Related Experiment Video
Updated: May 3, 2026
08:43
An Orthotopic Bladder Tumor Model and the Evaluation of Intravesical saRNA Treatment
Published on: July 28, 2012
14.7K
LUND-PROBE - LUND Prostate Radiotherapy Open Benchmarking and Evaluation dataset
Viktor Rogowski1,2, Lars E Olsson1,3, Jonas Scherman1
1Radiation Physics, Department of Hematology, Oncology, and Radiation Physics, Skåne University Hospital, Lund, Sweden.
Scientific Data
|April 11, 2025
Summary
A new dataset aids prostate cancer radiotherapy research by providing MRI/CT images and segmentations for 432 patients. It supports automated planning, segmentation, and deep learning model uncertainty analysis.
Area of Science:
- Medical Imaging
- Radiation Oncology
- Machine Learning
Background:
- Radiotherapy for prostate cancer requires accurate segmentation of target volumes and organs at risk (OARs) using CT and/or MRI.
- Manual segmentation, the gold standard, is time-consuming and labor-intensive, hindering machine learning applications.
Purpose of the Study:
- To present a comprehensive, publicly available clinical dataset for prostate cancer radiotherapy research.
- To facilitate advancements in automated treatment planning, segmentation accuracy, and deep learning model uncertainty assessment.
Main Methods:
- A dataset of MRI and synthetic CT (sCT) images, segmentations, and dose distributions for 432 prostate cancer patients was compiled.
- An extended dataset included deep learning (DL) segmentations, uncertainty maps, and expert-adjusted segmentations for 35 patients.
Main Results:
- The dataset provides a valuable resource for training and validating machine learning models in radiotherapy.
- It enables research into automated segmentation, inter-observer variability, and DL model uncertainty quantification.
Conclusions:
- This publicly accessible dataset accelerates research in medical imaging and prostate cancer radiotherapy.
- It supports the development of more efficient and accurate automated radiotherapy treatment planning systems.
