Skip to content
MRI Heart

CMRxRecon2025 Data4ProposalReview

CMRxRecon2025 demo dataset for proposal review (Data4ProposalReview)

Fully sampled raw k-space of cardiac and aortic MRI (cine, T1 and T2 mapping, flow, tagging, black blood), a small demo set released for reviewing the proposal of a MICCAI 2025 challenge on foundation models for cardiac MRI reconstruction.

Overview

Data4ProposalReview is a small demonstration set of cardiac MRI raw data published on Synapse by the organizers of CMRxRecon2025, a MICCAI 2025 challenge on foundation models for accelerated cardiac MRI reconstruction across centers, scanners and diseases. The organizers released it so that reviewers of the challenge proposal could inspect the data format. It is not the challenge's training or validation release, which sits in a separate Synapse dataset.

Composition

The directory listing on the dataset page shows six folders, each with five cases under TestSet/FullSample: Aorta (sagittal and transversal acquisitions), BlackBlood, Cine (long-axis, LVOT and, for four of the five cases, short-axis), Flow2d, Mapping (one T1 and one T2 mapping file per case) and Tagging. In total the listing holds 49 MATLAB files. Case numbers repeat across folders, but the page does not say whether equal numbers mean the same person, so the number of people is not known. The six zip archives total about 16.4 GB.

Acquisition

The files hold fully sampled multi-coil k-space. The page refers to the acquisition protocol of the earlier CMRxRecon papers and gives no scanner, field strength or site for these cases. For the full CMRxRecon2025 study the page announces 600 cases from five or more centers, with raw k-space, 24 auto-calibration lines and basic acquisition information.

Annotations

None. The data serve image reconstruction from undersampled k-space, and the fully sampled data act as reference.

Known limitations

The set is tiny and meant only as a preview. The file description section on the page (separate _ks, _calib and _info files) does not match the file names in the directory listing, so the exact file layout should be checked after download. Demographics, diagnoses, vendors and countries are not reported for these cases. The license is named only as "CC-BY" without a version, and access needs a Synapse account.

Cohort

Aggregate numbers from the sources below. Bars are relative to the largest value.

Contrast / sequence

Scans

Groups can overlap

  • Cine MRI 14 29%
  • T1 map 5 10%
  • T2 map 5 10%

Directory of data in Data4ProposalReview dataset page on Synapse (syn60059486), wiki and file listing

License and access

Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer

Access
Free registration

Create an account and accept terms online

Access page

Creative Commons Attribution 4.0 International

Use, share and adapt the data for any purpose, including commercial use, as long as you credit the creators.

The dataset page says the CMRxRecon dataset "follows the CC-BY license" and gives no version. The rule answers of CC BY 3.0 and 4.0 are the same in this index. Downloads need a free Synapse account. The CMRxRecon2025 challenge page asks users to cite the listed CMRx papers and says use is free "after the embargo period (Dec. 2025)".

Original license text Version read: 4.0 Checked 2026-10-07

What you can do

  • Yes
  • Yes
  • Yes
  • Yes
  • Yes

What you can share

  • Yes
  • Yes
  • Yes

What you must do

  • Yes
  • Share alike No
  • No
  • No
  • Manuscript review No
  • Release code No
  • Return results No
  • Delete after use No

Limits

  • No
  • Location limits No

Citation

Wang C, Lyu J, Wang S, et al. CMRxRecon: A publicly available k-space dataset and benchmark to advance deep learning for cardiac MRI. Scientific Data 11, 687 (2024). doi:10.1038/s41597-024-03525-4

Sources

Every number on this page comes from one of these documents. Each chart names the table or page it is taken from. The raw numbers are in stats.csv.