PANTHER Challenge Report: Cross-Domain Pancreatic Tumor Segmentation in Magnetic Resonance Imaging
Abstract
Accurate delineation of pancreatic tumors on Magnetic Resonance Imaging (MRI) is important for diagnosis, radiotherapy treatment planning, and outcome assessment, but remains challenging due to complex anatomy and subtle tumor appearance. In routine practice, tumor contours on MRI are produced manually, which is time-consuming and subject to inter-observer variability. Radiotherapy on MRI-Linear Accelerator (MRI-Linac) systems further requires fast and consistent Gross Tumor Volume (GTV) contours for online adaptation, yet most public pancreas tumor segmentation benchmarks focus on Computed Tomography (CT). The Pancreatic Tumor Segmentation in Therapeutic and Diagnostic MRI (PANTHER) challenge addresses this gap by benchmarking automatic pancreatic tumor segmentation on MRI. The dataset includes contrast-enhanced T1-weighted diagnostic MRI and T2-weighted MRI-Linac scans with expert pancreas and tumor annotations, organized into two tasks: (1) tumor segmentation on diagnostic MRI and (2) tumor segmentation on MRI-Linac images. Performance was evaluated using overlap metrics, distance-based metrics, and tumor volume error. The challenge attracted 285 registered participants, with 12 and 9 final submissions for Tasks 1 and 2, respectively. On diagnostic MRI, top methods achieved performance close to inter-reader agreement. Multi-reader analysis suggested that models often reproduced the contouring style of the training annotator, highlighting the importance of annotation quality and consensus. In contrast, performance on MRI-Linac images was lower and more heterogeneous, including cases of complete localization failure.
Related
- Multimodal Magnetic Resonance Imaging Biomarkers of Pediatric Traumatic Brain Injury Identified by Ensemble Learning Are Associated with Psychopathology
- Comment on "The utility of magnetic resonance imaging of the spinal cord in paediatric transverse myelitis: A systematic review
- Multi-parametric ex vivo magnetic resonance imaging for myelination in the developing mouse brain
- Evolution of deep learning for structural magnetic resonance imaging analysis in Alzheimer’s disease: from slice-level classification to brain structure modeling
- A swine model of infarct-related reentrant ventricular tachycardia: Electroanatomic, magnetic resonance, and histopathological characterization
BibTeX
@article{tarifa2026panther,
title = {{PANTHER Challenge Report: Cross-Domain Pancreatic Tumor Segmentation in Magnetic Resonance Imaging}},
author = {Amparo Betancourt Tarifa and Marcel Verheij and René Monshouwer and Hanne Heerkens and Faisal Mahmood and Uffe Bernchou and Emilie Helgesen Karlsson and Ömer Faruk Durugöl and Maximilian Rokuss and Yannick Kirchhoff and Cédric Hémon and Valentin Boussot and Jean-Claude Nunes and Jean-Louis Dillenseger and Chenyuan Bian and Libo Zhang and Yue Ning and Chuanyi Huang and Lisheng Wang and Kyriaki Kolpetinou and George Matsopoulos and John Hermans and Erik van der Bijl and Peter Koopmans},
journal = {Medical Image Analysis},
year = {2026},
month = jun,
doi = {10.1016/j.media.2026.104186},
url = {https://hal.science/hal-05671365v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - PANTHER Challenge Report: Cross-Domain Pancreatic Tumor Segmentation in Magnetic Resonance Imaging
AU - Amparo Betancourt Tarifa
AU - Marcel Verheij
AU - René Monshouwer
AU - Hanne Heerkens
AU - Faisal Mahmood
AU - Uffe Bernchou
AU - Emilie Helgesen Karlsson
AU - Ömer Faruk Durugöl
AU - Maximilian Rokuss
AU - Yannick Kirchhoff
AU - Cédric Hémon
AU - Valentin Boussot
AU - Jean-Claude Nunes
AU - Jean-Louis Dillenseger
AU - Chenyuan Bian
AU - Libo Zhang
AU - Yue Ning
AU - Chuanyi Huang
AU - Lisheng Wang
AU - Kyriaki Kolpetinou
AU - George Matsopoulos
AU - John Hermans
AU - Erik van der Bijl
AU - Peter Koopmans
PY - 2026
DA - 2026/06/23
JO - Medical Image Analysis
DO - 10.1016/j.media.2026.104186
UR - https://hal.science/hal-05671365v1
AB - Accurate delineation of pancreatic tumors on Magnetic Resonance Imaging (MRI) is important for diagnosis, radiotherapy treatment planning, and outcome assessment, but remains challenging due to complex anatomy and subtle tumor appearance. In routine practice, tumor contours on MRI are produced manually, which is time-consuming and subject to inter-observer variability. Radiotherapy on MRI-Linear Accelerator (MRI-Linac) systems further requires fast and consistent Gross Tumor Volume (GTV) contours for online adaptation, yet most public pancreas tumor segmentation benchmarks focus on Computed Tomography (CT). The Pancreatic Tumor Segmentation in Therapeutic and Diagnostic MRI (PANTHER) challenge addresses this gap by benchmarking automatic pancreatic tumor segmentation on MRI. The dataset includes contrast-enhanced T1-weighted diagnostic MRI and T2-weighted MRI-Linac scans with expert pancreas and tumor annotations, organized into two tasks: (1) tumor segmentation on diagnostic MRI and (2) tumor segmentation on MRI-Linac images. Performance was evaluated using overlap metrics, distance-based metrics, and tumor volume error. The challenge attracted 285 registered participants, with 12 and 9 final submissions for Tasks 1 and 2, respectively. On diagnostic MRI, top methods achieved performance close to inter-reader agreement. Multi-reader analysis suggested that models often reproduced the contouring style of the training annotator, highlighting the importance of annotation quality and consensus. In contrast, performance on MRI-Linac images was lower and more heterogeneous, including cases of complete localization failure.
SN - 1361-8415
SN - 1361-8423
ER -