Non‐Contrast <scp>MRI</scp> for Hepatocellular Carcinoma Surveillance and Diagnosis: A Systematic Review and Meta‐Analysis of Diagnostic Accuracy

BACKGROUND: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality worldwide, particularly among patients with chronic liver disease and cirrhosis. Non-contrast magnetic resonance imaging (NC-MRI) has emerged as a potential alternative imaging modality for HCC surveillance and diagnosis, avoiding the risks associated with contrast agents; however, its diagnostic performance has not been fully established. This systematic review and meta-analysis aimed to evaluate the diagnostic accuracy of NC-MRI for the detection and surveillance of HCC. METHODS: A comprehensive systematic search of five electronic databases (PubMed, EMBASE, Scopus, the Cochrane Library, and Web of Science) was conducted. Eligible studies evaluated NC-MRI for HCC diagnosis or surveillance and reported sufficient data to construct diagnostic accuracy measures. Data were extracted on study characteristics, patient populations, NC-MRI protocols, reference standards, and diagnostic outcomes, including sensitivity, specificity, diagnostic odds ratio (DOR), and likelihood ratios. Subgroup analyses were performed according to study design, reference standard, and imaging intent (diagnostic vs. surveillance). RESULTS: Twenty-one studies comprising 7412 participants were included. The pooled sensitivity and specificity of NC-MRI for HCC detection were 84% and 94%, respectively. The pooled DOR was 78 (95% CI: 50-122), with a positive likelihood ratio (LR+) of 13 (95% CI: 9-19) and a negative likelihood ratio (LR-) of 0.17 (95% CI: 0.13-0.22). Subgroup analyses demonstrated higher diagnostic performance in prospective studies and when NC-MRI was applied for surveillance purposes. Studies using histopathology as the reference standard also showed high diagnostic accuracy, with a DOR of 74. CONCLUSION: NC-MRI demonstrates consistently high diagnostic accuracy across diverse study settings and represents a promising contrast-free imaging option in selected high-risk populations. However, its role in population-based surveillance requires cautious interpretation, particularly for early-stage disease and warrants further prospective validation.