A 25‐Year‐Old Man with a Posterior Fossa Mass

Clinical History

A 25‐year‐old gentleman presented with short history of headache and gait disturbances. Radiology revealed a heterogeneously enhancing mass in the fourth ventricle extending from midbrain to pons. CT scan showed a hyperdense mass in the fourth ventricle. There was no calcification or cysts seen within the lesion. MRI showed a well‐defined mass in the fourth ventricle. The lesion was isointense on T1, hyperintense on T2 (Figure 1A) and enhanced heterogeneously on contrast administration (Figure 1B). The contrast images showed tumor infiltration into brain stem suggesting a poor plane of cleavage, better appreciated on sagittal MRI (Figure 1C). FLAIR MRI showed slightly hyperintense lesion suggesting a cellular tumor (Figure 1D). Though the lesion showed no calcifications or cysts or areas of bleed, the enhancement pattern and infiltration in the fourth ventricular floor suggested a radiological diagnosis of ependymoma. The lesion was operated through a midline suboccipital craniotomy. The lesion was yellow gray, not very vascular, and was attached to the brain stem. A thin sliver of tumor adherent to the brain stem was left. Near total resection could be achieved.

Figure 1.

 

Microscopic Pathology

The histopathological examination revealed characteristic features of a tumor, which was defined by perivascular pseudorosettes, elongated round to oval uniform nuclei, present in a fibrillary background (Figure 1E). Amidst these areas, interestingly, there were numerous islands composed of pale central nuclear‐free neuropil‐rich zones (Figure 1F,G). These on the periphery were bordered by similar tumor cells (Figure 1H). The transition between these islands and tumor cells was abrupt. On a low magnification, these islands were strikingly similar to pseudo‐palisading necrosis of high‐grade gliomas. Tumor cells were immunoreactive for glial fibrillary acidic protein (GFAP) with perivascular accentuation (Figure 1I). Epithelial membrane antigen (EMA) revealed distinct dot‐like positivity within the tumor cells (Figure 1J). The pale neuropil‐like islands were intensely positive with synaptophysin (Figure 1K). No immunostaining for NeuN was observed. Mitotic activity was brisk with intermediate to high Ki‐67 labeling index. Necrosis and vascular proliferation were absent. What is your diagnosis?

Diagnosis

Anaplastic ependymoma (WHO Grade III) with neuropil‐like islands.

Discussion

Neuronal differentiation in ependymoma, especially in the form of these distinct neuropil‐like islands, is extremely rare and finds its mention in the literature in only three case reports to date 1, 2, 3. It was first characterized by Gessi et al in 2005 and was thought to induct the tumor in the category of mixed glio‐neuronal tumor 2. Subsequently, it was better defined in a paper by Rodriguez and colleagues who identified such differentiation at histology, by immunohistochemistry alone or at ultrastructural level in their ependymoma cohort 3. Their findings suggested this differentiation to be an unusual feature both at histology and immunohistochemical level, and ultrastructural features were defined to be the most reliable criterion for its presence. Such a differentiation occurring in a pure glial tumor supported the speculations of its origin from cancer stem cells which were substantiated in landmark paper by Taylor et al 4. These synaptophysin/MAP‐2 positive neuropil islands bordered by tumor cells are a potential diagnostic pitfall as it morphologically mimics pseudo‐palisading necrosis of high‐grade gliomas, at low magnification. Another differential diagnostic possibility at this location is a ganglioglioma, which is characterized by biphasic histology with the presence of dysmorphic ganglion cells. Immunohistochemistry is of special value, as it readily resolves the differentials. Characteristic dot‐like positivity for EMA and perivascular accentuation for GFAP are key features of ependymoma.

Neuronal differentiation in ependymoma has so far been observed in supratentorial examples while the index case is in infratentorial location. Given the rarity, the clinical behavior and prognostic impact of such histologic feature are difficult to predict; however, the earlier examples had similar association with anaplastic features as seen in the index case. Hence a close and extended follow‐up is advised. Interestingly, the tumor recurrences were devoid of such neuropil–like islands.

More recently, with molecular stratification of supratentorial and infratentorial ependymomas, the genetic landscape of these unusual morphological feature observed in anaplastic ependymoma should be explored. These are likely to be enriched/upregulated for genes associated with neuronal differention. In conclusion, neuronal differentiation is a very rare and an elusive feature; awareness about its occurrence is critical to avoid misdiagnosis. The description of these features in a small case series with follow‐up data will elucidate its clinical significance in ependymal tumors.