Multiparametric quantification and visualization of liver fat using ultrasound

Area of Science:

• Medical Imaging
• Hepatology
• Biophysics

Background:

• Traditional B-scan ultrasound has limitations for assessing liver steatosis.
• Clinicians need to integrate multiple ultrasound parameters for accurate assessment.
• Novel multiparametric approaches are needed for simple clinical evaluation of liver fat.

Purpose of the Study:

• To develop an integrated multiparametric ultrasound approach for assessing liver fat.
• To determine the optimal subset of ultrasound parameters for predicting liver fat fraction.
• To enable simple and rapid clinical assessment of hepatic steatosis.

Main Methods:

• Measured 13 ultrasound and shear wave elastography parameters in 30 subjects.
• Used MRI-derived proton density fat fraction (MRI-PDFF) as the reference standard.
• Analyzed all combinations of parameters to find the optimal predictive subset.

Main Results:

• A combination of four key ultrasound parameters achieved a correlation >0.93 with MRI-PDFF.
• The optimal multiparametric model showed an AUC approaching 1.0 for steatosis classification.
• A visual color overlay strategy was developed to display fat distribution.

Conclusions:

• An integrated multiparametric ultrasound approach can accurately estimate liver fat content.
• As few as four parameters are sufficient for reliable steatosis assessment.
• This method offers a rapid, inexpensive, and visually intuitive tool for clinical use.