- Deep learning offers potential for automated and efficient image analysis.
Purpose of the Study:
- To evaluate and validate PanSegNet, a deep learning algorithm for pediatric pancreas segmentation on MRI.
- To assess PanSegNet's performance in children with acute pancreatitis (AP), chronic pancreatitis (CP), and healthy controls.
Main Methods:
- Retrospective collection of 84 pediatric MRI scans (2-19 years) from healthy children and those with AP/CP.
- Manual pancreas segmentation by pediatric radiologists, confirmed by a senior radiologist.
- Quantitative assessment of PanSegNet segmentations using Dice Similarity Coefficient (DSC) and Hausdorff distance (HD95).
Main Results:
- PanSegNet achieved high DSC scores: 88% (controls), 81% (AP), and 80% (CP).
- HD95 values demonstrated good segmentation accuracy across all groups.
- Strong agreement was observed between automated and manual pancreas volumes (R² = 0.85 controls, 0.77 diseased).
Conclusions:
- PanSegNet is the first validated deep learning tool for pediatric pancreas MRI segmentation.
- The algorithm demonstrates expert-level performance in segmenting pancreata in healthy and diseased pediatric populations.
- The tool and annotated dataset are publicly available to advance research in pediatric pancreatic imaging.
Objective:
Our study aimed to evaluate and validate PanSegNet, a deep learning (DL) algorithm for pediatric pancreas segmentation on MRI in children with acute pancreatitis (AP), chronic pancreatitis (CP), and healthy controls.
Methods:
With IRB approval, we retrospectively collected 84 MRI scans (1.5T/3T Siemens Aera/Verio) from children aged 2-19 years at Gazi University (2015-2024). The dataset includes healthy children as well as patients diagnosed with AP or CP based on clinical criteria. Pediatric and general radiologists manually segmented the pancreas, then confirmed by a senior pediatric radiologist. PanSegNet-generated segmentations were assessed using Dice Similarity Coefficient (DSC) and 95th percentile Hausdorff distance (HD95). Cohen's kappa measured observer agreement.
Results:
Pancreas MRI T2W scans were obtained from 42 children with AP/CP (mean age: 11.73 ± 3.9 years) and 42 healthy children (mean age: 11.19 ± 4.88 years). PanSegNet achieved DSC scores of 88 % (controls), 81 % (AP), and 80 % (CP), with HD95 values of 3.98 mm (controls), 9.85 mm (AP), and 15.67 mm (CP). Inter-observer kappa was 0.86 (controls), 0.82 (pancreatitis), and intra-observer agreement reached 0.88 and 0.81. Strong agreement was observed between automated and manual volumes (R2 = 0.85 in controls, 0.77 in diseased), demonstrating clinical reliability.
Conclusion:
PanSegNet represents the first validated deep learning solution for pancreatic MRI segmentation, achieving expert-level performance across healthy and diseased states. This tool, algorithm, along with our annotated dataset, are freely available on GitHub and OSF, advancing accessible, radiation-free pediatric pancreatic imaging and fostering collaborative research in this underserved domain.
Related Concept Videos
01:24Magnetic Resonance Imaging
01:27Imaging Studies IV: Magnetic Resonance Imaging