Liver fat quantification by dual-echo MR imaging outperforms traditional histopathological analysis

Area of Science:

• Radiology and Imaging Science
• Hepatology and Liver Disease Research
• Medical Diagnostics

Background:

• Accurate quantification of liver fat content (LFC) is crucial for diagnosing and managing liver diseases.
• Histopathologic analysis, the traditional standard, can be subjective and time-consuming.
• Non-invasive imaging techniques like MRI offer a promising alternative for LFC assessment.

Purpose of the Study:

• To evaluate the accuracy of dual-echo (DE) magnetic resonance imaging (MRI) for quantifying liver fat content (LFC).
• To compare DE MRI with and without fat and water separation against histopathologic analysis.
• To assess the performance of DE MRI in both in vitro phantoms and patients undergoing liver surgery.

Main Methods:

• A 1.5-T MRI scanner was used with a 3D DE MRI sequence.
• Liver fat content was estimated using fat fractions from in-phase/out-of-phase images (MRI(IP/OP)) and Dixon-based fat-only/water-only images (MRI(DIxON)).
• In vitro phantoms and 43 liver biopsies from 22 surgical patients were analyzed, with histopathology serving as the reference standard.

Main Results:

• Both MRI(IP/OP) and MRI(DIxON) showed significant correlation with titrated LFC in vitro (r = 0.993, P < .001).
• In vivo, DE MRI methods correlated significantly better with computer-based histology (P < .001) than traditional cell-count fraction.
• MRI(IP/OP) exhibited smaller measurement bias (4.8%) compared to MRI(DIxON) (21.1%) and histopathology.

Conclusions:

• Dual-echo MRI enables accurate quantification of liver fat content in a surgical patient population.
• DE MRI without fat and water separation provides superior accuracy and minimal measurement bias for LFC quantification.
• DE MRI outperforms traditional histopathologic analysis for assessing liver fat.