Spinal Ependymomas. Part 1: Intramedullary Ependymomas

Jörg Klekamp, MD


Neurosurg Focus. 2015;39(2):e6 

In This Article


A total of 1447 patients presented at the Nordstadt Hospital in Hannover, Germany, between 1980 and 2003 or the Christliche Krankenhaus in Quakenbrück, Germany, from 2004 to 2014 with tumors of the spinal canal and their cases were entered into the spinal cord database. All patients agreed in writing upon admission that their medical data may be used for scientific purposes in anonymous form. Data prior to 1991 were collected retrospectively. Since 1991, all patients have been continuously entered in the database. Apart from general patient data and specific features of each spinal cord pathology, the neurological examinations before surgery, before discharge from the hospital, 3 months postoperatively, and yearly thereafter by outpatient visits and questionnaires were analyzed for individual symptoms according to a scoring system[13,15] (Table 1) and the McCormick Scale.[23] Among 1447 patients with spinal tumors, 309 patients presented with intramedullary tumors. Of these, 100 patients with ependymomas underwent a total of 102 operations since 1982. Unless a patient appeared unfit to undergo an operation, surgery was recommended to all symptomatic patients with a de novo tumor. In patients with recurrent tumors, surgery was restricted to patients with evidence of tumor growth or progressive neurological symptoms.

Before the introduction of MRI, the diagnosis was based on CT with contrast or myelography in 3 patients, while the remainder presented with MRI.

All patients were operated on with intraoperative neurophysiological monitoring of somatosensory-evoked potentials. Monitoring of motor-evoked potentials was added routinely in 2011. Patients with cervical tumors were operated on while in the semisitting position and all patients with other tumors were in the prone position. Anesthetists were asked to maintain mean arterial blood pressure at 85 mm Hg throughout the intradural dissection. The exposure involved laminotomies with reinsertion of the lamina for children, while laminectomies were employed for adults up to the mid 1990s, when lamina were reinserted with increasing frequency in adults as well. Since 1995, all lamina are reimplanted. The microsurgical technique employed was based on the work of Malis[20] and Stein[29] and has been published in detail previously[11,14] (Fig. 1). It was general practice to determine the exact extension of the tumor with ultrasonography before myelotomy. With the exception of exophytic growing ependymomas (Fig. 2), all were removed via a midline approach (Fig. 1). As of 1994, the cord was closed with 7–0 pial sutures after tumor removal.

Figure 1.

Case of a 56-year-old woman with an ependymoma at the C5–6 level, with a 2-year history of local pain followed by a slight gait ataxia. A: Sagittal T1-weighted MR image with contrast displays an ill-defined lesion with diffuse boundaries. B: Sagittal T2-weighted MR image demonstrates a clearly demarcated tumor. C–l: Photographs showing intraoperative views throughout tumor resection. In the semisitting position, dura and arachnoid opening displays the midline of the spinal cord marked by a small sulcus (C). After opening of the pia mater, the posterior tracts are spread apart displaying the small vessels in their course along the posterior midline septum of the spinal cord on either side (D). This vascular display marks the anatomical midline of the cord. Once the tumor is reached, the pia is sutured to the dura (E). With continuous debulking a good cleavage plane can be dissected on the right side. The upper pole of the tumor has been dissected free, displaying the ventral part of the tumor bed (F). Small tumor-feeding arteries from the anterior spinal cord artery are isolated, coagulated, and cut (G). In this fashion, the tumor has been completely removed (H). Hemostasis is achieved by repeated irrigation avoiding bipolar coagulation. The cord is closed with 7–0 pia sutures (I). Two years after surgery, the patient's neurological state is unchanged. She developed a pain syndrome, which is well controlled with medication. J and K: Sagittal T1-weighted MR image (J) with contrast showing no tumor remnant or recurrence, and axial T2-weighted MR image (K) showing only minor adhesions between spinal cord and dura in the midline.

Figure 2.

Case of a 47-year-old woman with an exophytic ependymoma at the C1–2 level, with a 3-month history of a slight tetrapa-resis. A-C: MR images showing the tumor with varying degrees of enhancement with a clear demarcation. Sagittal T1-weighted MR image (A) and sagittal (B) and axial (C) T2-weighted MR images showing a small cyst in the lower part of the tumor and the exophytic growth of this lesion on the left side. D-F: Photographs showing intraoperative views of the patient in the semisitting position and the tumor can be seen protruding out of the cord on the left side (D). The pia has been incised along the tumor border on the right side (E). With debulking of the tumor a cleavage plane could be followed, as described in Fig. 1, without requiring an additional myelotomy. The tumor has been completely resected (F). G and H: Postoperative sagittal (G) and axial (H) T2-weighted MR images showing the complete resection. The patient's neurological condition remained unchanged without developing a neuropathic pain syndrome.

The following categories were distinguished according to the amount of resection: a gross-total resection (GTR) was defined as no tumor remnant detectable at the end of surgery as well as in the postoperative MRI. A subtotal resection was defined as a complete resection of the tumor mass with a small remnant detectable on the postoperative MRI. A partial resection was defined as an incomplete resection of the tumor mass, while operations with less than 50% reduction of the tumor mass were classified as biopsies.

The 102 operations were divided among 15 neurosurgeons. Three groups of surgeons were distinguished according to the number of operations performed on all intramedullary tumors: Group A, fewer than 20 operations (11 surgeons, 16 operations); Group B, 20–30 operations (2 surgeons, 15 operations); Group C, more than 70 operations (2 surgeons, 71 operations). The overall study period was divided into 3 periods: Period A, 1982–1990 (n = 24); Period B, 1991–2000 (n = 42); Period C, 2001–2014 (n = 36).

For statistical tests of significance, Student t-tests and chi-square tests were employed. Means are given with the standard deviation. Results of multiple regression analyses are provided with b-values to indicate the impact of individual factors, the correlation coefficients of the corresponding regression equations, and their p values. Long-term follow-up was analyzed with Kaplan-Meier statistics[9] to determine the percentages of patients with tumor recurrences. Differences were considered statistically significant, if a p value ≤ 0.05 was reached. For all statistical analyses the software package "PC-Statistik" version 5.0 was employed (Hoffmann-Software). Surgical morbidity was defined as a permanent aggravation or a new, permanent deficit of motor function, gait, or sphincter function within 30 days after the operation. Sensory functions, dysesthesias, and pain were analyzed separately. A postoperative neuropathic pain syndrome was defined as postoperative neuropathic pain requiring medication. A postoperative myelopathy was defined as a neurological deterioration in the absence of tumor growth or recurrence.