Abstract: Objective The aim of this study was to compare CTP score, end-stage liver disease model (MELD), MELD combined with indocyanine green (ICG) clearance test (MELD-ICG) score, chronic liver dysfunction (CLD) score and modified albumin bilirubin (mALBI) scores for predicting early liver dysfunction in patients with primary liver cancer (PLC) undergoing transhepatic arterial chemoembolization (TACE). Methods A retrospective analysis was performed in 87 PLC patients underwent TACE treatment in our hospital from December 2018 through December 2019, and the preoperative CTP, MELD, MELD-ICG, CLD and ALBI scores were calculated. The area under receiver operating curve (AUC) was applied to predict the incidence of liver dysfunction. Results 27 (31.0%) out of 87 patients with PLC after TACE developed postoperative hepatic insufficiency in our series; the preoperative CTP, MELD, MELD-ICG, CLD and ALBI scores in the 27 patients with hepatic insufficiency were (6.9±1.1), 12.0 (10.0 -14.0), 27.4 (20.2-35.0), (1.5±0.9) and (2.6±1.5), significantly higher than in 60 patients with compensated liver functions; the AUC of CTP, MELD, MELD-ICG, CLD and mALBI were 0.708, 0.787, 0.827, 0.735 and 0.723, with the AUC of MELD-ICG in predicting postoperative hepatic insufficiency highest [0.827 (95% CI: 0.730-0.899), significantly higher than CTP or CLD or mALBI .Conclusion The five scoring systems of CTP, MELD, MELD-ICG, CLD and mALBI are all efficacious tools for predicting hepatic dysfunction after TACE in patients with PLC, and the MELD-ICG score has the best predictive value and needs further investigation.

Key words: Primary liver cancer, Transhepatic arterial chemoembolization, Liver dysfunction, Liver function predicting scoring system