1568 / 1708
Discussion
In our study, patients treated for histologically proven
MPE had excellent long-term survival rates: the
10-year OS rate was 97%, and the 10-year PFS and LC
rates were 62% and 72%, respectively. Most recur-
rences occurred within the primary site of disease. No
recurrences occurred outside the neural axis. Most
importantly, our data support the notion that patients
who initially receive adjuvant radiotherapy to maxi-
mize local control have improved PFS rates.
Our results compare favorably with those of other
series in the literature that have focused on spinal
ependymomas, including MPEs. For example, 10-year
survival rates of 75% to 100% have been reported by
other authors [9, 14, 16, 18, 25–27] (Table 3).
In the current study, no significant effect of the
different initial surgical extents on OS or PFS was
found. However, in patients who had either GTR or
STR, adjuvant radiotherpy showed better LC rates
(90% for both) at 10 years, compared with GTR or
STR alone (58% and 0%, respectively). This result was
comparable to that reported by Waldron et al. [18],
who likewise observed no influence of the extent of
resection. These authors also showed no recurrences in
11 patients with low-grade tumors treated with GTR
and adjuvant RT; similarly, in our study, only 1 (10%)
of 10 patients who had GTR plus adjuvant RT had
disease recurrence. Furthermore, Shaw et al. [25]
observed recurrences in three (44%) of seven patients
treated with GTR and postoperative RT. In a Mayo
Clinic series on MPEs, only 7 (16%) of the 45 patients
who underwent GTR had recurrences [14]. Several
authors have demonstrated the importance of the
extent of surgical resection in determining recurrence
patterns and do not support the use of adjuvant RT
after GTR in spinal ependymoma [20, 23, 28, 29].
In our experience, MPEs can be technically tricky to
resect completely because of the anatomic complexity
of the cauda equina. Intraoperative ultrasonography
can be quite helpful in disclosing occult foci hiding
behind the cauda equina [19]. The main asset to a safe
and effective operation in this disease, however, re-
mains an experienced, persistent, technically skillful
surgeon [30]. Recently, the use of monitoring tech-
niques, such as the recording of somatosensory evoked
potentials, and the earlier diagnosis made possible by
CT and MRI have made tumors (including MPEs)
Table 2
Patients who failed the primary treatment
Patient Age
(years)
Tumor
location
Primary
treatment
Site of
recurrence
Time to
recurrence
(months)
Salvage
therapy/outcome
(survival duration after salvage)
1
22
L1–4
STR
a
and local RT
Distant
78
Surgery, CSI, and CHT/Dead-unknown
(27 years)
2
14
T6–L4
GTR
and local RT
Local
15
Surgery/NED (10 years)
3
27
L1–3
STR
and local RT
Distant
11
Surgery, whole spine RT,
and CHT/AWD (21 years)
4
31
L3–4
STR
Local
20
Surgery and local RT/NED (19 years)
5
20
L1–3
Bx
and local RT
Local
40
Surgery/NED (5 years)
6
16
L2
GTR
Local
5
Surgery and RT/NED (10 years)
7
28
L2
STR
Local
55
Surgery and RT/NED (10 years)
8
19
T5–7/T12–L1 GTR
Distant and local
11
CSI and CHT/AWD (11 years)
9
35
T12–L3/L5–S1 GTR
Distant
8
CSI/NED (7 years)
10
33
S1–5
GTR
Local
70
RT/NED (6 years)
11
60
T12–L4
STR and CSI
Local
5
CHT/DOD (2.5 years)
12
26
T9–L2
GTR
Local
23
RT/AWD (4 years)
a
STR = Subtotal resection; RT = Radiotherapy; CSI = Craniospinal irradiation; CHT = Chemotherapy; GTR = Gross total resection;
NED = No evidence of disease; AWD = Alive with disease; BX = Biopsy; DOD = Dead of disease
0
2
4
6
8
10
12
14
16
18
Pre-radiotherapy
Last Follow-up
Number of Patients
B
C
D
E
Fig. 4
Neurologic function according to Frankel classification
J Neurooncol (2006)
123