Overview

Schema Version

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

Name

UCSD postoperative glioblastoma MRI with multishell diffusion (RSI) cohort

Link

https://dx.doi.org/10.1148/ryai.230489

Indexing

Keywords: glioblastoma, multishell diffusion, restriction spectrum imaging, RSI, DSC perfusion, cellularity, infiltrative tumor, nonenhancing tumor, enhancing tumor, segmentation, nnU-Net, overall survival, progression-free survival, postoperative, preoperative

Content: NR, MR, OI, RS

RadLex: RID35976, RID43349, RID39536, RID6282, RID4044

Author(s)

Louis Gagnon

Diviya Gupta

George Mastorakos

Nathan White

Vanessa Goodwill

Carrie R. McDonald

Thomas Beaumont

Christopher Conlin

Tyler M. Seibert

Uyen Nguyen

Jona Hattangadi-Gluth

Santosh Kesari

Jessica D. Schulte

David Piccioni

Kathleen M. Schmainda

Nikdokht Farid

Anders M. Dale

Jeffrey D. Rudie

Organization(s)

University of California San Diego

Cortechs.ai

Medical College of Wisconsin

Pacific Neuroscience Institute and Saint John’s Cancer Institute at Providence Saint John’s Health Center

University of Pennsylvania

University of California San Francisco

License

Text: Data generated or analyzed during the study are available from the corresponding author by request.

Contact

Corresponding author: Louis Gagnon, email as listed in article: ac.lavalu@9.nongag.siuol

Funding

Salaries for two authors were partly funded by NIH Small Business Innovation Research grant R44NS120796 awarded to Cortechs.ai.

Ethical review

IRB-approved retrospective study at UCSD; informed consent waived; data were de-identified and collected according to HIPAA.

Comments

Retrospective, single-institution internal dataset used to develop and validate a deep learning model for segmentation of enhancing and nonenhancing cellular tumor and for outcome prediction in glioblastoma; four external datasets were used for testing/validation.

Date

Published: 2024-08-21

References

[1] Gagnon L, Gupta D, Mastorakos G, et al.. "Deep Learning Segmentation of Infiltrative and Enhancing Cellular Tumor at Pre- and Posttreatment Multishell Diffusion MRI of Glioblastoma". Radiology: Artificial Intelligence. 2024-08-21. doi:10.1148/ryai.230489. PMID: 39166970. PMCID: PMC11427928.

Dataset

Motivation

To train and validate a DL model to segment enhancing and nonenhancing cellular tumor, distinguish recurrent/residual tumor from posttreatment changes, and predict survival in glioblastoma.

Sampling

Patients with glioblastoma imaged January 2010–June 2022 at UCSD with multishell diffusion (RSI); pathology database searched to enrich for cases with pathology-proven progression and treatment-related changes.

Partitioning scheme

Internal UCSD data were split into training (183 timepoints) and validation (60 timepoints). Fivefold analysis performed for internal evaluation. External testing performed on four independent cohorts.

Missing information

Specific file formats, image counts, and per-sequence acquisition parameters are provided in supplemental appendices; not all are detailed in the main text.

Relationships between instances

Multiple timepoints per patient; multifocal disease present in some timepoints; enhancing cellular tumor (ECT) and nonenhancing cellular tumor (NECT) masks derived from total cellular tumor (TCT).

External data

External validation performed on MCW (postoperative), LUMIERE (postoperative), UPenn-GBM (preoperative), and UCSF-PDGM (preoperative) public/externally sourced datasets as described in the article.

Confidentiality

De-identified according to HIPAA.

Re-identification

Data were de-identified; informed consent waived by IRB.

Sensitive data

Clinical imaging data from patients with glioblastoma, de-identified.