Epithelioid Hemangioendothelioma, Multiple Focal Nodular Hyperplasias, and Cavernous Hemangiomas of the Liver

To the Editor.—Focal nodular hyperplasia (FNH) is a mass lesion composed of benign hepatocytes. Typically, a prominent branching artery is present within a central fibrous scar. Increased arterial flow is thought to be a possible cause of the hepatocellular hyperplasia.1 Most examples are idiopathic, occurring in a normal or nearly normal liver. Recent reports have described FNH and FNH-like lesions in a variety of different clinical and anatomic settings, including hypoplasia or agenesis of the portal vein,2,3 Budd-Chiari syndrome,4 cirrhosis of various etiologies,5,6 hereditary hemorrhagic telangiectasia,7 multiple FNH syndrome,3 and adjacent to fibrolamellar carcinoma8–13 and hydatid cyst.14 This myriad of associations and the report of Bralet et al,15 which appeared in a recent issue of the Archives, offer some insight into the pathogenesis of FNH. This letter reviews this information, adds a new case, and proposes a revised hypothesis for the pathogenesis of FNH.The patient Bralet et al described was a young woman treated with oral contraceptives who presented with a large epithelioid hemangioendothelioma in the left lobe and several metastatic deposits in the right lobe of her liver. There were 3 lesions of “FNH” in the right lobe and cells of epithelioid hemangioendothelioma in all 3 lesions. Two points should be mentioned.1. It is not clear that the regions of hepatocellular hyperplasia were actually FNH. The authors do not describe enlarged central vessels with a branched hierarchy accompanied by ductular proliferation. The case has similarities to the multiple FNH syndrome,3 in that the nodules are multiple and there is a hemangioma elsewhere in the liver. However, the pathogenesis of the hyperplastic nodules is likely to be different in this case. The close proximity of neoplastic cells and hepatocellular hyperplasia suggests the neoplasm played a role in causing the hyperplasia. In the absence of more details, it might be better to consider these nodules as peritumoral hyperplasia, as described below.2. The authors suggest that local obstruction of vessels (ie, ischemia) could account for the hyperplasia. This is in keeping with previous suggestions that atrophy appears to precede hyperplasia in nodular regenerative hyperplasia.16 However, parenchyma adjacent to primary and metastatic carcinomas usually shows atrophy rather than hyperplasia.17 What conditions are required to generate an exception to the usual atrophic response?A new case presented here offers additional clues (Figure 1). This patient had a solitary fibrolamellar carcinoma accompanied by a 14-mm-diameter FNH immediately adjacent to the carcinoma and 3 remote FNH lesions measuring 2, 2, and 11 mm in diameter. The carcinoma invaded many structures in the region of the largest FNH, including the portal and hepatic veins, lymphatics, and nerves. All 3 of the remote FNH lesions had central branched vessels and ductular proliferation in a fibrous matrix, and all were free of carcinoma cells. All 3 lesions had adjacent hepatic vein obliteration or stenosis, likely postthrombotic in origin. In one 2-mm FNH, an arteriovenous shunt was identified. Thus, local venous obstruction could have played a role in the development of the FNH response, as suggested by Lough et al18 for non–tumor-associated FNH. Organization of the thrombus may lead to formation of arteriovenous shunts that could augment angiogenesis in the supplying artery and its branches.Berman described FNH adjacent to the fibrolamellar type of hepatocellular carcinoma in 3 of 12 cases.8 Since then, single case reports have documented hyperplastic or nodular lesions immediately adjacent to various focal lesions, including fibrolamellar carcinoma,9–13 hemangioma,1,19 and echinococcal cyst.14 In one case, a 5-mm FNH lesion was noted remote from a large fibrolamellar carcinoma.20 These hyperplastic lesions have been illustrated in very few cases. Only 1 lesion has been demonstrated that grossly resembled FNH.14 One liver had a peripheral irregular zone of hyperplasia with ductular proliferation and radiating fibrous septa.11 One liver had a 1- to 2-cm band of peritumoral hepatocytes that resisted acetaminophen toxicity. In this band there was no fibrous septation; a few arteries were unaccompanied by ducts or portal veins.13 In none of these cases was arterial branching described.The presence of remote FNH lesions suggests an additional stimulus operating at a distance, such as systemic elevation of tumor-derived growth factors or sex hormones. The angiogenic substances, vascular endothelial growth factor or basic fibroblast growth factor, have been documented in endothelial neoplasms, including epithelioid hemangioendothelioma, angiosarcoma, and hemangioma.21–24 Transforming growth factor-β has been documented within fibrolamellar carcinomas.25 Focal nodular hyperplasia has a female predominance in most studies. There is evidence that estrogens may induce hepatic angiogenesis in the rat.26 Many patients with FNH have been exposed to oral contraceptives, including the patient with associated echinococcal cyst.14 The coincidence in some patients of multiple FNH with various vascular lesions and neoplasms, including vascular malformations, hemangioma, meningioma, astrocytoma, and vascular dysplasia, adds another dimension to the puzzle.3 Astrocytomas and meningiomas are responsive to female hormones,27,28 possibly through enhanced vascular endothelial growth factor production.29Considering this body of information, it appears that FNH and peritumoral hyperplasia are nonspecific responses to local increased arterial perfusion. This unity is the basis for the hypothesis presented diagrammatically in Figure 2. In brief, increased arterial flow may be generated by anomalous arteries or angiogenesis. Local angiogenesis may be stimulated by local factors (eg, local venous thrombosis, postthrombotic arteriovenous shunts, and tumor production of angiogenic factors) and augmented by systemic factors (eg, oral contraceptives, female gender, and systemic elevation of tumor-associated growth factors). Mild angiogenesis may account for peritumoral hyperplasia. Marked angiogenesis may cause the FNH phenotype with arterial branching, ductular proliferation, and fibrosis. The fibrosis may be secondary to flow-related injury.30

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.