Comparison of non-invasive serum markers for discriminating significant liver fibrosis in children with choledochal cyst

Objectives To compare the discriminatory performance of seven serum markers for significant liver fibrosis in children with choledochal cyst (CDC) and to further explore the relationship between gamma-glutamyl transpeptidase (GGT) level and histologically significant liver fibrosis. Methods We retrospectively collected clinical data of children with CDC who underwent radical surgery and intraoperative liver biopsy at Shanxi Provincial Children's Hospital from January 2020 to December 2025. Using METAVIR fibrosis stage as the histological reference standard, we compared seven serum markers: GGT, GGT to platelet ratio index (GPR), aspartate aminotransferase to alanine aminotransferase ratio (AAR), aspartate aminotransferase to platelet ratio index (APRI), Fibrosis-4 index (FIB-4), fibrosis cirrhosis index (FCI), and S-index. We evaluated the discriminatory performance of each marker using receiver operating characteristic (ROC) curves and analyzed the association between GGT and significant liver fibrosis using multivariate logistic regression. Results We included 219 children, of whom 106 (48.4%) had significant liver fibrosis. GGT showed the highest correlation coefficient with METAVIR fibrosis stage (rs = 0.638, P < 0.001) and the highest AUC (0.826, 95% CI: 0.770–0.882), with an optimal cutoff of 123.0 U/L, sensitivity of 89.6%, and specificity of 69.9%. The AUCs for GPR, S-index, FCI, APRI, AAR, and FIB-4 were 0.814, 0.809, 0.802, 0.771, 0.705, and 0.528, respectively. After Holm correction, GGT showed no statistically significant difference in AUC compared with GPR, FCI, S-index, or APRI, and only outperformed AAR and FIB-4. After adjusting for age, sex, maximum cyst diameter, jaundice history, and cholangitis history, each 10 U/L increase in GGT was associated with an OR of 1.041 (95% CI: 1.021–1.062, P < 0.001) for significant liver fibrosis; after excluding children with a history of preoperative cholangitis, the OR was 1.055 (95% CI: 1.023–1.088, P < 0.001). The proportions of significant liver fibrosis in the GGT < 100, 100–300, and >300 U/L groups were 19.0%, 56.3%, and 69.6%, respectively, showing a linear increasing trend ( χ 2 trend = 42.90, P < 0.001). In the exploratory analysis for ≥F3, the AUCs for FCI and GGT were 0.906 and 0.901, respectively. Conclusions In this cohort, GGT achieved the highest AUC but did not significantly outperform GPR, FCI, S-index, or APRI. GGT was independently associated with ≥F2 significant liver fibrosis, but this association may be partly influenced by biliary obstruction and cholangitis. GGT can serve as a serum marker to assist preoperative assessment of significant liver fibrosis, but it cannot replace histopathological evaluation.