Astroblastoma is a rare glial neoplasm, which was not initially recognized. Even after astroblastoma was first reported by Bailey and Cushing, its existence and origin remained controversial [8]. Based on microscopy observation, astroblastoma was initially classified as a transitional type between the astrocytoma and the glioblastoma multiform [9]. Some scientists believed that GFAP and vimentin positivity in astroblastoma supported the hypothesis that this tumor was derived from the cytogenetically more primitive astroblast [10, 11], or arose from a process of dedifferentiation involving mature astroglial cells [12]. However, an ependymal or tanycyte-derived origin of astroblastoma is considered by many authors based on the electron microscopic features [13, 14]. Tanycytes are suggested to be glial precursor cells that may occur during normal human embryogenesis, which explains the existence of congenital astroblastoma [15, 16]. Astroblastoma is often mixed with other types of tumor cells such as glioblastomas or anaplastic astrocytomas, and the pseudo-chrysanthemum cluster structure also appears in other tumors, leading to controversy about its existence [17].
With the advancement of gene identification technology, numerous specific genes have been identified in astroblastomas. The most frequent gene alterations detected were MN1 mutation, gain of chromosome arm 20q and chromosome 19, loss of 9q, 10 and X [18–20]. These findings suggested that astroblastomas represent a distinct entity with characteristic cytogenetic features that differ from those of ependymomas and astrocytomas. Because large sample gene detection studies have not been conducted for astroblastomas, there is no unified gene mutation spectrum. Moreover, tumors with histological features of astroblastoma may result from diverse and possibly distinct genetic events [21].
Perivascular pseudorosettes are also found in other tumors. So the histological differential diagnoses of astroblastomas include ependymomas and papillary meningiomas. The astroblastoma in our patient exhibited broad footplates as opposed to the tapering processes seen in ependymoma. Moreover, in contrast to ependymomas, the spaces between the pseudorosettes were often rarified. EMA expression was previously shown to be typically localized at the cell membrane in astroblastomas [22–24], which was also observed in our patient. In ependymoma, EMA is expressed along the luminal surface of some ependymal rosettes or manifests as dot-like perinuclear or ring-like cytoplasmic structures [25]. Therefore, we ruled out the diagnosis of ependymoma based on morphology and immunohistochemistry. The distinction between astroblastomas and papillary meningiomas is aided by immunohistochemical features such as positive GFAP staining in astroblastomas. Interestingly, the oligodendrocyte-like cells were outside the papillary areas in our patient, with clear boundaries in between. Meanwhile, neither IDH 1/2 mutation nor 1p/19q co-deletion was observed, so the diagnosis of oligodendroglioma was ruled out. Based on these findings, the tumor was diagnosed as astroblastoma with oligodendroglial-like cells.
Oligodendroglial-like cells can appear in many central nervous system tumors, such as dysembryoplastic neuroepithelial tumor (DNT) [26], rosette-forming glioneuronal tumor (RGNT) [27], papillary glioneuronal tumor [28], diffuse leptomeningeal glioneuronal tumor [29], and gangliogliomas (GGs) [30]. There is only one previous report by Lehman NL and Hattab EM that found scattered oligodendroglial-like cells between astrocytoma cells, which is different from our patient [31].
Different terminologies have been previously used to describe oligodendroglial abnormalities including oligodendroglial hyperplasia [32], clusters of oligodendroglia [33], oligodendroglial hamartoma [34], and oligodendroglial-like cells. These lesions may represent a spectrum of the same abnormality. However, the appearance of oligodendroglial-like cells under the microscope was not the same, visible oligodendroglial-like cells floated in the mucus-like matrix, or infiltrated the tumor tissue, or were arranged in bundles. Oligodendroglial-like cells and oligodendroglioma have different gene mutation spectra. Despite oligodendroglial-like morphology, oligodendroglial-like cells do not have chromosome 1p and 19q deletions or IDH1 mutation [35]. Our patient did not have IDH1/2 mutation, or 1p and 19q deletions. Moreover, there was a clear boundary between the tumor tissue and oligodendroglial-like cells. Hence, we hypothesized that the oligodendroglial-like cells in this case were not oligodendroglioma components, but oligodendroglial-like cell hyperplasia.