Development and validation of non-invasive models in predicting advanced fibrosis of choledochal cyst

Abstract

PurposeCholedochal cyst patients (CDCs) develop liver brosis, especially advanced brosis without prompt surgery. This study validated the aspartate aminotransferase (AST)-to-platelet ratio index (APRI) and Fibrosis-4 index (FIB-4) and constructed a model for predicting advanced brosis in pediatric CDCs. Methods 330 CDCs (advanced brosis: 34, Ludwig staging 3-4; non-advanced brosis: 296, Ludwig staging 0-2) between January 2020 and March 2022 were reviewed. The APRI and FIB-4 were validated. Relevant variables were analyzed by backward stepwise logistic regression. Enhanced bootstrap method was used for internal veri cation with 1000 samples. ResultsMean age at operation was 29.42 ± 35.15 months. The AUROCs of APRI and FIB-4 were 0.761 (0.673-0.850) and 0.561 (p = 0.246). An index, APAR (AST to prealbumin ratio), was constructed with AUROC of 0.776 (0.693-0.860). The AUROCs of APAR + APRI and APAR + FIB-4 were 0.791 (0.713-0.869) and 0.782 (0.699-0.865). No signi cant differences were in the AUROCs of the indices or their combinations. APAR and APRI could be used together to reduce the rate of misdiagnosis. The risk of advanced brosis varied from the different APAR and APRI scores. ConclusionBoth APAR and APRI were crucial to identify patients at high risk of advanced brosis for CDC.