Between 1997 and 2013, 33 patients with ependymomas underwent surgery in the Chonnam National University Hospital and Chonnam National University Hwasun Hospital. A retrospective review of medical records was performed in all patients. This study was approved by our Institutional Review Board (CNUHH-2017-084). Participants in research consented to being involved in this study. Pediatric patients were defined as patients under 18 years old.

Ependymomas were classified as World Health Organization (WHO) grade II ependymomas or grade III anaplastic ependymomas according to the WHO classification system. Data regarding patients' clinical characteristics, such as age, tumor location, extent of removal, pathologic grade, postoperative adjuvant treatment, PFS, and OS, were retrospectively analyzed. Postoperative adjuvant treatment was performed for all WHO grade III ependymoma patients. In WHO grade II ependymoma patients, adjuvant treatment was recommended according to the extent of resection. In adult grade II ependymoma patients with gross total resection (GTR), adjuvant treatment was not recommended. One adult grade III ependymoma patient with STR did not receive adjuvant treatment due to poor general condition. Radiotherapy was performed 4 weeks after the operation for consideration patient's general condition and surgical wound. If the patient's general condition was not good, we considered the patient's condition when determining treatment timing. Conventional external beam radiotherapy was performed in 13 patients. There was no radiation-induced necrosis during follow-up. Adjuvant chemotherapy was performed for a 15-month-old girl.

The high-risk group was defined as patients with WHO grade III tumors or subtotally resected tumors, and the other patients were assigned to the low-risk group. To evaluate efficacy of adjuvant treatment, adult patients were divided into three groups: high risk with adjuvant treatment, high risk without adjuvant treatment, and low risk without adjuvant treatment.

We evaluated the extent of removal by evaluating the intraoperative findings and postoperative contrast-enhanced computer tomography images. The extent of removal was classified as total (100%) tumor removal, subtotal (over 50%), or partial (lower than 50%). Tumor recurrence was based on the presence of a tumor on follow-up magnetic resonance images (MRI) or the development of symptoms. We performed follow-up brain MRI scans every 6 months. We confirmed tumor recurrence by brain MRI. The related symptoms were absent or present depending on the patients. The pattern of recurrence was classified into local, distant, or cerebrospinal fluid (CSF) dissemination. Local recurrence was defined as a recurrence at the primary tumor site, and distant recurrence was defined as spatially separated recurrence within the brain. CSF dissemination was defined as a recurrence in the ventricles, subarachnoid space, and spine.

Table 1 summarizes the patients' information. Of the six pediatric patients with ependymomas receiving treatment, one patient (16.6%) had a supratentorial tumor, and five patients (83.3%) had infratentorial tumors. Three patients (50.0%) had grade II ependymomas, and three patients (50.0%) had grade III. GTR was performed in two patients (33.3%) and STR in four (66.6%). Post-operative adjuvant therapy was administered to five patients. Radiation treatment was administered to four patients (66.6%), and one patient was treated with chemotherapy (16.6%). In the radiation treatment, three patients received only local brain radiation therapy consisting of 4,000, 4,500, and 5,400 cGy, respectively. One patient received local and craniospinal irradiation (CSI) during the same period. The radiation dose was 3,820 cGy in the local brain and 1,700 cGy in the CSI. Chemotherapy with vincristine, prednisolone, and carmustine was administered to one patient according to the CCG-9445 protocol of the Children's Cancer Group high-grade glioma study for 5 months. The mean follow-up duration was 48.5 months (range, 16.9–102.8 months). Tumor recurrence occurred in three patients (50%). Two patients showed local recurrence on follow-up at 32.1 months and 72.0 months, respectively. One patient showed local recurrence and CSF dissemination at 8.8 months.

A total of 27 adult patients with ependymoma were treated; 22 patients (81.4%) had WHO grade II ependymomas, and five (18.5%) had WHO grade III ependymomas. One patient with WHO grade II ependymoma demonstrated malignant transformation into WHO grade III ependymoma after recurrence. The tumor was located supratentorially in seven patients (25.9%), infratentorially in 13 (48.1%), and in the spinal cord in seven (25.9%). The tumor was totally resected in 15 patients (55.5%), and an STR was performed in 12 (44.4%). Postoperative adjuvant radiation therapy was administered to nine patients (33.3%). Eight patients received 4,500 cGy to 5,400 cGy local brain radiotherapy and one patient received 3,240 cGy CSI. The mean follow-up duration was 83.8 months (range, 3.8–182.2 months). Tumor recurrence occurred in seven patients (25.9%). Mean recurrence period of relapsed patients was 12.0 months (range, 1.6–46.6 months). Five patients developed local recurrence, and two patients developed local recurrence combined with CSF dissemination. Table 2 summarizes the information of relapsed adult patients.

We defined the mean range according to the follow-up duration and determined single-variable effects on PFS and OS via univariate analyses. The single variables were age, sex, tumor location, extent of removal, pathology, and adjuvant therapy. PFS was calculated from the dates of surgery to radiological progression or the last follow-up visit, and OS was calculated from the dates of surgery to death or final follow-up. Survival probability was calculated using the Kaplan-Meier method by performing comparisons with the log-rank test. All statistical analyses were performed with a significance level of p<0.05 using SPSS 21.0 (IBM Corp., Armonk, NY, USA).

Six patients were pediatric patients (mean age, 6.15 years; range, 1.3–11.0 years) with a male-to-female ratio of 2:4, and 27 were adults (mean age, 47.5 years; range, 19.0–70.0 years) with a male-to-female ratio of 13:14. The mean follow-up duration was 66.7 months (range, 3.9–144.1 months). In all patients, the mean PFS was 114.0±13.8 months and mean OS was 138.0±12.6 months. There was no statistical significance in the relationship between PFS and sex, tumor location, or pathology. Adult patients were found to be associated with a longer PFS (p=0.03) and OS (p=0.03) (Fig. 1A). Six pediatric patients had a mean PFS of 43.7±13.5 months and a mean OS of 58.1±13.7 months, while 27 adults had a mean PFS of 127.8±14.8 months and a mean OS of 151.2±12.5 months. Depending on the age and location, we divided the patients into four groups: pediatric-infratentorial, adult-supratentorial, adult-infratentorial, and adult-spinal cord. Pediatric-infratentorial group showed a trend of poor prognosis compared to adult-spinal cord group in terms of PFS (p=0.032) (Fig. 1B). PFS was 49.0±15.1 months in pediatric-infratentorial group, 85.5±26.3 months in adult-supratentorial group, 130.7±21.3 months in adult-infratentorial group, and 113.5±12.2 months in adult-spinal cord group.

Pediatric-infratentroial group was not statistical poor significance for OS compared adult-group (p=0.358). OS was 66.4±13.7 months in pediatric-infratentorial group, 129.3±27.4 months in adult-supratentorial group, 155.5±17.4 months in adult-infratentorial group, and 113.5±12.2 months in adultspinal cord group. There was no statistically significant relationship with sex, tumor location, extent of removal, or pathology. However, there was a statistically significant relationship with age.

These data were summarized in Table 3. Adult ependymomas were analyzed for clinical prognostic factors. For PFS, there was no statistically significant relationship with sex (p=0.64), tumor location (p=0.28), pathology (p=0.23), or postoperative adjuvant treatment (p=0.13). However, the extent of removal had a significant effect on PFS (p=0.04) (Fig. 1C); the mean PFS was 154.6±14.0 months for GTR, and 82.7±20.6 months for STR. The one-year PFS rate was 100% for GTR, and 75% for STR. For OS, there was no significant effect of sex (p=0.63), tumor location (p=0.78), pathology (p=0.20), or postoperative adjuvant treatment (p=0.31). A trend towards a better OS was found in patients who was underwent GTR (p=0.06); the mean OS was 172.5±9.3 months for GTR and 120.6±22.3 months for STR.

Twelve adult patients underwent total resection of the ependymoma without anaplastic pathology and adjuvant treatment. The mean age of the patients was 54.3 years (range, 20.0–68.07 years). The male-to-female ratio was 4:8. The tumor was located supratentorially in two patients (16.6%), infratentorially in four (33.3%), and in the spinal cord in six (50%). The mean follow-up duration was 93.5 months (range, 27.9–162.7 months). One out of the 12 patients developed local recurrence at the primary origin site.

The clinical course of the one case of recurrence is described (Fig. 2). A 36-year-old woman had a 3.7-cm-sized ependymoma in the fourth ventricle that was totally resected via a telovelar approach. One year later, follow-up MRI showed a 1-cm-sized recurrent lesion associated with enhancement of the fourth ventricle. Local brain radiotherapy (5,400 cGy) and spinal irradiation (2,160 cGy) were performed. Over the course of 92 months of follow-up after radiotherapy, no recurrence was seen.