Dr. Paulino and his coauthors present a thoughtful, comprehensive
survey of the literature regarding the evolving management of
intracranial germinoma. They review promising areas of clinical
research and acknowledge an increasing role for chemotherapy in the
initial management of this tumor.
They also refute the old wisdom that employed a test-screening dose
of radiotherapy to distinguish germinoma from other pineal region
tumors in order to spare patients the morbidity of surgery. The firm
insistence by these radiation oncologists that every patient with a
suprasellar or pineal region tumor is best managed with information
derived from a generous biopsy is reassuring. I wish to share some
additional reflections on how to advance clinical research in this area.
Factors That Promote Therapeutic Conservatism
1. The previous morbidity and mortality associated with operating
on patients with pineal region tumors were excessive and deterred the
traditional application of histology-based treatment planning.
These concerns are no longer valid. The surgical routes to the pineal
(transtentorial or supracerebellar) or posterior third ventricular
regions (transcallosal) are safely executed in most pediatric
neurosurgical centers. The use of neuronavigation systems guided by
magnetic resonance imaging (MRI) and the surgical microscope permits
the preoperative identification and intraoperative avoidance of
critical vascular and neural anatomy. Endoscopically derived biopsies
may be successfully obtained in patients with tumors that extend into
the posterior third ventricle, with the added advantage that the
surgeon can perform a third ventriculostomy during the same
procedure, thereby avoiding the need for a systemic shunt.
2. The disease is readily curable with radiotherapy.
Because the prognosis is so favorable with high doses and volumes of
radiotherapy, clinicians have been reluctant to alter the
conventional approach, which includes the ventricular field with or
without the craniospinal field for isolated pineal or suprasellar
primary tumors and craniospinal irradiation for disseminated disease
The authors of this review, in fact, concur that germinomas are best
treated in this fashion. This experience was based on an older
clinical experience in which patients developed recurrences outside
an involved field following either an inadequate or technically
suboptimal staging procedure that most likely underestimated regional
and metastatic involvement and included patients with tumors other
than pure germinomas.
3. Intracranial germinomas are uncommon.
Intracranial germ-cell tumors constitute less than 1% to 2% of
primary central nervous system (CNS) tumors in children, and < 60%
of germ-cell tumors are pure germinomas. Since the incidence of
germinomas peaks in the second and third decades of life, adult
neuro-oncology teams will manage 40% of patients. Not until recently
have multi-institution collaborations emerged, and these have
involved primarily pediatric consortia in the United States and
Europe and one cooperative group (the Pediatric Oncology Group
[POG]). To date, the only feasible protocol design has been a
single-arm, response-oriented, phase II study.
Management Issues for Future Study
1. Will MRI technology permit earlier diagnosis of patients with
smaller primary tumors of lower stage?
The endocrine manifestations of suprasellar germ-cell tumors may be
subtle and not easily recognized. Even when patients present with
diabetes insipidus, computed tomographic (CT) scans of the head may
be normal for several years before a contrast-enhancing suprasellar
mass is evident. Magnetic resonance imaging, which includes thin
sections through the suprasellar region, identifies pathologic
enhancement better than CT, especially in the region of the infundibulum.
Since most germinomas will eventually disseminate, either within the
ventricles or subarachnoid space, the earlier the diagnosis, the
lower is the incidence of dissemination. Early diagnosis alone will
spare patients significant treatment-related morbidity.
2. What is the highest cerebrospinal (CSF) level of beta-human
chorionic gonadotropin (beta-HCG) consistent with a favorable-risk germinoma?
Presently, we triage patients with primary CNS germ-cell tumors into
two categories based on histology and serum/CSF tumor markers
(alpha-fetoprotein [AFP]/beta-HCG). The high-risk group consists of
patients with a nongerminoma tumor component, any significant
elevation of AFP, or a modestly elevated CSF level of
HCG. A safe HCG level in the CSF is probably below 50 mIU/mL, but
neither this value nor any higher level has been systematically
studied with regard to prognosis in patients with histologically
3. What are the late effects of high-dose and large-volume brain
irradiation in long-term germinoma survivors?
The majority of germinoma patients will survive their brain tumors
and remain at continued risk for the progressive cognitive effects of
radiotherapy. There are very few long-term follow-up data on which to
assess the quality of these patients survival. Most of the
cognitive predictions are extrapolated from studies of other groups
of patients with favorable outcomes, such as those with
medulloblastomas. Such patients tend to be younger at diagnosis than
germinoma patients and to receive lower doses of cerebral irradiation
The risks of irradiating an adolescent and young adult population are
often minimized, but it is well recognized that older patients can
acquire similar cognitive deficits. Since many patients with
germinomas, even those with apparent pineal primary tumors, present
with endocrine deficiencies, it is difficult to assess the late
effects of radiation therapy alone on endocrine function.
4. What assumptions can be made regarding the similarity of
seminoma and dysgerminoma to their CNS counterpart, the germinoma?
If one could prove that the molecular biology and treatment
sensitivities of the systemic and CNS tumors are identical, the
clinical lore accrued in the management of a much larger group of
patients with systemic germ-cell tumors could be more readily transferred.
We already know that germinomas resemble of seminomas with respect to
their high radiation sensitivity and chemosensitivity, but we still
employ much higher doses of radiotherapy for germinomas than we do
for seminomas and dysgerminomas, and multimodality therapy for CNS
tumors is still in the experimental stages. Can germinoma be cured
with radiation doses in the range of 30 Gy? What is the smallest
radiation volume that will provide a high cure rate? Can we identify
on a molecular basis which germinomas can be cured with chemotherapy
alone? If so, what clinical and molecular factors are most reliable
in predicting responsivity to treatment?
5. What is the optimal role of chemotherapyto facilitate the
use of a reduced radiotherapy dose or to replace radiotherapy altogether?
Chemotherapy is likely to be highly effective in patients with
germinoma for several reasons. Patients with disseminated seminomas
can be cured with chemotherapy alone. Germinomas tend to arise in
CNS locations devoid of the blood-brain barrier, such as the pineal
region and infundibulum, so that the pharmacokinetics of chemotherapy
within the CNS should be similar to that in a systemic organ.
Pilot studies in patients with newly diagnosed or recurrent germinoma
have confirmed the high response rates to both single chemotherapy
agents and combination regimens. Special precautions need to be taken
when cisplatin (Platinol) and cyclophosphamide (Cytoxan, Neosar) are
administered to patients with diabetes insipidus, however. Although
most germinoma patients can be cured with either larger doses of
radiotherapy or chemotherapy alone, the morbidity or mortality will
be greater. I favor a treatment regimen that combines radiotherapy
and chemotherapy, but in doses that minimize acute and late effects
and produce cure rates equivalent to or better than historical expectations.
In pilot studies conducted in patients with localized disease, a
complete response to neoadjuvant, single-agent chemotherapy, such as
cyclophosphamide or carboplatin (Paraplatin), has permitted
reductions in the dose and volume of radiation therapy in over 70% of
patients without compromising survival.[5,6] The ventricles were not
routinely irradiated in these pilot studies, and the assumption is
that chemotherapy may have eliminated micrometastases within the
ventricles and subarachnoid space. It is also theoretically possible
that chemotherapy may reduce systemic micrometastases in patients
with ventricular-peritoneal (VP) shunts.
6. What type of consortium can best conduct prospective clinical trials?
Clearly, the rarity of germinomas is the rate-limiting factor in the
pace of clinical research. In the past, the cooperative groups have
been reluctant to engage in a phase III trial that would require more
than 5 years of accrual and another 5 years of surveillance before a
survival advantage could be confirmed. Prospective clinical trials
using either multimodality or single-modality (radiotherapy or
chemotherapy) approaches require either the development of a loosely
organized consortium or an institution with an unusually large volume
of patients, such as may exist in Japan.
A theoretically compelling phase III study that could be undertaken
by a multicooperative group consortium would be a comparison of
high-dose/volume radiation therapy alone vs reduced-dose/volume
radiation therapy and chemotherapy. The end points in this study
would be quality-of-life parameters, since the 5-year survival rate
in both arms is likely to exceed 90%. This concept is not likely to
be widely accepted because quality-of-life measures are
not particularly quantitative, and the pediatric cooperative groups
are just learning to use them. The most productive clinical trials
pursuit is to combine the expertise and accrual potential of the
Childrens Cancer Group (CCG) and POG or a similar
multi-institution consortium, as has been done successfully in Europe.
At present, phase II pilot studies are still the most feasible. A
logical next step could be to further escalate the intensity of
chemotherapy and reduce the dose and volume of radiotherapy. Outcome
measures should include quality of life and cognitive function, as
well as response and survival.
1. Sawamura Y, Ikeda J, Shirato H, et al: Germ cell tumors of the
CNS: Treatment considerations based on 111 cases and their long-term
clinical outcomes. Eur J Cancer 34:104-110, 1998.
2. Armstrong C, Mollman J, Corn B, et al: Effects of radiation
therapy on adult brain behavior. Neurology 43:1961-1965, 1993.
3. Bosl G, Motzer R: Testicular germ-cell cancer. N Engl J Med
4. Bajorin D, Sarosdy D, Mazumdar M, et al: Randomized trial of
etoposide and cisplatin vs etoposide and carboplatin in patients with
good-risk germ cell tumors: A multi-institutional study. J Clin Oncol
5. Allen J, Kim J, Packer R: Neoadjuvant chemotherapy for newly
diagnosed germ-cell tumors of the central nervous system. J Neurosurg
6. Allen J, DaRosso R, Donahue B: A phase II trial of preradiation
carboplatin in newly diagnosed germinoma of the central nervous
system. Cancer 74:940-944, 1994.
7. Calaminus G, Bamberg M, Baranzelli M: Intracranial germ cell
tumors: A comprehensive update of the European data. Neuropediatrics