Abstract: Objective To explore the risk assessment and pre-warning model of injection within Venenum Bufonis (VBI) in the treatment of liver cancer, so as to demonstration research for clinical practice. Methods The randomized controlled trials (RCTs) of VBI in the treatment of liver cancer were collected by searching comprehensively and systematically in journal literature databases such as CNKI, Wanfang Database, SinoMed, PubMed since the establishment of the database until January 31, 2022. By using the multi-criteria decision analysis model, the benefit-risk assessment (BRA) of VBI in the treatment of liver cancer under different kinds of medication conditions was taken. Then based on BRA values, the VBI risk pre-warning level could be predicted by the entropy weight and matter-element extension model. Results 64 RCTs of VBI in the treatment of liver cancer were included. The BRA value of the treatment method without transarterial chemoembolization was 81.55% higher than that of combined use; the BRA value of Cinobufasu injection was 100% higher than the Chansu injection or Delisheng injection; the conventional dose group was better than the off-label dose group; and the difference between 28 days and long course of treatment was 19 [95%CI（9.42, 28.88）]. Based on the BRA results, the course of treatment need to be first-level pre-warning response. Conclusion Multi-criteria decision-making and entropy weight matter-element extension model can be used as risk assessment and pre-warning model for herbs, which provide basis for traditional Chinese medicine pharmacovigilance.

Key words: Huachansu injection, Delisheng injection, Venenum Bufonis injection, liver cancer, risk assessment, risk pre-warning, pharmacovigilance