AIMS: The aim of this study was to characterize and delineate the broad histological spectrum of hepatic angiomyolipoma (AML) and to obtain a better understanding of its clinicopathological diagnosis by reviewing a large series of AMLs. METHODS AND RESULTS: According to the proportions of three histological components, AML could be classified into 10 types; 36 of 55 tumours (65%) were classified as myomatous, eight as myoangiomatous, six as mixed (conventional), two as lipomatous, two as myolipomatous, and one as lipomyomatous. The morphology of smooth muscle cells (SMCs) in AML was quite variable, giving a wide variety of growth patterns. Fourteen (25%) of 55 tumours showed severe cellular atypia, and invasive growth of tumour cells was found in 69% of the tumours. However, except for two autopsy cases, all of the patients were in good health at follow-up, without metastases. Immunostaining for HMB-45 was positive in all cases. We found frequent lymphocyte infiltration, and the occasional presence of abnormally large blood vessels around the tumour at the tumour-background liver interface, features which have not been reported previously. CONCLUSIONS: The majority of hepatic AMLs were myomatous in type, showing variable cellular morphology and growth patterns. Cellular atypia and invasive growth were frequent, indicating that hepatic AMLs often show malignancy-like histological features. Although the majority of cases behave as benign tumours, AML should be considered to have uncertain malignant potential, and careful follow-up of patients is recommended. Immunostaining for HMB-45 is specific for AML, and establishes the diagnosis. The occasional presence of abnormally large blood vessels around the tumour may give new insights into the evaluation of findings from imaging.
AIMS: The aim of this study was to characterize and delineate the broad histological spectrum of hepatic angiomyolipoma (AML) and to obtain a better understanding of its clinicopathological diagnosis by reviewing a large series of AMLs. METHODS AND RESULTS: According to the proportions of three histological components, AML could be classified into 10 types; 36 of 55 tumours (65%) were classified as myomatous, eight as myoangiomatous, six as mixed (conventional), two as lipomatous, two as myolipomatous, and one as lipomyomatous. The morphology of smooth muscle cells (SMCs) in AML was quite variable, giving a wide variety of growth patterns. Fourteen (25%) of 55 tumours showed severe cellular atypia, and invasive growth of tumour cells was found in 69% of the tumours. However, except for two autopsy cases, all of the patients were in good health at follow-up, without metastases. Immunostaining for HMB-45 was positive in all cases. We found frequent lymphocyte infiltration, and the occasional presence of abnormally large blood vessels around the tumour at the tumour-background liver interface, features which have not been reported previously. CONCLUSIONS: The majority of hepatic AMLs were myomatous in type, showing variable cellular morphology and growth patterns. Cellular atypia and invasive growth were frequent, indicating that hepatic AMLs often show malignancy-like histological features. Although the majority of cases behave as benign tumours, AML should be considered to have uncertain malignant potential, and careful follow-up of patients is recommended. Immunostaining for HMB-45 is specific for AML, and establishes the diagnosis. The occasional presence of abnormally large blood vessels around the tumour may give new insights into the evaluation of findings from imaging.