The aims of this study were to identify inflammation factors in hepatoblastoma tissue that correlated with different clinical characteristics, and to explore the probability as predictive biomarkers for diagnosis and prognosis.SELDI-TOF-MS was performed to screen protein peaks that were significantly highly expressed in tumor tissue compared with adjacent liver tissue. After removing proteins larger than 30kDa, the targeted peaks were separated by solid phase extraction and tricine–SDS-PAGE. Protein fragments produced by in-gel digestion were identified by LC–MS/MS. Immunohistochemical assays further confirmed these results. Overall survival curves were graphed by Kaplan–Meier method and multivariate analysis was performed by Cox proportional hazards regression model.Three protein peaks (m/z 12,138, m/z 13,462, and m/z 15,120) that were significantly upregulated in the tumor tissue were identified as macrophage migration inhibitory factor (MIF), chemokine (C-X-C motif) ligand 7 (CXCL7), and interleukin 25 (IL-25). These factors were closely related to clinical stage, lymph node metastasis, vascular invasion and serum AFP level. High expression of each inflammatory marker indicated poor prognosis. Multivariate analysis suggested that MIF, CXCL7, and IL-25 were prognostic factors independent of patient sex, age and tumor histological type.MIF, CXCL7, and IL-25 might be considered as effective inflammation factors for diagnosis and prognosis of hepatoblastoma and as potential novel treatment targets through inhibition of inflammatory function.Prognosis studyLevel I.