ABSTRACT: Introduction: Epithelioid hepatic angiomyolipoma (Epi-HAML) can easily be misdiagnosed as a malignant tumor such as hepatocellular carcinoma (HCC) because of the low-fat content on imaging. We analyzed and compared the magnetic resonance imaging (MRI) features of Epi-HAML and HCC, which would aid in disease diagnosis. Methods: We included 30 pathologically confirmed patients with Epi-HAML and 80 with HCC, who underwent both MRI unenhanced scan and three-phase contrast-enhanced MRI scan. The clinical and MRI features of the two groups were summarized and analyzed. Results: Epi-HAML showed significant differences compared to HCC group in terms of clinical features such as sex preference, age, concomitant diseases (hepatitis B and cirrhosis), and elevated plasma alpha-fetoprotein (AFP) (P < 0.001). In addition, there were statistically significant differences between both tumor types with regard to conventional MRI findings such as a solitary tumor (100 vs. 83.8%, P = 0.018), well-defined (93.3 vs. 71.3%, P = 0.027), mild hyperintensity (40.0 vs. 3.7%, P < 0.001) on DWI with high b-value, fat within the tumor (43.3 vs. 8.8%, P < 0.001), and rare necrosis (3.3 vs. 26.3%, P = 0.016). Besides, Epi-HAML displayed significant differences compared to HCC in terms of contrast-enhanced MRI characteristics such as draining hepatic vein (30.0 vs. 3.8%, P < 0.001), portal vein tumor thrombus (0 vs. 13.8%, P = 0.033), hypointensity at delayed phase (70.0 vs. 95%, P = 0.001), intra-tumor vessel at delayed phase (36.7 vs. 10.0%, P = 0.003), pseudocapsule (20.0 vs. 78.8%, P < 0.001), and prolonged enhancement (56.7 vs. 1.2%, P < 0.001). Conclusion: Epi-HAML frequently occurs in middle-aged women and usually lacks characteristic clinical symptoms. Typically, Epi-HAML presents as an isolated and well-defined tumor with rich vasculature. Specific MRI features such as intra-tumor fat, intra-tumor vessel, draining hepatic vein, prolonged enhancement, and lack of capsule may contribute to a more confident diagnosis of Epi-HAML.
