Conventional and artificial intelligence-based computed tomography and magnetic resonance imaging quantitative techniques for non-invasive liver fibrosis staging

•CT and MRI-based methods have potentials to stage liver fibrosis quantitatively and reproducibly.•MRE is the most accurate method currently and recommended by multiple clinical societies.•The CT LSN score, volumetric assessment and AI-based method have high accuracy, but need additional validation.•There are critical questions that must be considered before the development and validation of non-invasive methods. Chronic liver disease (CLD) ultimately develops into liver fibrosis and cirrhosis and is a major public health problem globally. The assessment of liver fibrosis is important for patients with CLD for prognostication, treatment decisions, and surveillance. Liver biopsies are traditionally performed to determine the stage of liver fibrosis. However, the risks of complications and technical limitations restrict their application to screening and sequential monitoring in clinical practice. CT and MRI are essential for evaluating cirrhosis-associated complications in patients with CLD, and several non-invasive methods based on them have been proposed. Artificial intelligence (AI) techniques have also been applied to stage liver fibrosis. This review aimed to explore the values of conventional and AI-based CT and MRI quantitative techniques for non-invasive liver fibrosis staging and summarized their diagnostic performance, advantages, and limitations.