Pancreatic tumor segmentation on T2W MRI :PANTHER challenge baseline

About
Interfaces
This algorithm implements all of the following input-output combinations:
Challenge Performance
Date | Challenge | Phase | Rank |
---|---|---|---|
April 16, 2025 | PANTHER | Open Development Phase Task 2 | 1 |
Model Facts
Summary
This is the baseline algorithm for Task 2 of the PANTHER challenge. The algorithm is based on the nnU-Net framework (v2) and consists of a two step approach for pancreatic tumor segmentation on T2-weighted MR-Linac MRIs, inspired by the baseline model of the PANORAMA challenge. First a segmentation of the pancreas is obtained from downsampled images using MRSegmentator. Based on this segmentation a region of interest (ROI) around the pancreas is cropped from the original input T2W MRI. The cropped ROIS are then used to train a nnU-Net algorithhm. The process is shown in the figure below.
Figure: Pipeline of the baseline algorithm for Task 2 of the PANTHER challenge.
Mechanism
This algorithm automatically generates a segmentation mask for primary pancreatic tumors on T2-weighted MR-Linac images.
- Input: T2-weighted, MR-Linac MRIs
- Output: Segmentation mask delineating pancreatic tumor regions
The model was trained using the 50 annotated MR images from the PANTHER Public Training Dataset.
Validation and Performance
Uses and Directions
Purpose:
This algorithm was developed for research purposes only and is designed to automatically segment pancreatic tumors in T2W MR-Linac images.
Intended Use:
- The model is intended exclusively for T2W MR-Linac images.
- It should only be applied to image data that match the imaging characteristics of the training set.
Benefits:
- Automatic segmentation of primary pancreatic tumors.
Target Population:
This algorithm was trained on a cohort of patients with confirmed primary pancreatic tumors, treated with an UNITY MR-Linac system. It is meant to support radiation oncologists delineating pancreatic tumors for daily treatment using T2W MR-Linac images and is not intended for use on other imaging modalities or sequences.
General Use:
Radiation oncologists can leverage this model as an initial delineation tool to support and streamline daily radiotherapy planning.
Before Using This Model:
It is recommended to validate the model retrospectively and prospectively on a cohort that reflects your local patient population to confirm its performance and clinical applicability.
Safety and Efficacy Evaluation:
To be determined.
Warnings
Risks:
Even when used appropriately, reliance on this pancreatic tumor segmentation algorithm may lead to mistreatment due to incorrect predictions and missed tumor regions. Errors in segmentation can compromise radiotherapy planning and treatment, which poses significant risks to patient health.
Inappropriate Settings:
This model was developed exclusively for T2-weighted MR-Linac (UNITY) images and has not been trained on other sequences or on data from different MR-Linac manufacturers. Consequently, applying the algorithm to such images may result in unreliable predictions. It should not be deployed in clinical practice without thorough local validation.
Generalizability:
Since the model was trained on a specific set of MR-Linac images, its performance may diminish when applied to images from different sources or acquired using different protocols. It is critical to validate the model on the intended patient population before it is applied in an external setting.
Common Error Messages
The inference progress for this algorithm is incorrectly captured as "error messages", resulting in each case being predicted with "Succeeded, with warnings". This does not mean actual warnings occurred, although this can still be the case.
Information on this algorithm has been provided by the Algorithm Editors, following the Model Facts labels guidelines from Sendak, M.P., Gao, M., Brajer, N. et al. Presenting machine learning model information to clinical end users with model facts labels. npj Digit. Med. 3, 41 (2020). 10.1038/s41746-020-0253-3