1575 / 1708
last follow-up as compared to 49 of 73 (67%) patients who
received a higher dose (2). Stuben
et al.
reported a similar
dose-response; patients that received doses greater than 45
Gy had significantly improved progression free survival
rates (14). We did not observe a dose-response; therefore, a
total dose of 50.4 Gy in 28 fractions is reasonable. There are
series reporting favorable outcomes for patients with low-
grade (including myxopapillary) tumors of the lumbar spine
or cauda equina having undergone en bloc complete resec-
tion alone (15–19); therefore, the option of observation
following en bloc complete resection for patients with low-
grade spinal tumors is a reasonable one. In this setting, it is
advisable to obtain a post-operative MRI to verify the
absence of locally persistent tumor before withholding ra-
diotherapy (20).
In the present series, the survival rate of patients with
high-grade tumors was quite poor (27% at 5 years). Fortu-
nately, only a minority of patients present with high-grade
lesions. Potential methods of improving radiotherapy in-
clude modifications of field designs, doses, and fraction-
ation pattern. Regarding field designs, craniospinal axis
radiotherapy is reasonable for those with high-grade
ependymomas because of their high risk of leptomeningeal
failure (41% in our series). Although we were unable to
show that craniospinal irradiation provided a survival ad-
vantaged, the number of patients (10) with high-grade tu-
mors was too small to clearly examine this issue, and 7 of
the 10 did receive craniospinal irradiation. Vanuytsel
et al.
reported that craniospinal irradiation was associated with
better survival rates for patients with high-grade tumors
(11). Our finding that 9 of the 10 patients with high-grade
tumors developed local failure indicates that the radiation
doses used (24–55.2 Gy in patients with high-grade tumors,
median 43 Gy) were insufficient to achieve local control.
The dose of radiation administered to the primary site could
potentially be increased above conventional levels (to 59.4–
64.8 Gy) with the following RT techniques: 3-D treatment
planning utilizing small conformal beams, dynamic confor-
mal RT, or stereotactic RT. The use of systemic therapy can
be also considered in patients with high-grade tumors.
Bloom
et al.
found that the use of chemotherapy improved
survival rates in this setting (21). It is possible that combi-
nations of conventional therapy or possibly novel therapies
will result in more favorable outcomes.
Major limitations of this study are biases which may have
been introduced by the retrospective nature of the data
collection and the relatively small number of patients with
high-grade tumors. However, this is one of the larger studies
in the literature and it has long-term follow-up (median
follow-up was greater than 10 years).
In summary, ependymomas are uncommon gliomas af-
fecting all levels of the central nervous system. Low-grade
tumors are more common than high-grade lesions. The
present study indicates that postoperative radiotherapy of
low-grade lesions resulted in high survival rates. The 5-year
and 10-year actuarial survival rates were 87% and 79%,
respectively. Myxopapillary ependymomas occurred in the
lumbar spine region and were associated with an excellent
prognosis; the 5-year survival rate was 100%. Patients with
high-grade lesions had a poor prognosis due to a high risk of
both local failure and craniospinal axis seeding. The results
of aggressive therapy with resection and craniospinal irra-
diation were unfavorable in this group; therefore, new forms
of therapy will be required to improve the prognosis of
patients with high grade ependymomas.
Fig. 6. Overall survival by primary tumor location.
Fig. 7. Overall survival by histologic type of tumor.
Fig. 8. Overall survival of patients with myxopapillary ependymo-
mas (all considered grade 1) as compared to those with other tumor
types divided by grade.
957
Ependymomas
●
S. E. S
CHILD
,
et al.