BACKGROUND: This meta-analysis aimed to summarize the diagnostic accuracy and cut-off values of cytokeratin (CK) 18 measurements, specifically M30 and M65, as candidate biomarkers for the pathological evaluation of biopsy specimens used to stage liver inflammation and fibrosis in patients with chronic liver diseases. METHODS: Databases were searched for studies collected up to January 11th, 2025. Pooled sensitivity, specificity, area under the receiver-operating characteristic curves, and mean cut-off values were calculated using random-effects models regardless of heterogeneity. A meta-regression analysis and subgroup analysis were performed to explore heterogeneity. RESULTS: Sixty-three studies comprising 9137 patients were included. The summarized AUROC curve of CK18 M30 for the diagnosis of significant liver inflammation, fibrosis ≥F1, ≥F2, ≥F3, and =F4 according to the METAVIR score system were 0.82, 0.75, 0.78, 0.78 and 0.76, with mean cut-off values of 264.3, 188.0, 276.9, 322.8 and 169.4 U/L. For M65, the summarized AUROC curve for detecting significant liver inflammation, fibrosis ≥F1, ≥F2, and =F4 were 0.79, 0.70, 0.76, 0.64 and 0.72, with mean cut-off values of 541.1, 417.6, 500.1, 424.6 and 674.0 U/L. The subgroup analyses implied that ethnicity may be the primary factor related to heterogeneity in CK18 M30 when applied to detect significant inflammation. Asian patients had values 79.7 U/L higher than those of non-Asian patients (p = 0.0157). CONCLUSIONS: CK18 M30 and M65 have clinically meaningful accuracy as alternative diagnostic tools for determining liver inflammation and fibrosis using biopsy specimens of patients with steatotic liver disease or viral hepatitis. REGISTRATION: PROSPERO registration number: CRD42022364598.