SPINE Volume 34, Number 21, pp E756 –E760
©2009, Lippincott Williams & Wilkins
Long-Term Surgical Outcomes for Myxopapillary
Ependymomas of the Cauda Equina
Masaya Nakamura, MD, Ken Ishii, MD, Kota Watanabe, MD, Takashi Tsuji, MD,
Morio Matsumoto, MD, Yoshiaki Toyama, MD, and Kazuhiro Chiba, MD
Study Design. Retrospective case series.
Objective. To evaluate our treatment strategy for
myxopapillary ependymomas of the cauda equina.
Summary of Background Data. Some patients with
myxopapillary ependymoma develop cerebrospinal fluid
(CSF) dissemination leading to poor prognosis. Because
of the rarity of this tumor, there is no consensus on its
optimal treatment options and prognosis.
Methods. We reviewed 25 cases of myxopapillary
ependymomas, treated surgically between 1972 and
2005. The mean postoperative follow-up period was 10.4
years. The effects of surgical margins at surgery and post-
operative radiotherapy on tumor recurrence and progno-
sis were investigated.
Results. In 15 patients, total resection achieved (6 cases
of en bloc resection without postoperative radiation, and in
9 cases piecemeal resection) was followed by whole brain
and spinal cord radiation or local irradiation. Fourteen of
these patients survived without tumor recurrence. In 1 case
of total resection without radiotherapy, the tumor capsule
was violated intraoperatively and local recurrence occurred
2 years after surgery. In 4 patients, the tumors were removed
subtotally. Of these, 2 patients who received radiation (24 Gy)
only to the whole brain and spinal cord developed recur-
rence, and 2 who received whole brain and spinal cord
radiation (24 Gy) supplemented with local radiation (46 Gy)
developed no recurrence. The remaining 6 patients received
partial resection after local radiation (40 –50 Gy) alone, and
all 6 died of CSF dissemination.
Conclusion. The results of the present study indicate
that the surgical margin obtained at the initial surgery and
the extent and amount of postoperative radiation can be
crucial factors determining the prognosis of patients with
myxopapillary ependymoma. Although this tumor is his-
tologically benign, CSF dissemination can occur once tu-
mor capsule is violated, before or during surgery. Therefore,
early diagnosis is essential, and a therapeutic strategy in-
cluding radiotherapy, on the assumption that this tumor is
malignant, should be established.
Key words: myxopapillary ependymoma, cauda equine,
surgical outcome, radiation. Spine 2009;34:E756 –E760
From the Department of Orthopedic Surgery, School of Medicine, Keio
University, Shinjuku, Tokyo, Japan.
Acknowledgment date: December 12, 2008. First revision date: March
13, 2009. Acceptance date: March 13, 2009.
The manuscript submitted does not contain information about medical
device(s)/drug(s).
No funds were received in support of this work. No benefits in any
form have been or will be received from a commercial party related
directly or indirectly to the subject of this manuscript.
Supported by IRB at Keio University.
We obtained signed patient consent forms except from the 6 dead patients.
Address correspondence and reprint requests to Masaya Nakamura,
Department of Orthopedic Surgery, School of Medicine, Keio Univer-
sity, 35 Shinanomachi, Shinjuku, Tokyo, 160 – 8582, Japan; E-mail:
masa@sc.itc.keio.ac.jp
E756
Myxopapillary ependymoma, which frequently arises in
the terminal filum of the spinal cord, was first reported
by Kernohan as a subtype of ependymoma in 1932.1
Myxopapillary ependymoma is a grade 1 ependymoma
according to the World Health Organisation classifica-
tion and is considered to have a good prognosis. Despite
its benign histologic characteristics, some patients de-
velop cerebrospinal fluid (CSF) dissemination and have a
poor prognosis.2,3 Because of its rarity, there is no con-
sensus on the optimal treatment options and prognosis
of this tumor.4 –7 The present study was undertaken to
retrospectively review the long-term surgical results of
myxopapillary ependymoma, to identify an optimal
treatment strategy for this disease.
Materials and Methods
Twenty-five patients, 13 males and 12 females with myxo-
papillary ependymoma, who visited our hospital between
1972 and 2005, were included in the present study. All pa-
tients underwent primary surgery at our hospital and their
age at surgery ranged from 14 to 58 years (mean: 33 years).
In all cases, the tumor was surgically resected through a
posterior approach. In cases in which en bloc resection of the
tumor together with the capsule was impossible, radiother-
apy was administered after surgery. Although local radia-
tion alone was administered before 1985, whole-brain and
spinal cord radiation in addition to local radiotherapy was
used after 1985. The postoperative follow-up periods ranged
from 3.5 to 25.2 years (mean: 10.4 years). The effects of
surgical margins and postoperative radiotherapy on progno-
sis, including tumor recurrence, survival of patients, and
neurologic recovery, were analyzed retrospectively. The sur-
gical margin was determined from the surgical reports or
postoperative imaging studies. The surgical margin was clas-
sified as a total resection if the surgeon had described a
complete removal of the tumor or there was no evidence of a
residual tumor on postoperative magnetic resonance imag-
ing. The surgical margin was classified as a subtotal or par-
tial resection if the surgeon resected more than 90% or less
than 90% of the tumor respectively, which was confirmed by
follow-up imaging studies. Neurologic symptoms were eval-
uated according to the method reported by Schweitzer and
Batzdorf.4 The outcome was classified as “excellent” if the
patient was symptom-free or had minor symptoms not af-
fecting daily living; “good” if the patient was stable but with
a fixed neurologic deficit, such as sphincter dysfunction or
weakness; and “poor” if the patient had continuing active
disease or remain handicapped. Pearson correlation coeffi-
cient between the size of tumors and the duration of symp-
toms was analyzed by GraphPad Prism (GraphPad software,
California).
 Surgical Outcome of Myxopapillary Ependymomas • Nakamura et al E757

15 12

10
Patient’s number

(numbers of vertebra)
10

Size of tumors
6 R=0.543

4 Capsule rapture (+)

5
Capsule rapture (-)
2

0
0
0 10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70
Duration of symptoms (months)
Patient’s age
Figure 3. Correlation between tumor size and duration of symp-
Figure 1. Age distribution of patients at the time of surgery. toms. There was a positive correlation between the tumor size and
the duration of symptoms (R ⫽ 0.543).

Results
underwent a second operation (total resection) and re-
Clinical Data
ceived radiotherapy (24 Gy for the whole brain and spi-
The age at surgery ranged from early teen to the sixth
nal cord, and 46 Gy for the local site). To date, the
decade, but the third decade of life was the most common
patient has remained tumor-free, 12 years after the reop-
(11 cases) (Figure 1). For the tumor size, the mean num-
eration. No sign of recurrence has been detected in any of
ber of affected vertebrae was 4.2. In 9 cases, the tumor
the 5 patients in whom the tumor was removed without
was located cranial to the conus medullaris, and there
capsule violations.
were skip lesions in 5 cases (Figure 2). There was a pos-
The tumor was totally resected in a piecemeal fashion
itive correlation between the tumor size and the duration
using a cavitron’s ultrasonic surgical aspirator (CUSA) in
of symptoms (Pearson correlation coefficient ⫽ 0.543)
9 cases, in which the tumor was found to have extruded
(Figure 3). It is notable that capsule ruptures were ob-
from the capsule and adhered directly to the cauda
served in all cases in which the tumors extended over
equina. In these cases, postoperative radiotherapy (24
more than 3 vertebrae. Even in patients with a tumor
Gy for the whole brain and spinal cord, and 46 Gy for the
covering less than 2 vertebrae, there were 2 cases of cap-
local site) was supplemented, and none of these patients
sule rupture (Figure 3).
had recurrence to date.
Effects of Surgical Procedures and Postoperative The tumor was subtotally removed (over 90%) with
Radiotherapy on Tumor Recurrence CUSA in 4 patients because of strong adhesion of the
The details of the surgical procedures and radiotherapy tumor to the cauda equina. All 4 of these patients under-
are summarized in Table 1. En bloc total resection in- went postoperative radiotherapy (in 2 cases, 24 Gy for
cluding the tumor capsule was achieved in 6 cases. Post- the whole brain and spinal cord only, and in 2 cases, an
operative radiotherapy was not administered in these additional 46 Gy for the local site). One of the 2 patients
cases. In one of these cases, the capsule was partially receiving radiation only to the whole brain and spinal
violated during surgery, and local tumor recurrence was cord developed recurrence 2 years after surgery. This
found 2 years after the primary surgery. This patient patient underwent a second operation and has remained
recurrence-free to date, 8 years after the reoperation. In
Th6
7
8 Table 1. Surgical Procedures and Radiotherapy
9
10 Total Resection
(Level of the tumors)

11 Sub-Total Partial
12 En bloc Piece by Piece Resection Resection
L1
2 No. case 6 9 4 6*
3 Radiation
4 Local 0 9 2 6
5 Whole brain and 0 9 4 0
S1 spinal cord
2
Local recurrence 1 0 2 6
Dissemination 0 0 0 6
3
*Cases before 1985.
Figure 2. Locations of tumors.
E758 Spine • Volume 34 • Number 21 • 2009
1 total tumor resection. Four cases of a piecemeal total
.8
Survival rate
.6
.4
.2
0 Representative magnetic resonance (MR) images of a tu-
0 50 100 150 200
(Months)
Figure 4. Cumulative survival curve.
the second patient receiving radiation for the brain and
spinal cord, a recurrence was detected 3 years after sur-
gery, and the patient is under observation. The 2 patients
receiving the additional 46 Gy local site radiation have
no signs of tumor recurrence.
In 6 patients who underwent surgery before 1985, the
tumor removal ended up in partial resection (⬍90%)
because CUSA was not available until 1985. These 6
patients underwent postoperative local radiotherapy
alone (40 –50 Gy); however, all 6 had recurrences. In 2
patients, reoperation was not possible. The remaining 4
patients underwent reoperation 3, 4, 6, and 16 years
after the initial surgery. All 4 patients, however, devel-
oped another recurrence 6 months to 3 years after the
reoperation. A third operation was not possible in 2 pa-
tients, but was performed in the other 2 patients, al-
though only partial resection was possible in either of
them.
Six of the 25 patients died of CSF dissemination, making
the 5-year survival rate 80% (Figure 4). A more detailed
analysis of the survival revealed that there were 16 cases of
continuous disease free, 2 of no evidence of disease, 1 alive
with disease, and 6 dead of disease. All of the patients who
died of CSF dissemination were those who had received
local radiotherapy after partial tumor resection.
Functional Outcome
Functional assessments on 19 survival cases revealed that
10 cases (40%) were rated as excellent, 5 (20%) as good,
and 4 (16%) as poor (Table 2). Analysis of the relation-
ship between the functional outcome and the surgical
Table 2. Functional Outcomes
Total Resection
Sub-Total
En bloc Piece by Piece Resection
Excellent 6 4 0 0
Good 0 4 1 0
Poor 0 1 3 0
Death 0 0 0 6
dead
procedure revealed that all 10 cases rated excellent had a
alive
resection and one of a subtotal resection were rated as
good, because all of these patients had residual dysuria.
Four cases rated as poor included one case of piecemeal
total resection and 3 of subtotal resection. Thus, the
functional outcome of myxopapillary ependymoma de-
pended on the surgical margin.
Magnetic Resonance Imaging and
Histologic Examination
mor without a capsule rupture revealed a well-
250 300
circumscribed mass, which was hypointense on T1-
weighted and hyperintense on T2-weighted images with a
homogenous enhancement by gadolinium (Figure 5). It is
noteworthy that most neurinomas at the cauda equina
show a heterogeneous enhancement by gadolinium,
which is an important difference for the differential diag-
nosis. There was no difference in the histologic charac-
teristics among the tumors of different sizes. All of the
tumors showed typical features of myxopapillary
ependymoma, that is, marked mucous changes of the
stroma and papillary arrangement of tumor cells (Figure
6). The MIB-1 indexes of all tumors were below 2%, and
none showed findings of malignancy, such as nuclear
atypism or a high frequency of nuclear division.
Discussion
Time Course of Tumor Growth
In our cases, the age at surgery ranged from the early
teens to the 60s, but the majority were in the third decade,
consistent with the previous reports.4 –7 There was a posi-
tive correlation between the tumor size and duration of
illness, and capsule rupture was found in all of those ex-
tending over 3 vertebral levels, suggesting that these tumors
grow over time and eventually penetrate the capsule. In all
cases, the localization of the tumors included the L2–L4
levels, suggesting that these tumors initially develop at
this level, growing slowly in the craniocaudal direction
over time, and that once the tumor perforates the cap-
sule, it spreads along the cauda equina and occupies the
dural canal, eventually leading to CSF dissemination.2,3
There was no difference in histologic features among the
tumors of different sizes, and all tumors showed a low
growth rate and absence of malignant characteristics.
According to these findings, early diagnosis before the
perforation of the tumor capsule is an important deter-
minant for good prognosis.
Effect of Surgical Margin and Radiotherapy
on Prognosis
Regarding the prognosis of patients with myxopapillary
Partial ependymoma, Sonneland et al stated that completeness
Resection Total
of tumor resection is the major factor determining the
10 likelihood of postoperative tumor recurrence.5 In the
5 present study, long-term survival without tumor recur-
4 rence was obtained in all 5 patients in whom complete
6
resection without capsule violation was achieved. In 1
 Surgical Outcome of Myxopapillary Ependymomas • Nakamura et al E759
Figure 5. Representative MR im-
ages of a tumor without a cap-
sule rupture revealed a well-
circumscribed mass, which was
hypointense on T1-weighted (A)
and hyperintense on T2-
weighted (B) images with a ho-
mogenous enhancement by gad-
olinium (C).
case, the capsule was violated intraoperatively however,
because the tumor was removed completely, no radio-
therapy was administered, and local recurrence occurred
2 years after the surgery. This indicates that even after
complete tumor resection together with the capsule, if
the capsule is violated, recurrence can occur, and post-
operative radiotherapy, therefore, should be considered
in such cases.
Radiotherapy has been considered to be effective
against this tumor.8 Chan et al9 reported that postoper-
ative intracranial metastasis was seen in 3 of 5 patients
who did not receive radiotherapy even after complete
tumor resection, and that radiation to the entire brain
Figure 6. Histology (HE staining). All of the tumors showed typical
features of myxopapillary ependymoma, that is, marked mucous
changes of the stroma and papillary arrangement of tumor cells.
and spinal cord (including local radiation; total dose
30 –50 Gy) prevented further tumor progression. In the
present study, recurrence was seen in 2 of 4 patients who
underwent radiation to the whole brain and spinal cord
after subtotal tumor resections. The recurrence in these 2
patients seems to be attributable to an inadequate
amount of local radiation (24 Gy). Our experience sup-
ports the view proposed by Chan et al that 30 to 50 Gy of
local radiation is necessary in cases in which total tumor
resection could not be achieved.
In our study, most patients who were managed before
the introduction of MRI and CUSA had a tumor affecting
more than 5 vertebrae, and all of them died of CSF dis-
semination, because the tumor resection was incomplete
and only local radiation (40 –50 Gy) was administered
after surgery. The results of this study indicate that it is
ideal to remove the tumor en bloc, without violating the
capsule. However, in cases of large tumors in which the
capsule is perforated due to delayed diagnosis, it is advis-
able to resect the tumor as completely as possible using
CUSA and to administer postoperative radiation to the
whole brain and spinal cord (24 Gy) in addition to a suffi-
cient amount of radiation at the local site (40 –50 Gy).
Although functional outcomes were favorable in the
cases of total resection, residual dysuria was often seen in
the cases of total resection but with a capsule rupture.
These findings indicate that surgical treatment of this
tumor before the rupture of the capsule is important
from the viewpoint of functional prognosis. To this end,
early diagnosis and early surgery are essential.
The present results indicate that the extent of tumor
spread and the completeness of tumor resection at the
initial surgery, rather than the histologic grade of the
tumor, are crucial factors determining the prognosis of
patients with myxopapillary ependymoma. Although
this tumor is histologically benign, CSF dissemination
can occur once the capsule is violated, before or during
surgery.10,11 Therefore, a therapeutic strategy including
E760 Spine • Volume 34 • Number 21 • 2009
radiotherapy, on the assumption that this tumor is ma-
lignant, should to be established.
Key Points
● The surgical margin obtained at the initial surgery
and the extent and amount of postoperative radiation
can be crucial factors determining the prognosis of
patients with myxopapillary ependymoma.
● Although myxopapillary ependymoma is histo-
logically benign, CSF dissemination can occur once
the capsule is violated, before or during surgery.
● Myxopapillary ependymoma should be treated
assuming that this tumor is malignant.
References
1. Kernohan JW. Primary tumors of the spinal cord and intradural filum ter-
minale. In: Penfield W, ed. Cytology and Cellular Pathology of the Nervous
System. Vol. 3. New York, NY: Paul B Hoeber; 1932:993–1025.
2. Davis C, Barnard RO, Path MR. Malignant behavior of myxopapillary
ependymoma. J Neurosurg 1985;62:925–9.
3. Plans G, Brell M, Cabiol J, et al. Intracranial retrograde dissemination in
filum terminale myxopapillary ependymoma. Acta Neurochir (Wien) 2006;
148:343– 6.
4. Schweitzer JS, Batzdorf U. Ependymoma of the cauda equina region: diag-
nosis, treatment, and outcome in 15 patients. Neurosurgery 1992;30:202–7.
5. Sonneland PR, Scheithauer BW, Onofrio BM. Myxopapillary ependymoma.
Cancer 1985;56:883–93.
6. Lonjon M, Von Langsdorf D, Lefloch S, et al. Factors influencing recurrence
and role of radiotherapy in filum terminale ependymomas: 14 cases and
review of the literature. Neurochirurgie 2001;47:423–9.
7. Celli P, Cervoni L, Cantore G. Ependymoma of the film terminale. Acuta
Neurochir (Wien) 1993;124:99 –103.
8. Shaw EG, Evans RG, Scheithauer BW, et al. Radiotherapeutic management
of adult intraspinal ependymomas. Int J Radiat Oncol Biol Phys 1986;12:
323–7.
9. Chan HS, Becker LE, Hoffman HJ, et al. Myxopapillary ependymoma of the
filum terminale and cauda equina in childhood: report of seven cases and
review of the literature. Neurosurgery 1984;14:204 –10.
10. Schiffer D, Chio A, Giordana MT, et al. Histologic prognostic factors in
ependymoma. Childs Nerv Syst 1991;7:177– 82.
11. Fassett D, Pingree J, Kestle AW. The high incidence of tumor dissemination
in myxopapillary ependymoma in pediatric patients. J Neurosurg 2005;102:
59 – 64.