Hepatic Angiomyolipoma: Clinical Features and Imaging Findings of Quantitative Contrast-Enhanced Ultrasound Perfusion Analysis and Magnetic Resonance Imaging

Objectives: To investigate whether quantitative contrast-enhanced ultrasound (CEUS) analysis and magnetic resonance imaging (MRI) can differentiate hepatic angiomyolipoma (AML) from hepatocellular carcinoma (HCC) and hepatic epithelial angiomyolipoma (EAML) from nonepithelial AML.

Methods: Retrospective analysis of 26 CEUS lesions pathologically confirmed as AML and 60 patients with confirmed HCC. Clinical and ultrasound (US) features of AML versus HCC and EAML versus nonepithelial AML were analyzed. Quantitative parameters, including arrival time, ascending slope, time to peak, peak intensity (PI), fall time (FT), and mean transit time (MTT), were analyzed. Diagnostic performance of MRI and CEUS was compared.

Results: Angiomyolipoma was significantly different from HCC in US findings of solitary tumor (P < .001), well-defined border (P = .025), and hyperechogenicity (P < .001). Tumor types were significantly different in US findings of centripetal filling pattern (P < .001), less hypoenhancement in the CEUS delayed phase (P = .04), high PI (P = .007) and ascending slope (P = .006), and short FT (P = .001) and MTT (P < .05). The sensitivity and specificity of quantitative parameters were 61.5% to 96.2% and 55.0% to 71.7% (area under the curve, 0.642-0.773). An early draining vein and tumor vessels, intratumor fat, absence of pseudocapsules, mild signal intensity on diffusion-weighted imaging, and hypointensity in the MRI delayed phase were helpful for differentiation of AML from HCC (P < .05). The positive predictive values of CEUS and MRI were 46.3% to 100% and 42.9% to 75%, with no significant difference (P > .05). Epithelial AML was more likely to be hypoechoic with a centripetal filling pattern than nonepithelial AML (P < .05). The MTT and FT of EAML were shorter than those of nonepithelial AML (P < .05). The PI of EAML was significantly stronger than that of nonepithelial AML (P < .01). The sensitivity and specificity of quantitative parameters were 68.7% to 100.0% and 60.0% to 100.0% (area under the curve, 0.825-0.941). Less intratumor fat was helpful for differentiation of EAML from nonepithelial AML (P < .05). The positive predictive values of CEUS and MRI were 77.8% to 100% and 85.7%, with a significant difference (P < .05).

Conclusions: Quantitative CEUS findings could be valuable in differentiating EAML from HCC.