Temperature-corrected proton density fat fraction estimation using chemical shift-encoded MRI in phantoms

Area of Science:

• Medical Imaging
• Quantitative MRI
• Biomarker Development

Background:

• Chemical shift-encoded MRI (CSE-MRI) is a standard for quantifying proton density fat fraction (PDFF), a key biomarker for hepatic steatosis.
• Temperature fluctuations are known to introduce bias in PDFF estimations, particularly in phantom studies, limiting accuracy.

Purpose of the Study:

• To develop and evaluate strategies for correcting temperature-induced biases in PDFF estimation using CSE-MRI.
• To assess the effectiveness of these correction strategies across different MRI systems, field strengths, and protocols.

Main Methods:

• Computer simulations were used to model temperature effects on PDFF estimation with magnitude-, complex-, and hybrid-based CSE-MRI methods.
• Temperature-controlled phantom experiments were conducted to validate the proposed correction method.
• The method was further evaluated on multi-center, multi-vendor phantom data.

Main Results:

• Simulations and experiments confirmed that increased temperature deviation from the assumed value elevates PDFF bias.
• The proposed correction method significantly reduced PDFF bias and variability for magnitude-based CSE-MRI across various conditions.
• Complex and hybrid CSE-MRI methods demonstrated minimal bias and variability, both before and after correction.

Conclusions:

• Temperature correction effectively mitigates temperature-related PDFF bias and variability in magnitude-based CSE-MRI.
• The developed method improves PDFF estimation accuracy across diverse MRI vendors, sites, field strengths, and protocols, even without prior temperature data.