Introduction

Spinal myxopapillary ependymoma (SME) refers to a rare pathological variant of ependymoma, which is predominantly located in the cauda equina and filum terminale regions of the spinal cord. Although SME is generally considered as a benign entity with a slow-growing nature, histologically corresponding to WHO grade I, local recurrence, metastasis, and even cerebrospinal fluid (CSF) dissemination have been reported in many cases [9, 15, 22]. According to previous reports, the overall recurrence rate of SME ranged from 7 to 37% [14]. The optimal therapeutic strategy for SMEs remains controversial. The surgical outcomes have not been well established, and some scholars recommend adjuvant radiotherapy and/or chemotherapy [3, 4, 21]. Nakamura et al. proposed that radiation on the whole brain and spine may be helpful to prevent the local recurrence and CSF dissemination, when the tumor capsule is ruptured intraoperatively [21]. Moreover, sparse case reports demonstrated that temozolomide had a certain efficacy for controlling the recurrent SME [7, 9]. In literatures, studies regarding SMEs were limited to single-case reports or small case series. The purpose of this large-scale, single-center study was to investigate the clinicoradiological characteristics and surgical outcomes of SMEs.

Materials and methods

Patients

This retrospective study enrolled a total of 34 consecutive patients diagnosed with SME between Jan 2007 and June 2016. SME was diagnosed based on pathological evidence, and the histopathological slides were reassessed by two neuropathologists independently. Individual demographic and clinical data were collected, including sex, age at diagnosis, onset symptoms, duration of symptoms, and physical signs.

The study was approved by the Institutional Review Board and Ethics Committee of Beijing Tiantan Hospital.

Radiological evaluation

Magnetic resonance imaging (MRI) with gadolinium-contrast enhancement was performed as the standard radiological examination pre- and postoperatively in all cases. The radiological features, including tumor location (spinal segment), size (the maximum diameter), cystic or solid appearance, signal intensity, tumor margins, and contrast enhancement, were analyzed independently by two neurosurgeons.

Treatment

All patients underwent laminotomy and microsurgical removal via the posterior median approach, with intraoperative neurophysiologic monitoring of somatosensory- and motor-evoked potentials. No patients received preoperative radiotherapy. Gross total resection (GTR) was defined as gross macroscopical removal of the tumor and no evidence of a residual tumor on postoperative MRI. Subtotal resection (STR) was defined as resection of the majority of the tumor with a small residual portion. Postoperatively, none of the patients received adjuvant chemotherapy. Among the patients undergoing STR, 8 of them received postoperative adjuvant radiotherapy which was delivered for 35 days with daily fractions of 115–150 cGy (total dose range, 40–52 Gy). Radiation fields included the tumor plus one to two vertebral bodies above and below the tumor.

Prognostic assessment

Follow-up data for all patients were obtained during individual office visits, with a mean follow-up time of 41.5 months (standard deviation = 19.9, range 21–85). The patients’ postoperative status was investigated by physical examination and gadolinium-enhanced MRI, and McCormick classification was applied to assess neurological function [1, 19]. Radiological and functional assessments were performed preoperatively, at discharge, at 3 months postoperatively, and annually thereafter.

Statistical analysis

SPSS 24.0 software (IBM Corp., Armonk, NY, USA) was used for statistical analyses. Fisher’s exact test was used for the comparison of categorical variables. Probability (p) values ≤ 0.05 were considered significant.

Results

Clinical features

There were 21 males and 13 females, yielding a male-to-female ratio of 1.6:1. The average age at diagnosis was 29.97 ± 12.46 years (range 8–59 years); there were 9 pediatric patients (26.5%) and 25 adult patients (73.5%). The clinical onset symptoms included pain (28/34 cases; 82.4%), weakness (15/34; 44.1%) and numbness (7/34; 20.6%) in extremities, and sphincter dysfunction (9/34; 26.5%). The duration of symptoms preceding the initial diagnosis ranged from 1 month to 6 years (mean, 20.7 ± 16.5 months). The demographic and clinical profiles were summarized in Table 1.

Table 1 Demographic and clinical characteristics
Full size table

Radiological characteristics

The tumor locations included lumbar segments (18/34 cases; 52.9%), thoracolumbar segments (8/34; 23.5%), lumbosacral segments (6/34; 17.6%), and thoraco-lumbo-sacral segments (2/34; 5.9%). On MRI, all SMEs appeared heterogeneously hypointense on T1-weighted imaging and heterogeneously hyperintense on T2-weighted imaging. After Gd-diethylenetriaminepentaacetic acid (Gd-DTPA) administration, all tumors showed heterogeneously remarkable enhancement. All of these tumors were solid and well-demarcated, and no remarkable cystic component was noted. Rostral or caudal syringomyelia was noted in 20 cases (58.8%). The preoperative suspected diagnoses included spinal ependymomas, dermoid cysts, schwannomas, and teratomas. The radiological images of representative cases were presented in Fig. 1.

Fig. 1
figure 1

Magnetic resonance images of a representative case with SME. Magnetic resonance imaging revealed an intraspinal lesion (arrow) at the L2 level, which was heterogeneously hyperintense on T2-weighted imaging (a) and heterogeneously enhanced on T1-weighted contrast-enhanced imaging (b). Postoperative imaging showed the lesion was completely removed (c)

Full size image

Intraoperative findings and surgical outcomes

According to the operative documents, no evoked potentials were lost during the surgery. The SMEs were grayish-red in color and soft in texture, with a rich blood supply. The tumors were encapsulated, with close adhesion to the surrounding cauda equina and/or filum terminale. Macroscopically, the SMEs showed a sausage-like appearance (Fig. 2). GTR was achieved in 18 patients (52.9%), and STR was achieved in 16 patients (47.1%).

Fig. 2
figure 2

Intraoperative findings. Intraoperatively, we found the SME was grayish-red with complete capsule, and it showed a sausage-like appearance

Full size image

Postoperatively, all the patients were uneventful. After an average follow-up period of 41.5 months, radiological recurrence was noted on MRI in 5 of the 8 patients (62.5%) who underwent STR but no adjuvant radiotherapy. The age of patients with recurrence was 8, 17, 18, 18, and 29 years, respectively. The time interval between operation and recurrence was 37, 41, 50, 62, and 71 months, respectively. The sizes of recurrent tumors were all less than 2 cm. There was no tumor progression in the patients who underwent GTR (n = 18) or combined STR and radiotherapy (n = 8) (Fig. 3a). Fisher’s exact test showed incomplete resection (STR) was significantly associated with tumor recurrence (p < 0.05).

Fig. 3
figure 3

Follow-up data. a Eighteen patients underwent gross total resection (GTR) and none of them received postoperative radiotherapy (RT); among these patients, no tumor recurrence was noted during the follow-up period. Sixteen patients underwent subtotal resection (STR) and 8 of them received postoperative RT; during the follow-up period, tumor recurrence was noted in 5 of the 8 patients who underwent STR but no RT. b Preoperatively, 12 patients were at grade I of the McCormick classification, 20 at grade II, and 2 at grade III; during the follow-up period, 31 patients were at grade I, 2 at grade II, and 1 at grade III

Full size image

Regarding the neurological function, preoperatively, 12 patients were at grade I of the McCormick classification (neurologically normal or mild deficits), 20 at grade II, and 2 at grade III (grades II and III represent moderate to severe neurological deficits); during the follow-up period, 31 patients were at grade I, 2 at grade II, and 1 at grade III (Fig. 3b). Fisher’s exact test showed surgical treatment can significantly improve the neurological function of patients with SMEs (p < 0.05).

Discussion

Ependymomas are tumors of glial origin arising from the ependymocytes, which line the CSF-filled ventricles in the brain and the central canal of the spinal cord. Ependymomas are the most common primary tumors of the spinal cord, occupying approximately 60% of all spinal neuroepithelial neoplasms [15]. Pathologically, spinal ependymomas can be divided into four subtypes: conventional ependymomas (cellular, papillary, clear cell, and tanycytic), subependymomas, myxopapillary ependymomas, and anaplastic ependymomas. According to the 2007 and 2016 CNS WHO classifications, myxopapillary ependymomas are classified as grade I low-grade tumors [17, 25]. SME was firstly designated as a distinct clinicopathological entity by Kernohan in 1932 [13]. As reported, SMEs were estimated to account for about 27% of all spinal ependymomas [23].

In literatures, SMEs were found most commonly in the fourth decade of life with a male preference, and 8–20% of them occurred in pediatric patients [10, 23]. In the current study, our demographic results were consistent with previous findings. Due to its slow-growing nature, SME usually has a long clinical course [5]. The average symptom duration of ependymomas was reported to be about 1 year [27]; in our study, the mean symptom duration of SMEs was much longer (20.7 months). Clinical manifestations of SMEs are nonspecific, including local pain, sensorimotor disturbance, and sphincter dysfunction. As SMEs most frequently occur in the cauda equina and filum terminale regions of the spinal cord, hypoesthesia in the saddle area (perineum and anus) and bladder-rectal dysfunction as well as sexual dysfunctions are quite common. Once the nerve root is affected, the patients can manifest as intractable lumbosacral or lower-extremity pain. In the late stage of the disease, the patients can develop flaccid paralysis [27]. In the current series, pain is the most common onset symptom, which may be related to the involvement of surrounding nerves [24].

In fact, almost all the previously described SMEs as well as SMEs in our study were located in the lumbar spinal canal, and some scholars hypothesized that these entities arise from the lumbar thecal sac in proximity to the conus medullaris, cauda equina, and filum terminale [15, 20]; nevertheless, some intracerebral counterparts have also been reported [16, 18, 26, 28]. Considering myxopapillary ependymomas have a potential for metastasis and CSF dissemination, the diagnosis of primary intracerebral myxopapillary ependymomas should be differentiated from metastatic or disseminated tumors. Noteworthily, on MRI, SMEs are well-demarcated unlike conventional ependymomas. This radiological feature was consistent with macroscopical findings that SMEs were encapsulated with a sausage-like appearance.

Currently, surgical resection is the mainstream modality for the treatment of SMEs. The goal of surgery is decompression of the spinal cord and maximal safe resection of the tumor. In the previous literature, few studies analyzed the functional outcomes after surgical resection. Herein, we found surgical treatment could significantly improve the neurological function of patients with SMEs. Therefore, timely surgical decompression should be highlighted for the functional rehabilitation. With the development of neurophysiological monitoring and microneurosurgical techniques, the rate of total resection has been significantly increased with a lower incidence of complications. However, due to the close adhesion to surrounding nerves, complete removal of SMEs may still be challenging. In our study, only 52.9% of the patients gained GTR. Previous studies indicated that the extent of resection in the initial surgery is a prognostic factor for spinal cord ependymomas [12]. Feldman et al. performed a systematic review which enrolled 28 articles describing 475 patients, and they found that the overall recurrence rate after GTR (15.5%) was significantly lower than that after STR (32.6%); additionally, they also noted significantly higher recurrence rates in pediatric patients compared with adults (40.5% vs 23.4%) [8]. Consistently, in our study, all the five patients with recurrence were treated by STR, and four of them were pediatric. The role of radiotherapy for treating SMEs remains controversial. Feldman’s systematic review found that adjuvant radiotherapy was not associated with a decrease in recurrence rates [8], while the experience from M.D. Anderson Cancer Center showed that the addition of postoperative radiotherapy to surgery was associated with significantly higher 10-year progression-free survival rates (75% for surgery and postoperative radiotherapy vs. 37% for surgery alone) and 10-year local tumor control rates (86% for surgery and postoperative radiotherapy vs. 46% for surgery alone) [2]. Our results also support that tumor recurrences may occur late (> 10 years) with adjuvant radiotherapy. The optimal dose of radiation for SMEs is under debate as well. In previous studies, a dose of 40–50 Gy has been widely used [6, 29], and a dose-response effect has been reported by Garcia [11]. Akyurek et al. proposed that doses greater than 40 Gy had significant efficacy in improving progression-free survival rates [2]. In our study, there was no tumor progression in the patients who underwent STR with postoperative radiotherapy (40–52 Gy). The definitive role of radiotherapy still needs further research. However, considering the potential risk of recurrence and metastasis, radiotherapy should be recommended for patients who undergo incomplete resection.

Conclusions

SMEs have some characteristic features on MRI which can facilitate the preoperative diagnosis. Surgical resection can significantly improve the clinical symptoms. Complete surgical resection is the best treatment of choice with a favorable outcome. In cases of incomplete resection, postoperative radiotherapy should be highlighted.