Comments on imaging and management of hepatic hemangiomas

Abstract

Sir, I read with interest the case report by van der Meijs and colleagues [1] regarding a neonate with an extensive “hepatic hemangioendothelioma.” I would like to congratulate the authors for their relentless attempts to help this patient and for reporting their experience. While one cannot argue in principle with the diagnosis of liver hemangioma and the major systemic and local burden of this extensive lesion, I would disagree with some of the management measures provided to this child. There are a few key points on which I would like to comment. First, the proper name for these benign liver tumors is simply “liver hemangiomas”. This name was adopted by the International Society for the Study of Vascular Anomalies based on the classification of vascular anomalies by Mulliken and Glowacki in 1982 [2]. “Hemangioendothelioma” of the liver typically refers to the adult-onset hepatic epithelioid hemangioendothelioma, a rare vascular mesenchymal neoplasm with borderline malignant potential [3]. Though still in common use in the published literature, using this term to refer to the commonly encountered infantile lesions can be confusing and is better avoided in favor of “liver hemangioma.” Second, a CT-guided biopsy was performed on this child before an MRI study was obtained. Liver hemangiomas, as the authors demonstrated with the Doppler US study, are very vascular and percutaneous biopsies carry a significant risk of bleeding, especially if the infant is coagulopathic. The MR imaging features, shown in Fig. 2, are classic for liver hemangiomas. Had the MRI study been performed earlier, the biopsy could have been entirely avoided. The deterioration of the clinical status which developed following the liver biopsy still could have been related to bleeding. The authors stated that “the possibility of a bleed as a result of the biopsy was excluded,” without providing any further details. Bleeding can occur within the lesion and cause serious distension and consumptive coagulopathy. Regarding the alpha-fetoprotein (AFP) level in this neonate, it is worth mentioning that the AFP reference range for normal, full-term neonates and infants is 105–22,000 μg/l [4]. The AFP level of 42,683 μg/ l reported in this paper might be within the normal range and does not, based on this number alone, justify performing a biopsy. The authors’ description of the lesions as “partially cystic” is one of the commonly seen inaccuracies with US studies for liver hemangiomas. The “involuting” center of these lesions can be quite hypoechoic and frequently mistaken for cystic changes. As demonstrated by the MR image provided, the lesions were completely solid. In an attempt to achieve a rapid decrease in tumor size, the authors started the neonate on high-dose prednisone. Why not initiate steroid treatment as early as the diagnosis was made? In addition, the patient was simultaneously given vincristine. The patient developed severe side effects to this treatment (pulmonary hypertension and severe myocardial hypertrophy). It is unclear if or when these drugs were discontinued. Four days following the initiation of medical treatment the authors repeated the MRI study, which demonstrated no response. The more commonly followed protocol is to examine the patient for evidence of regression (or stabilization) of the hemangioma 2 weeks after the initiation Pediatr Radiol (2009) 39:637–638 DOI 10.1007/s00247-009-1244-3