AI Boosts Accuracy in Diagnosing Liver Fibrosis in MASH, New Study Finds
Amsterdam, Netherlands – Artificial intelligence (AI) is showing promise in improving the consistency and accuracy of liver fibrosis assessment in patients with metabolic associated steatohepatitis (MASH), according to research published in Journal of Hepatology this month. A study led by Abdurrachim and colleagues demonstrates that a digital pathology platform utilizing AI significantly enhances agreement among pathologists when staging fibrosis, notably in early disease stages.
Currently, variation in assessments – both between diffrent pathologists (inter-assessor variation) and within the same pathologist over time (intra-rating variation) - poses a notable challenge in evaluating liver biopsies for MASH. This inconsistency can hinder clinical trial results, impacting patient selection and accurate measurement of treatment response.
The study involved 120 digitized liver biopsies from two trials, analyzed by four specialized hepatopathologists. researchers employed a platform based on Unstained Second Harmonic Generation/Two-PHOTON Excitation Fluorescence (SHG/TPEF) images and quantitative fibrosis (QFibrosis) values generated by AI. Results showed a significant improvement in the Kappa statistic for inter-assessor agreement in fibrosestaging with AI support, especially for fibrosis stages F0-F2.
Specifically, AI support increased agreement among assessors for identifying study participants eligible for inclusion (Fibrosestadium F2-F3) from 45% to 71%, for exclusion (Stadium F0/F4) from 38% to 55%, and for evaluating treatment response from 49% to 61%. A majority of the pathologists – at least three out of four – found the SHG/TPEF images, continuous QFibrosis values, and QFibrosis stages useful in 83%, 55%, and 38% of cases, respectively.
The research, published online November 2024, suggests that AI-powered digital pathology tools could become a valuable aid for pathologists in the diagnosis and management of MASH, potentially streamlining clinical trials and improving patient care.
Source: Abdurrachim D, Lek S, Lin Ong CZ, et al.Utility of AI digital pathology as an aid for pathologists scoring fibrosis in MASH. J hepatol. 2025; 82: 898-908. https://doi.org/10.1016/j.jhep.2024.11.032
