Multi-Site Fat-Water Phantom MRI Data for PDFF Validation

Six sites contributed chemical shift-encoded MRI data of a phantom containing 11 vials with fat concentrations from 0% to 100%. The data was acquired using different MRI vendors and two field strengths (1.5T and 3T) between October 2014 and December 2015 to validate proton-density fat-fraction quantification techniques. Diego Hernando from the University of Wisconsin–Madison submitted the accompanying manuscript for publication in January 2016.

Use Cases

• Validate proton-density fat-fraction (PDFF) mapping algorithms based on the provided complex-valued multi echo images.
• Assess the reproducibility of fat quantification across different MRI vendors and field strengths using the multi-site acquisition protocol.
• Benchmark R2* (1/T2*) decay rate estimation methods against the provided 'fwmc_r2star' reconstruction results.
• Study the impact of acquisition parameters like echo polarity and precession direction on fat-water separation.

Strengths

• Data was acquired at six different sites with multiple MRI vendors, allowing for reproducibility analysis.
• The phantom includes 11 vials with precisely known fat concentrations ranging from 0% to 100%.
• Includes data from two field strengths (1.5T and 3T) and two protocols per field strength.
• Provides both raw complex-valued image data and processed reconstruction results (PDFF, R2*, water, and fat maps).

Limitations

• The total number of images, rows, or file size is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred from the description or the MAT-file structure after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

University of Wisconsin–Madison, Diego Hernando.

Collection Method

Acquired via chemical shift-encoded MRI scans of a fat-water phantom at six different sites.

Time Range

October 2014 to December 2015.

Freshness

Data collection concluded in December 2015; accompanying manuscript submitted in January 2016.

Geography

Multi-site (specific sites not named).