ObjectivesThe aim of the present study was to describe the experience at a single institution in the management of hepatic epithelioid hemangioendothelioma (HEHE).MethodsWe included 51 patients with histologically confirmed HEHE. We performed log-rank (Cox–Mantel) survival analyses using Kaplan–Meier methods to test differences in survival between patients in different groups. Univariate Cox regression analyses and multivariate proportional hazards regression model were carried out to identify independent prognostic factors.ResultsDifferent imaging modalities were used to diagnose HEHE with various presentations. Liver resection (LR), liver transplantation (LT), systemic treatment (ST), and surveillance had been used in our study. A significant difference was noted between the LR group and the surveillance group with respect to mean survival (p = 0.006), as was in the LR group and the ST group (p = 0.036), and in surgical approach (LR and LT) and nonsurgical approach (ST and surveillance) (p = 0.008). The mean survival between the ST group and the surveillance group was not significantly different (p = 0.851). LR (p = 0.010) and surgical approach (p = 0.014) were favorable predictors of outcome, while macrovascular invasion (MaVI) (p = 0.037), lung metastasis (p = 0.040), and surveillance (p = 0.033) were poor prognostic factors in univariate analysis. Multivariate analysis showed that LR (p = 0.010) and surgical approach (p = 0.014) were independently associated with good OS, while surveillance (p = 0.033) was independently associated with poor OS. After adjusting for confounding factors, patients in the LR group have much better OS than those in the surveillance group (p = 0.013). However, there was no significant difference in OS between the LR group and ST group (p = 0.254), as was in the ST group and the surveillance group (p = 0.857).ConclusionsThe definitive diagnosis of HEHE was dependent on histopathology, and it was not possible to make a specific diagnosis without biopsy because the radiological findings were similar to those in some hepatic malignancies. ST was not recommended for patients who were not candidates for surgical approaches, and surgical approaches should be warranted regardless of disease stage. The retrospective nature and the small size of the data limited the generalizability of the study, designing a worldwide database that contains all data about patients with HEHE independent of their therapy, which was highly recommended.