Child pages
  • The Río Hortega University Hospital Glioblastoma dataset: a comprehensive collection of preoperative, early postoperative and recurrence MRI scans (RHUH-GBM)

Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.



Image Modified


This data collection consists of multiparametric MRI scans of 40 adult patients with histopathologically confirmed WHO grade 4 astrocytoma, who underwent surgery at the Río Hortega University Hospital in Valladolid, Spain, between January 2018 and December 2022. The dataset encompasses 600 MRI series, covering three time points: preoperative, early post-operative (less than 72 hours after surgery), and the follow-up scan, at which recurrence is diagnosed. Patients included in the sample underwent gross total resection (GTR) or near total resection (NTR), defined as having no residual tumor enhancement and an extent of resection of more than 95% of the initial enhancing volume, respectively. The modified Response Assessment in Neuro-Oncology criteria (RANO) were used to define tumor progression.

The dataset contains T1-weighted (T1w), T2-weighted (T2w), Fluid Attenuated Inversion Recovery (FLAIR), T1w contrast-enhanced (T1ce) sequences, and diffusion-weighted imaging-derived apparent diffusion coefficient (ADC) maps. It also includes clinical and demographic data, IDH status, treatment information, and volumetric assessment of the extent of the resection. Moreover, the dataset comprises expert-validated segmentations of tumor subregions (e.g., enhancing tumor, necrosis, peritumoral region), generated through computer-aided methods from preoperative, postoperative, and follow-up scans.

This dataset is unique in its inclusion of patients who underwent extensive resection of > 95% of the enhancing tumor. It also stands out from other publicly available datasets by providing early postoperative studies and segmentations, filling the gap in preoperative-focused datasets. By making these data publicly available, the scientific community can analyze recurrence patterns in patients who underwent total or near-total resection and develop new registration and segmentation algorithms focused on post-surgical and follow-up studies.


  • This work was partially funded by a grant awarded by the "Instituto Carlos III, Proyectos I-D-i, Acción Estratégica en Salud 2022" under the project titled "Prediction of tumor recurrence in glioblastomas using magnetic resonance imaging, machine learning, and transcriptomic analysis: A supratotal resection guided by artificial intelligence," reference PI22/01680.