Hepatic angiomyolipoma: clinical, imaging and pathological features in 178 cases.

Abstract

The aim of the study was to review the main clinical, imaging and pathological features of hepatic angiomyolipoma (HAML). We retrospectively analyzed the imaging, pathological and clinical features of 178 patients who underwent surgical resection for HAML. Forty-three males and 135 females with a median age of 43.5 years (range: 17-76 years) were enrolled in the study. Routine blood tests, including those for α-fetoprotein, carcinoembryonic antigen, and carbohydrate antigen determinant, were normal. Liver function was classified as Child-Pugh A for all patients. Twenty-one patients were positive for hepatitis B surface antigen (HBsAg), while two had liver cirrhosis. Seventy-five lesions were located in the right hepatic lobe, 42 in the left hepatic lobe, 12 in the hepatic caudal lobe, and 49 in both left and right lobe. All patients underwent examination by ultrasound sonography (US), and additional examinations by computed tomography (CT) and magnetic resonance (MR) were carried out for 115 and 81 patients, respectively. Various diseases including HAML, liver cancer, liver hemangioma, hepatic lipoma, hepatic hamartoma, focal nodular hyperplasia and liver adenoma were diagnosed before operation. CT and MRI are more conclusive than US (US vs CT: χ(2) = 17.72, P = 0.00; US vs MRI: χ(2) = 12.14, P = 0.00), but for the sensitivity of these three imaging modalities, MRI performs the best (US vs MRI: χ(2) = 5.938, P = 0.015; CT vs MRI: χ(2) = 2.544, P = 0.111). All the patients underwent liver resection (96 right hepatic lobectomy, 41 left hepatic lobectomy and nine caudal lobe hepatic lobectomy). Several sampled tissues were positive for HBsAg (21/168, 12.5 %), β-catenin (4/41, 9.8 %), CD-34 (119/165, 72.1 %), HMB-45 (162/169, 92.9 %), SMA (145/158, 91.8 %) and vimentin (20/27, 74.1 %). The diagnosis of HAML remains difficult and mainly relies on pathological findings. For the imaging modalities, we recommend MRI examination because of its better definition and sensitivity. Although HMB-45 is considered the main histopathological biomarker for HAML, SMA may be also helpful in the diagnostic process (HMB-45 vs SMA: χ(2) = 2.37, P = 0.123). When the diagnosis of HAML is suspicious or hypothesized, surgical resection remains the recommended strategy due to the difficulty in diagnosis preoperatively and the potentially invasive growth of the lesion.