BrainMetShare
Dataset Description
A brain MRI dataset to develop and test improved methods for detection and segmentation of brain metastases. The dataset includes 156 whole brain MRI studies, including high-resolution, multi-modal pre- and post-contrast sequences in patients with at least 1 brain metastasis accompanied by ground-truth segmentations by radiologists.
Canonical URL
https://stanfordaimi.azurewebsites.net/datasets/ae0182f1-d5b6-451a-8177-d1f39f01…
Full Description
About 2% of all patients with a primary neoplasm will be diagnosed with brain metastases at the time of their initial diagnosis. As we are getting better at controlling primary cancers, even more patients eventually present with such lesions. Given that brain metastases are often quite treatable with surgery or stereotactic radiosurgery, accurate segmentation of brain metastases is a common job for radiologists. Having algorithms to help detect and localize brain metastasis could relieve radiologists from this tedious but crucial task. Given the success of recent AI techniques on other segmentation tasks, we have put together this gold-standard, labeled MRI dataset to allow for the development and testing of new techniques in these patients with the hopes of spurring research in this area.
Dataset Details
This is a dataset of 156 pre- and post-contrast whole brain MRI studies in patients with at least 1 cerebral metastasis. Mean patient age was 63±12 years (range: 29–92 years). Primary malignancies included lung (n = 99), breast (n = 33), melanoma (n = 7), genitourinary (n = 7), gastrointestinal (n = 5), and miscellaneous cancers (n = 5). The specific primary malignancies for each case are included in an excel sheet that can be downloaded with the data. 64 (41%) had 1–3 metastases, 47 (30%) had 4–10 metastases, and 45 (29%) had >10 metastases. Lesion sizes varied from 2 mm to over 4 cm and were scattered in every region of the brain parenchyma, i.e., the supratentorial and infratentorial regions, as well as the cortical and subcortical structures. It includes 4 different 3D sequences (T1 spin-echo pre-contrast, T1 spin-echo post-contrast, T1 gradient-echo post (using an IR-prepped FSPGR sequence), T2 FLAIR post) in the axial plane, co-registered to each other, resampled to 256 x 256 pixels. The nominal in-plane resolution is 0.94 mm and the through-plane resolution is 1.0 mm. Standard dose (0.1 mmol/kg) gadolinium contrast agents were used for all cases. All the images have been skull-stripped by using the Brain Extraction Tool (BET) (Smith SM. Fast robust automated brain extraction. Hum Brain Map. 2002;17:143–155). The brain masks were generated from the precontrast T1-weighted 3D CUBE imaging series using the nordicICE software package (NordicNeuroLab, Bergen, Norway) and propagated to the other sequences.
Assignment of Labels
For 105 cases, we include radiologist-drawn segmentations of the metastatic lesions, stored in folder ‘mets_stanford_release_train’. The segmentations were based on the T1 gradient-echo post-contrast images. The remaining 51 cases are unlabeled and stored in ‘mets_stanford_release_test’. There are 5 folders for each subject in the training group – folder ‘0’ contains T1 gradient-echo post images; folder ‘1’ contains T1 spin-echo pre images; folder ‘2’ contains T1 spin-echo post images; folder ‘3’ contains T2 FLAIR post images; folder ‘seg’ contains a binary mask of the segmented metastases (0, 255). There are 4 folders for each subject in the testing group, which are labelled identically, except for the absence of folder ‘seg’.
Additional Information
More detailed information on this dataset and the Stanford group’s initial performance on this data set can be found in Grøvik et al., Deep Learning Enables Automatic Detection and Segmentation of Brain Metastases on Multisequence MRI, JMRI 2019; 51(1):175-182.
We would like to thank the team involved with labeling and preparing the data and for checking it for potential PHI: Darvin Yi, Endre Grovik, Elizabeth Tong, Michael Iv, Daniel Rubin, Greg Zaharchuk, and Ghiam Yamin, and the Division of Neuroimaging at Stanford for supporting this project.
Grøvik et al., Deep Learning Enables Automatic Detection and Segmentation of Brain Metastases on Multisequence MRI, JMRI 2019; 51(1):175-182 also available on ArXiv (https://arxiv.org/abs/1903.07988).
Terms & Conditions
Stanford University School of Medicine Brain Mets Dataset Research Use Agreement
By registering for downloads from Brain Mets Dataset, you are agreeing to this Research Use Agreement, as well as to the Terms of Use of the Stanford University School of Medicine website as posted and updated periodically at http://www.stanford.edu/site/terms/.
1. Permission is granted to view and use Brain Mets Dataset without charge for personal, non-commercial research purposes only. Any commercial use, sale, or other monetization is prohibited.
2. Other than the rights granted herein, the Stanford University School of Medicine (“School of Medicine”) retains all rights, title, and interest in the Brain Mets Dataset.
3. You may make a verbatim copy of the Brain Mets Dataset for personal, non-commercial research use as permitted in this Research Use Agreement. If another user within your organization wishes to use the Brain Mets Dataset, they must register as an individual user and comply with all the terms of this Research Use Agreement.
4. YOU MAY NOT DISTRIBUTE, PUBLISH, OR REPRODUCE A COPY of any portion or all of the Brain Mets Dataset to others without specific prior written permission from the School of Medicine.
5. YOU MAY NOT SHARE THE DOWNLOAD LINK to the Brain Mets dataset to others. If another user within your organization wishes to use the Brain Mets Dataset, they must register as an individual user and comply with all the terms of this Research Use Agreement.
6. You must not modify, reverse engineer, decompile, or create derivative works from the Brain Mets Dataset. You must not remove or alter any copyright or other proprietary notices in the Brain Mets Dataset.
7. The Brain Mets Dataset has not been reviewed or approved by the Food and Drug Administration, and is for non-clinical, Research Use Only. In no event shall data or images generated through the use of the Brain Mets Dataset be used or relied upon in the diagnosis or provision of patient care.
8. THE Brain Mets DATASET IS PROVIDED "AS IS," AND STANFORD UNIVERSITY AND ITS COLLABORATORS DO NOT MAKE ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, NOR DO THEY ASSUME ANY LIABILITY OR RESPONSIBILITY FOR THE USE OF THIS Brain Mets DATASET.
9. You will not make any attempt to re-identify any of the individual data subjects. Re-identification of individuals is strictly prohibited. Any re-identification of any individual data subject shall be immediately reported to the School of Medicine.
10. Any violation of this Research Use Agreement or other impermissible use shall be grounds for immediate termination of use of this Brain Mets Dataset. In the event that the School of Medicine determines that the recipient has violated this Research Use Agreement or other impermissible use has been made, the School of Medicine may direct that the undersigned data recipient immediately return all copies of the Brain Mets Dataset and retain no copies thereof even if you did not cause the violation or impermissible use.
In consideration for your agreement to the terms and conditions contained here, Stanford grants you permission to view and use the Brain Mets Dataset for personal, non-commercial research. You may not otherwise copy, reproduce, retransmit, distribute, publish, commercially exploit or otherwise transfer any material.
Limitation of Use
You may use the Brain Mets Dataset for legal purposes only.
You agree to indemnify and hold Stanford harmless from any claims, losses or damages, including legal fees, arising out of or resulting from your use of the Brain Mets Dataset or your violation or role in violation of these Terms. You agree to fully cooperate in Stanford’s defense against any such claims. These Terms shall be governed by and interpreted in accordance with the laws of California.