Optimization of measurements with an ultrasound attenuation coefficient algorithm for quantifying liver fat

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Hepatology

Background:

• Ultrasound attenuation coefficient (AC) measurement methods vary, impacting liver fat assessment accuracy.
• Fixed regions of interest (ROIs) may not be optimal for all patient anatomies.
• Comparing fixed vs. movable ROIs is essential for improving ultrasound-based liver steatosis detection.

Purpose of the Study:

• To evaluate if a movable ROI improves ultrasound AC measurement correlation with MRI proton density fat fraction (MRI-PDFF).
• To assess the impact of ROI positioning on the diagnostic performance of AC measurements for liver steatosis.
• To determine the optimal placement for a movable ROI in ultrasound AC measurements.

Main Methods:

• Retrospective multicenter study involving 750 participants.
• Ultrasound AC measurements using an Arietta 850 system with stored raw data.
• Reprocessing data with a standard fixed ROI (35-75 mm), a shorter fixed ROI (45-75 mm), and movable ROIs (30 mm length) positioned at 20 mm and 25 mm below the liver capsule.
• Correlation analysis with MRI-PDFF using Spearman rho and performance assessment via AUROC.

Main Results:

• A movable ROI positioned 25 mm below the liver capsule (ROI-setting4) showed the highest correlation with MRI-PDFF (0.80).
• ROI-setting4 demonstrated superior performance across all steatosis grades, with AUROC for S>0 of 0.92.
• This optimal movable ROI setting significantly outperformed others in participants with obesity and skin-to-liver distances > 25 mm.

Conclusions:

• A movable ROI significantly enhances both the correlation between ultrasound AC and MRI-PDFF and the diagnostic performance for liver steatosis.
• Optimal performance is achieved with a movable ROI positioned 25 mm below the liver capsule.
• Fixed ROI methods may lead to decreased performance, particularly in patients with greater skin-to-liver distances.