DIAGNOSIS: PARAGANGLIOMA OF THE CAUDA EQUINA, GANGLIOCYTIC VARIANT
The clinical and radiological findings, as well as the first-glance morphological impression led to the preliminary assumption, that the tumor in this 63-year old woman was an ependymoma. In particular, the perivascular pseudorosettes, ependymal canal-like structures and the monomorphic cellular pattern seemingly supported this diagnosis. However, the presence of a tumor capsule, the compartmentalization of tumor cells into zellballen surrounded by reticulin fibers, and the presence of ganglionic cells were untypical of ependymoma and instead were indicative of paraganglioma with a gangliocytic component. Immunoreactivity of the tumor cells for neuroendocrine antigens, the detection of GFAP-positive sustentacular cells and the ultrastructural confirmation of neurosecretory granules substantiated this diagnosis.
Paragangliomas arise from specialized neural crest cells associated with autonomic ganglia throughout the body. Most frequently, they occur in the glomus caroticum or the glomus jugulare. In the CNS, paragangliomas are almost exclusively located in the cauda equina (1). The first case of spinal paraganglioma was reported in 1970; interestingly, this tumor was described as "secretory ependymoma" (2). To date, more than 120 cases of paragangliomas of the CNS have been reported. Paragangliomas of the cauda equina usually occur in adults; the mean age at presentation is 47 years (3) with a reported age range of 12-71 years (1, 4). The male:female ratio is 1.4:1. Cauda equina paragangliomas correspond to WHO grade I and are usually cured by surgical resection. The recurrence rate is approximately 4%, and CSF-seeding or distant metastases are extremely rare (1). In a study of 31 cases, ganglionic differentiation was detected in 45% of the tumors (3). In another study of 30 patients with spinal paraganglioma, ganglion cells were detected in only one case (5). Clinically, the most common symptom of cauda equina paraganglioma is lower back pain accompanied by sciatica. Endocrine activity of these tumors is extremely rare, and only two cases with functional hormonal activity have been reported (6, 7). Radiologically, paragangliomas of the cauda equina are typically hypo- or isointense to the spinal cord on T1-weighted images, hyperintense on T2-weighted images, and strongly contrast-enhancing (1). Serpentine vessels, as observed in the present case, are not uncommon and are considered a major clue to the diagnosis of a highly vascular lesion (8).
The histogenesis of cauda equina paragangliomas remains enigmatic. Interestingly, Caccamo et al. have described a cauda equina tumor with both ependymal as well as paraganglionic differentiation, confirmed both immunohistochemically and ultrastructurally (9). The authors speculated that this tumor arose from elements normally found in the human filum terminale, such as ependymal cells, ganglionic neurons and neuroblasts. Interestingly, in teleost fish, ependymal cells lining the central canal of the distal spinal cord, can differentiate into large neurosecretory cells (10). Moreover, in human newborns the existence of ependymal cells with neurosecretory activity has been reported in the filum terminale (2). In rodents, cerebral ependymal cells have even been described as origin of neural stem cells (11). Despite these studies, any transdifferentiation-hypotheses regarding ependymal and neuroendocrine/neuronal cells in humans are at present highly speculative.
To conclude, the clinical, radiographic and morphological similarity between ependymomas and paragangliomas in the cauda equina region has repeatedly led to substantial diagnostic confusion. Ependymal tumors are far more common in this region than paragangliomas, which can therefore easily be overlooked on conventionally stained sections. Immunohistochemical and/or ultrastructural analyses are essential to distinguish these tumors in the cauda equina region.
Contributed by Jakob Matschke, Manfred Westphal, and Katrin Lamszus