Comparative integrated molecular analysis of intraocular medulloepitheliomas and central nervous system embryonal tumors with multilayered rosettes confirms that they are distinct nosologic entities

Abstract

Intraocular medulloepithelioma (IO MEPL) is an uncommon embryonal neuroepithelial neoplasm of the eye. These ocular neoplasms have been compared with intracranial medulloepitheliomas or other histologic variants of CNS embryonal tumor with multilayered rosettes (CNS ETMR) due to their morphological mimicry. We performed comprehensive molecular analysis to explore the histogenetic and biologic relationships between 22 IO MEPL and 68 CNS ETMR. Routinely prepared paraffin-embedded samples were assessed for genome-wide methylation profiles using the Illumina Methylation 450k BeadChip array. We identified strong cytogenetic and epigenetic differences between ocular neoplasms and CNS ETMR. None of the IO MEPL cases displayed the ETMR-specific amplification of the C19MC locus. Instead, cytogenetic analysis of the IO MEPL showed numerous copy number aberrations which involved either whole chromosomes or chromosomal arms; recurrent aberrations in these tumors affected chromosomes 1p, 4, 8 and 16p. DNA methylation patterns were also strikingly different between these two tumor entities, suggesting that they do not share common origins and biological behaviors. Comparative cluster analysis of 198 pediatric CNS tumors and 22 IO MEPL revealed a clear demarcation of the CNS ETMR and IO MEPL profiles from other CNS entities. In conclusion, although IO MEPL shares some histopathological features with CNS ETMR, they manifest striking molecular diversity at the cytogenetic and epigenetic levels. Consequently they deserve a separate nosologic designation in future tumor classifications, where CNS MEPL could be designated as a histological variant of CNS ETMR.

Nonteratoid Medulloepithelioma Presenting in a 78-Year-Old Male

“…In two large series, Korshunov et al [13] have recently proved that intraocular medulloepitheliomas and embryonal tumors of the central nervous system with multilayered rosettes are distinct nosologic entities, and Sahm et al [14] found a high prevalence of DICER1 and KMT2D somatic mutations. Both the punch and FNA biopsies of our tumor showed trisomy of chromosome 3, which was rare in those series.…”

“…The cytogenetics of medulloepithelioma are only recently beginning to be elucidated, with many genetic alterations including abnormalities in chromosomes 1, 4, 6, 8, 15, and 16 [11,12,13]. In two large series, Korshunov et al [13] have recently proved that intraocular medulloepitheliomas and embryonal tumors of the central nervous system with multilayered rosettes are distinct nosologic entities, and Sahm et al [14] found a high prevalence of DICER1 and KMT2D somatic mutations.…”

Early Neuroblastic and Astrocytic Differentiation Demonstrated Immunohistochemically in a Small Intraocular Medulloepithelioma

“…A similar tumor is found in the brain but has different genetic and epigenetic findings [2,3,4,5]. For every 100-150 retinoblastomas 1 ocular medulloepithelioma may occur.…”

On the Classification and Grading of Medulloepithelioma of the Eye

“…The cerebral medulloepithelioma is often large at diagnosis and is by definition classified as a high-grade malignant tumor although long-term survivors have been described. Only recently has the demonstration of molecular diversity at the cytogenetic and epigenetic levels enabled the identification of brain and ocular medulloepithelioma as separate nosologic entities [5,6]. Strict separation between a benign and malignant variant of intraocular medulloepithelioma does not reflect the clinical behavior of these tumors for as far as it can be deduced from small case series and case reports [7,8].…”

“…This notion is supported by the recently published molecular data on intraocular medulloepithelioma, which showed clustering in the same part of the dendrogram tree using unsupervised cluster analysis. There were no additional subclusters within the intraocular medulloepithelioma cohort separating histologically malignant from nonmalignant cases [6]. Neither could single nucleotide polymorphism array detect differences in copy number aberrations between benign and malignant medulloepithelioma.…”