The role of magnetic resonance imaging in children with intraocular tumors and simulating lesions.

PURPOSE: To evaluate the role of pre- and post-contrast magnetic resonance (MR) studies in children with intraocular tumors and simulating lesions.

METHODS: Patients younger than 12 years of age with the diagnosis of an intraocular tumor or simulating lesion with a thickness greater than 2.0 mm underwent pre- and post-contrast MR studies with surface coil. All post-contrast-enhanced images were performed with fat suppression techniques.

RESULTS: There were 40 children with the clinical diagnosis of an intraocular lesion as follows: retinoblastoma (n = 22), Coats disease (n = 5), ciliary body medulloepithelioma (n = 2), primary hyperplastic persistent vitreous (PHPV) (n = 2), retinal capillary hemangioma (n = 3), massive retinal gliosis (phthisis bulbi) (n = 2), uveal melanoma (n = 1), ciliary body leiomyoma (n = 1), retinopathy of prematurity (ROP) with total retinal detachment (n = 1), and post-traumatic retinal gliosis (n = 1). In the authors' series, solid intraocular tumors greater than 2.0 mm in thickness such as retinoblastoma, ciliary body medulloepithelioma, leiomyoma, choroidal melanoma, and retinal capillary hemangioma appeared hyperintense on T1-weighted images and hypointense on T2-weighted images. Secondary serous or exudative retinal detachment, regardless of the underlying etiology (neoplasm, Coats disease, PHPV, phthisis bulbi, ROP), showed hyperintensity of the subretinal space on both T1- and T2-weighted images. Retinal gliosis was hypointense on both T1- and T2-weighted images. On contrast-enhanced T1-weighted images with fat suppression techniques, minimal to marked enhancement was observed in retinoblastoma, medulloepithelioma, retinal capillary hemangioma, leiomyoma, and choroidal melanoma. No enhancement was demonstrated in the subretinal space of Coats disease, PHPV, ROP, and in retinal gliosis. Calcification was identified in 54% of retinoblastoma tumors on MR sequences. Calcification showed low signal intensity on both T1- and T2-weighted images, with lack of enhancement on contrast-enhanced sequences, thus differentiating it from surrounding tumor or tissue necrosis. The associated serous or exudative subretinal fluid secondary to intraocular tumors or simulating lesions did not demonstrate enhancement after contrast administration, therefore differentiating it from the causative lesion.

CONCLUSION: Pre- and post-contrast MR studies allowed differentiation of solid intraocular tumors such as retinoblastoma, medulloepithelioma, retinal capillary hemangioma, leiomyoma, and choroidal melanoma from intraocular lesions with primary retinal detachment such as Coats disease, PHPV, massive retinal gliosis (phthisis bulbi), ROP, and associated subretinal fluid or hemorrhage. Heterogeneity within retinoblastoma suggests foci of tumor necrosis and/or calcification. The various solid intraocular tumors were not reliably differentiated from one another based on MR features.