Abstract Metabolic dysfunction-associated steatotic liver disease (MASLD), previously termed nonalcoholic fatty liver disease (NAFLD), is the most prevalent chronic liver disorder globally and is strongly associated with obesity, type 2 diabetes mellitus, and the metabolic syndrome. Accurate and timely assessment of hepatic steatosis and fibrosis is critical for risk stratification and therapeutic monitoring. However, liver biopsy, the current reference standard, is invasive, subject to sampling variability, and unsuitable for repeated assessments of a large population. Non-invasive imaging modalities have emerged as potential alternatives, offering reproducibility and possibility for quantitative evaluation. Ultrasound-based techniques are focused on qualitative and sometimes quantitative assessment of liver fat, while elastography techniques are focused on liver stiffness, which is associated with fibrosis stage. Magnetic resonance techniques offer highly reproducible evaluation of steatosis, metabolic alterations, and fibrosis, with magnetic resonance imaging-proton density fat fraction regarded as the reference standard for fat quantification and magnetic resonance elastography providing the highest accuracy for fibrosis staging. Computed tomography also enables liver fat quantification; however, its use is limited mainly because of exposure to ionizing radiation. Recent artificial intelligence (AI) applications are oriented towards enhancing diagnostic accuracy and automate quantitative assessment while supporting longitudinal monitoring. Despite these advancements, further research is required to validate emerging techniques, standardize acquisition protocols, and ensure widespread clinical implementation. This review summarizes and synthesizes current evidence on non-invasive imaging diagnostic techniques for MASLD, highlighting their strengths, limitations, and potential integration into routine clinical practice in order to reduce dependence on invasive liver biopsy.
Lamprinakos et al. (Mon,) studied this question.