REPLY

We thank Dr. Gomori for his comments regarding our case report of a patient with an intraventricular melanoma. We agree that imaging of intracranial melanomas is complex. Melanotic melanomas have a typical appearance (bright on T1- and dark on T2-weighted images), whereas amelanotic melanomas have fewer typical features (mildly hypointense on T1- and mildly hyperintense on T2-weighted images). These features may be complicated by the presence of hemorrhage that happens in both melanotic and amelanotic melanomas. Once a hemorrhage has occurred, its MR imaging features are less specific, and the differential diagnosis expands. At 1.5 T, intracellular methemoglobin may have similar characteristics to melanotic melanoma. In the article by Atlas et al, the incidence of hemorrhage in metastatic brain melanomas was not given (1). To the best of our knowledge, the incidence of hemorrhage in these tumors is not well established. According to Enochs et al, the T1 shortening in melanotic melanomas is owing to binding of metallic ions such as Fe3⁺, Mn2⁺, and Cu2⁺ to melanin (2). In addition, aggregation of melanin in macromolecular particles contributes to this feature.

Although no intratumoral hemorrhage was found in our case (except that presumed to be due to surgical manipulation), we agree with Dr. Gomori that there is a suspicion for hemorrhage in the left occipital horn. The presence of microhemorrhages, not visible as discrete lesions on CT or MR imaging, may contribute to the appearance of some melanomas on imaging studies. In our patient, we believe that the imaging features were from hypercellularity, compactness, and the presence of melanin rather than hemorrhage. We disagree with his comment regarding the lack of support, from a histologic standpoint, in the diagnosis of melanoma. We clearly showed that, using HMB-45, the tumor strongly showed immunoreactivity. This marker is very specific for melanomas and nevus precursor cells (3). Occasionally, this marker may show cross-reactivity with angiomyolipomas and carcinomas. In our patient, we also used immunohistochemical markers for cytokeratin, leukocyte common antigens (marker for lymphoma), and glial fibrillary acidic protein, and all were negative.