DIAGNOSIS
Ganglioneurocytoma, WHO grade 2.
DISCUSSION
Neurocytomas are uncommon WHO grade II tumors typically arising near the foramen of Monro, and in this location are designated central neurocytoma (CN) (3). Extraventricular neurocytoma (EVN) has been reported at various locations in brain (1,2) but EVN in spinal cord (7-9) is rare; only 16 previous cases have been reported in the English literature (9), typically arising as an intramedullary lesion in the cervico-thoracic segment of the spinal cord. Oddly, there is a strong male predilection for spinal EVNs (2), in contrast with other EVNs which exhibit a slight female predilection. EVN likely arises from neuronal precursor cells surrounding the region of central canal in fetal life, but there is no direct evidence of its lineage; the molecular underpinnings of CN/EVN have yet to be elucidated.
CN and EVN typically show uniform round cells with neurocytic differentiation; the cytoplasm often merges imperceptibly in a neuropil-like pattern. Occasionally there is mature ganglion cell differentiation admixed with a neurocytic neoplasm, to which the designation 'ganglioneurocytoma' has been applied (5). While 4/16 cases of EVN reported in the spinal cord showed ganglionic differentiation (25%), this is somewhat less frequent than EVNs located in other regions (66%) (9). Herein, we report another rare case of spinal neurocytoma with ganglionic differentiation arising from the cervical spinal cord.
EVN may elude diagnosis since it has clinical, radiological, and pathological similarities to other intramedullary spinal tumors, particularly ependymoma and also oligodendroglioma. Our case was initially diagnosed as ependymoma on imaging, on frozen sections, and by the referring pathologist on permanent sections even with immunohistochemistry, since background GFAP expression had been attributed to the neoplastic cells. The distinction from ependymoma and oligodendroglioma can be suspected by careful examination of the neuropil-like matrix, and immunohistochemical evidence of neuronal origin is definitive. Perivascular ('pseudo') rosette formation is a classic but non-specific feature of ependymoma, which tends to exhibit a coarser glial fibrillary matrix, with orthogonal orientation of processes towards the surrounded vessels. Because of their consistent presence in ependymoma, perivascular rosettes may lead to a 'first impression diagnosis', but an appropriate differential diagnosis should be constructed to avoid an erroneous pitfall. GFAP expression should be assessed carefully to determine whether the cells of interest, rather than intervening reactive processes, are responsible for any observed immunopostivity; synaptophysin is the most suitable and reliable immunohistochemical marker of CN (10) and EVN; however oligodendrogliomas with neurocytic differentiation and synaptophysin expression have been documented (6).
In general, CN and EVN have an excellent prognosis (7). More aggressive behavior has been linked to an increased Ki-67 index, greater than 2% being considered as clinically significant in one laboratory (4), although some reports indicate all EVNs show greater than 3% positive nuclei (1); whether Ki-67 will be a reliable indicator of aggressive behavior thus remains unresolved.
Postoperatively, our patient developed some sensory and motor deficits in both upper and lower limbs and neuropathic pain. Two and half years post-operatively she was admitted to the rehabilitation centre with functional decline relating to weakness. After several months of rehabilitation, symptoms improved significantly. Multiple follow up MRI have shown no evidence of residual or recurrent tumor with only post-operative myelomalacia of the cervical spinal cord at the C2 level, and is now more than 4 years since surgery. While maximal safe resection of CN and EVN is the treatment of choice, extensive surgical resection may not be appropriate (9); several reports suggest a long survival period can be expected even after subtotal resection of these tumors as they have indolent nature and low proliferative potential (7-9). Adjuvant radiotherapy is not indicated for typical spinal EVN, but can be used in exceptional cases (4).
In conclusion, EVN or ganglioneurocytoma of the spinal cord, albeit rare, should be considered in assessing a heterogeneously enhancing intramedullary spinal tumor, especially in view of the clinical and morphological resemblance to more common tumors. EVN has a favorable prognosis and even after subtotal resection of these tumors, can have long postoperative period without recurrence without adjuvant treatment. Our case describes the seventeenth case of EVN of the spinal cord, and only the fifth case justifying the designation 'ganglioneurocytoma'.
REFERENCES
Contributed by Namita Sinha, MD, Jill Wooff, Jai J S Shankar, Peter Gorman, Robert J Macaulay