S2534 Diagnostic Accuracy of Artificial Intelligence in the Diagnosis of Liver Fibrosis: A Systematic Review and Meta Analysis

Introduction: Chronic liver disease (CLD) is a global health burden where early detection of liver fibrosis is critical for prognosis and management. Conventional methods like biopsy are invasive, while serum biomarkers often lack precision. Recent studies have demonstrated that machine learning algorithms applied to MRI data can enhance diagnostic accuracy for fibrosis detection. This meta-analysis assesses the performance of AI-based imaging tools in diagnosing liver fibrosis in CLD. Methods: The study follows the PRISMA guidelines and multiple major medical databases were searched by creating a Boolean search string to retrieve and select articles which assessed the accuracy of Machine learning, Deep learning and Neural Network models in diagnosis of Liver Fibrosis. The R Studio package was used to evaluate the diagnostic test accuracy. The Meta, Metadata and Mada packages were utilized to evaluate Pooled Area Under Curve and Pooled Accuracy based on the imaging modality used. The heterogeneity of the studies was assessed using the I2 test. Results: This study included a total of 17 studies where 40073 images from 17459 patients were analyzed by artificial intelligence software. The pooled Area Under the Curve (AUC) Set 1 was 0.98 (I2 = 0.6% ,95% CI [0.96,1.00]) and in Set 2 was 0.90 (I2 = 0.0%, 95% CI [0.84-0.96]). The Pooled accuracy was 87% (I2 = 98.4%, 95% CI [0.80-0.91]); and for respective subgroups: Pooled Accuracy specifically in patients who underwent Ultrasound was 81% (I2 = 86.6%, 95% CI [0.72-0.88]), in patients undergoing CECT was 87% (I2 = 98.5%, 95% CI [0.81-0.92]) and in patients who underwent MRI was 89% (I2 = 97.9%, 95% CI [0.68-0.97]). Conclusion: The high pooled AUC and Accuracy highlights the efficacy of Artificial Intelligence in identifying liver fibrosis proving its role in improvement of detection and management of liver fibrosis. The highest accuracy was obtained through AI analysis of MRI images in comparison with CECT and Ultrasound images. This suggests that Machine Learning Algorithms using MRI images show highest Accuracy in the detection of Liver Fibrosis.