Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

Longitudinal posttreatment diffuse glioma MRI dataset (UCSF 2018–2019)

Link

https://doi.org/10.1148/ryai.210243

Indexing

Keywords: posttreatment glioma, longitudinal assessment, nnU-Net, segmentation, UCSF, brain MRI, FLAIR, enhancing tumor, necrotic core, peritumoral edema

Content: MR, NR, OI

RadLex: RID35976, RID4025, RID35806, RID10317, RID39563, RID10782, RID4865

SNOMED: 1157043006, 443936004, 1163375002, 1156974002

Author(s)

Jeffrey D. Rudie

Evan Calabrese

Rachit Saluja

David Weiss

John B. Colby

Soonmee Cha

Christopher P. Hess

Andreas M. Rauschecker

Leo P. Sugrue

Javier E. Villanueva-Meyer

Organization(s)

Department of Radiology and Biomedical Imaging, University of California, San Francisco

Department of Radiology, Hospital of the University of Pennsylvania

License

Text: Dataset license not specified in article

Contact

Jeffrey D. Rudie

Funding

American Society of Neuroradiology Foundation Grant in Artificial Intelligence.

Ethical review

HIPAA-compliant, institutional review board–approved study with waiver of written consent at UCSF.

Comments

Retrospective single-site cohort used to train and test 3D nnU-Net models for posttreatment diffuse glioma segmentation and longitudinal change assessment.

Date

Published: 2022-08-03

References

[1] Rudie JD; Calabrese E; Saluja R; et al.. "Longitudinal Assessment of Posttreatment Diffuse Glioma Tissue Volumes with Three-dimensional Convolutional Neural Networks". Radiology: Artificial Intelligence. 2022-01-01. doi:10.1148/ryai.210243. PMID: 36204543. PMCID: PMC9530762.

[2] Bakas S; Reyes M; Jakab A; et al.. "Identifying the best machine learning algorithms for brain tumor segmentation, progression assessment, and overall survival prediction in the BRATS challenge". arXiv 1811.02629. 2018-01-01. Available from: https://arxiv.org/abs/1811.02629

[3] Menze BH; Jakab A; Bauer S; et al.. "The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)". IEEE Trans Med Imaging. 2015-01-01. PMID: 25494501. PMCID: PMC4833122. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4833122/

Dataset

Motivation

To develop and evaluate neural networks for segmentation and longitudinal assessment of posttreatment diffuse glioma on MR imaging.

Sampling

Consecutive discrete patients undergoing diffuse glioma posttreatment follow-up MRI at UCSF between January 2018 and December 2019; four initially selected patients were excluded due to missing images.

Partitioning scheme

Random split at the patient level: 198 patients (396 images) for training; 100 patients (200 images) for test.

Missing information

Dataset release status, de-identification details, and data file formats are not specified.

Relationships between instances

Each patient has two consecutive posttreatment MRI time points; longitudinal networks used registered and subtraction images between time points.

Noise

Heterogeneity in posttreatment appearance; potential technical differences between time points noted as challenging for segmentation-based longitudinal classification.

External data

The publicly available BraTS 2020 training dataset was used to train an initial network for preliminary segmentations of posttreatment MR images.

Confidentiality

HIPAA compliant; retrospective single-site cohort with IRB approval and waiver of written consent.

Sensitive data

Clinical brain MRI of patients with diffuse glioma; includes posttreatment imaging.