The vast majority of intramedullary spinal cord tumors are gliomas, comprised of ependymomas and astrocytomas. While the astrocytic tumors form about 25 percent of these gliomas, they must be still regarded as rare. The CBTRUS (2000) indicates that the mean age adjusted incidence rate of spinal cord tumors of all histologies is 0.56 per 100 000. However, experience indicates the majority of these tumors are extramedullary tumors of the meninges and spinal nerves, and a more precise estimate of the incidence of intramedullary tumors is sparse. A Norwegian population based data set analyzed by Helseth and Mørk (1989) indicates a prevalence of 0.15/100 000 in females and 0.09/ 100 000 in males. Astrocytomas are one-third as common as intramedullary ependymomas. In children under the age of 15, astrocytomas account for 90 percent of the intramedullary tumors that come to clinical attention (Miller and McCutcheon, 2000) and approximately 25 percent of all spinal tumors (Guidetti and Fortuna, 1974). Because intramedullary spinal cord astrocytomas more often affect the upper segments of the cord, they may be more frequent than the ependymomas in that region, while the latter predominate in the caudal segments. Occasionally the upper cervical segments may be involved in a spreading growth that arose primarily in the brainstem. In the series of Slooff and colleagues (1964), 35 of their 66 astrocytomas (Grades I and II) arose in the thoracic segments, 11 in the cervical segments and 10 in the cervicothoracic region. Another series of cases showed an equal distribution of astrocytomas between the cervical and thoracic cord, with only one of 79 tumors confined to the lumbar region (Minehan et al., 1995). Other studies have further supported that, in children and adolescents, the cervical segments seem to be the most
frequently involved (Reimer and Onofrio, 1985; Rossitch et al., 1990). The rarity of these lesions, particularly in children, has made generalizations on prognostic factors and therapeutic outcomes of limited reliability (Bouffet et al., 1998). However, age seems to be an independent prognostic variable even when gender, grade, therapy, and procedure are also considered (Santi et al., 2003).