Primary spinal intramedullary primitive neuroectodermal tumor.

Though primitive neuroectodermal tumor has been considered to arise from a neoplastic transformation of primitive neuroepithelial cells with propensity to involve any part of the central nervous system, a primary intramedullary spinal primitive neuroectodermal tumor is very unusual. The authors describe a case of an 18-year-old female with conus intramedullary tumor diagnosed to be primary spinal primitive neuroectodermal tumor following histopathological examination after surgery. The diagnosis of such a tumor is very crucial as the management strategies for these are relatively unclear and are associated with a poorer outcome compared to the other common intramedullary spinal tumors.

Introduction

Ever since the term primitive neuroectodermal tumor (PNET) has been proposed by Hart and Earle to define a group of malignant neoplasms of presumed neural crest origin, a number of cases of the same have been described at various locations.[1] Primary PNET of the spine is unusual with very few cases reported in the literature.[2-4] Occurrence of primary spinal PNET in an intramedullary location is further uncommon[4]; more so in children. Very few case reports of primary pediatric spinal intramedullary PNET are reported in the literature.[3-10] We describe a case of an 18-year-old female with conus intramedullary tumor diagnosed to be a spinal primary PNET following surgery The diagnosis was not suspected until the same was confirmed following final histopathological examination. The importance of recognizing the same lies in the different management strategy and therapeutic outcome compared to the usual intramedullary spinal tumors for which surgery forms the cornerstone of treatment and are associated with a better clinical outcome compared to spinal PNET.

Case Report

An 18-year-old female presented with complaints of low back ache and weakness involving bilateral lower extremities for 3 months along with bowel bladder involvement for 1 week. Examination revealed tenderness in lower back with hypotonic and paraplegia (0/5 Medical Research Council grade) in bilateral lower limbs. Ninety percent sensory loss to all modalities of sensation was present with no evidence of sacral sparing. Magnetic resonance imaging (MRI) of the spine revealed an intramedullary lesion, predominantly isointense on TIW MRI with homogenous contrast enhancement and interspersed cystic/necrotic areas from D11-L2 [Figures 1 and 2]. A differential diagnosis of common intramedullary tumors namely ependymoma (myxopapillary variant), paraganglioma, other glial neoplasms, metastases or any inflammatory pathology was considered. Laminectomy and tumor decompression was performed. . Histopathology revealed a highly cellular tumor, consisting mainly of small round to oval cells with hyperchromatic nuclei and remarkably scanty cytoplasm along with the presence of Homer-Wright pseudorosettes; the tumor cells being immunopositive for CD 99 [Figure 3]. Bone scan and an intracranial MRI were done which did not reveal any evidence of tumor elsewhere and a diagnosis of primary intramedullary PNET was made. Post-operative course was uneventful with minimal improvement in power and the patient is awaiting adjuvant chemoradiotherapy.

Figure 1

T1W Magnetic resonance imaging showing a predominantly isointense intramedullary tumor

Figure 2

Sagittal (a) and axial (b) contrast enhanced Magnetic resonance imaging showing the homogenously enhancing intramedullary tumor with interspersed cystic/necrotic areas

Figure 3

Microphotograph showing a highly cellular tumor, consisting mainly of small round to oval cells with hyperchromatic nuclei and remarkably scanty cytoplasm along with the presence of Homer-Wright pseudorosettes; the tumor cells being immunopositive for CD 99 (inset)

Discussion

PNETs are malignant tumors composed of highly undifferentiated neuroepithelial cells comprising of various lesions that are histologically similar and usually occur in the pediatric age. These are located mainly intracranially with management being surgical excision followed by craniospinal radiotherapy and a radiation boost to the primary tumor with or without chemotherapy. It has been postulated that PNETs arise from neoplastic transformation of primitive neuroepithelial cells in subependymal zones which are present in all areas of the central nervous system and may explain the presence of PNETs at locations other than the cerebellum.[211] Due to unknown reasons, primary PNET involving the spine is rare and can be intramedullary, intradural–extramedullary, or epidural, with the majority arising from the cauda equina.[2-412-18]

Drop metastasis from primary intracranial tumors disseminating via cerebrospinal fluid accounts for majority of spinal PNET.[2] Primary intraspinal PNETs are rare. Though, these tumors differ from central PNET in their immunohistochemical profile, genetic backgrounds and potential to metastasize outside CNS; the overall prognosis and survival is equally dismal.[2-4] Intramedullary primary PNET can present in both children and adults with no sex predilection across the whole spinal column with preferential involvement of the thoracic region.[34] Intramedullary location of a primary PNET is uncommon especially in children,[3-10] where glial neoplasms account for majority of intramedullary tumors. Only eight cases of clearly documented purely intramedullary spinal PNETs have been reported in the pediatric age group.[3-10] Though spinal PNET behaves differently from that of spinal cord astrocytoma or ependymoma, the common tumors at this location, the imaging features does not differentiate between them. This precludes the preoperative diagnosis of spinal PNET which usually is based on the histopathological finding with confirmation by immunohistochemical analysis. However, due to rarity of spinal cord PNET, the possibility of metastasis from a central or peripheral PNET should be ruled out before arriving at the diagnosis of primary spinal PNET as was done in the described case. Magnetic resonance imaging of the whole brain and spine or a FDG PET can be performed to exclude these possibilities.

The management and prognosis of this highly uncommon lesion in this location is different from the more common tumors involving this region and is uniformly poor. In view of the good clinical response to chemoradiation in medulloblastoma or supratentorial PNET the same should be considered even in spinal PNET, not withstanding the fact that complete excision is difficult to achieve in this eloquent location.

To conclude, PNET forms a very unusual differential for intramedullary spinal tumors. Recognition of this entity at this unusual location is important for prognostication and management as these tumors involve multidisciplinary treatment with poor overall outcome compared to other intramedullary tumors of the spinal cord which it can resemble radiologically.