In people with liver fibrosis, artificial intelligence systems exhibit the ability to offer automated diagnosis, staging, and risk stratification.

The objective of a systematic review was to offer a detailed summary of existing artificial intelligence (AI) implementations and evaluate the precision of existing AI applications in automatically diagnosing liver fibrosis.

A comprehensive search was conducted across databases including PubMed, Cochrane Library, EMBASE, and WILEY, utilizing predetermined keywords. The screening process involved evaluating articles for their relevance to AI applications in the diagnosis of liver fibrosis.

Exclusion criteria were applied, which included editorials, studies written in languages other than English, pediatric studies, conference presentations, letters to the editor, abstracts, case reports, and animal studies.

Overall, 24 articles that examined the automated imaging-based diagnosis of liver fibrosis were identified. Among these articles, 6 studies focused on liver ultrasound images, 7 studies analyzed computer tomography images, 5 studies examined magnetic resonance images, and six studies analyzed liver biopsies.

The studies included in the systematic review demonstrated that AI-assisted non-invasive techniques achieved comparable accuracy to human experts in detecting and staging liver fibrosis. However, further validation through clinical trials is necessary before implementing these findings into clinical practice.

AI systems were promising in the diagnosis of liver fibrosis, surpassing the limitations of non-invasive diagnostic methods.