CNS embryonal tumors (CET) other than medulloblastomas (MB) and atypical teratoid/rhabdoid tumors (AT/RTs), previously designated as \'central nervous system primitive neuroectodermal tumors\' (\'CNS PNETs\'), are a heterogenous subset of tumors with poorly defined diagnostic criteria. Other than the subset of embryonal tumor with multilayered rosettes (ETMR) defined by C19MC amplification, most CETs are diagnosed by exclusion of other molecularly defined entities and histological mimics including MB, AT/RTs, and high-grade gliomas, and termed as CET, not otherwise specified (NOS) in the 2016 WHO classification.
To reclassify \'CNS PNETs\' as per WHO 2016, and estimate the true proportion of CET, NOS in a tertiary healthcare setting, and to evaluate the diagnostic utility of C19MC amplification, Lin28A and Olig2 expression in the subclassification of CETs.
Previously diagnosed cases of \'CNS PNETs\' (2002-2016) were first evaluated by immunohistochemistry (IHC) for MIC2, RelaA, L1CAM, IDH1R132H, H3K27M, H3G34R, H3G34V, INI1, and BRG1 proteins and by fluorescence in-situ hybridization (FISH) for EWSR1 translocation to exclude histological mimics. The selected CETs (case cohort) and 79 histological mimics (comparison cohort) comprising of MB, AT/RT, pineal parenchymal tumors, Ewing sarcoma, esthesioneuroblastoma, intraocular medulloepithelioma, and H3G34R mutant high-grade glioma were subject to IHC for Olig2 and Lin28A, and FISH for C19MC amplification.
Twenty-two cases of \'CNS PNETs\' were retrieved, all of which were negative for the first panel of markers and showed retained INI-1/BRG1 expression. Three of them (3/22, 13.6%) showed C19MC amplification (ETMR, C19MC-altered) with ETMR histology, Lin28A positivity, and Olig2 negativity. Among the remaining 19 CETs, one showed medulloepithelioma histology (Medulloepithelioma, NOS) and remaining were non-descript small round cell tumors (CET, NOS), all negative for Lin28A. Olig2 was positive in only 3 CETs (13.6%), all being CET, NOS. All tumors in the comparison cohort were negative for C19MC amplification, Lin28A and Olig2 except for 27% of ATRTs that were Lin28A positive.
ETMR, C19MC-altered constitute less than 14% of CETs, with majority remaining uncharacterized as CET, NOS. Lin28A is 100% sensitive for the detection of C19MC amplification; however, its specificity is limited by its expression in ATRTs. Olig2 expression is seen only in a small subset of CET, NOS and is of limited diagnostic utility.