Bridging Science and Clinical Practice: How to Use Molecular with Brain Cancer

CENTRAL NERVOUS SYSTEM TUMORS
Bridging Science and Clinical
Practice: How to Use Molecular
Markers When Caring for a Patient
with Brain Cancer
CHAIR
Martin J. Van den Bent, MD
Erasmus University Medical Center
Rotterdam, Netherlands
SPEAKERS
Monika E. Hegi, PhD
University of Lausanne Hospitals
Lausanne, Switzerland
Myrna Rachel Rosenfeld, MD, PhD
University of Pennsylvania
Philadelphia, PA
MYRNA R. ROSENFELD
Bridging Science and Clinical Practice: How to Use Molecular
Markers When Caring for a Patient with Brain Cancer
Myrna R. Rosenfeld, MD, PhD
OVERVIEW
Technical advances in genomic and proteomic profiling and bioinformatics have resulted in the identification of a large number of
possible prognostic, predictive, and diagnostic molecular markers in glial tumors. Increasingly, clinical trials are incorporating tissue
analyses to prospectively and retrospectively study the value of these and yet-to-be defined markers. Once validated, markers form
the basis for increasingly stringent classification schemes and the development of personalized, targeted therapies. Descriptions of
molecular marker findings, many not validated as clinically relevant, are filling pathology reports, and patients arrive to clinic with the
latest journal article requesting marker assessment. Although some practitioners may choose to incorporate these findings into clinical
decision making, empirical data supporting these decisions is limited to a few specific circumstances. This article reviews three
markers— codeletion of 1p/19q, methylation of the O6-methylguanine-DNA methyltransferase (MGMT) promoter, and the presence of
IDH1/2 mutation—for which there exists high or moderate levels of evidence of current clinical utility for guiding diagnostic, prognostic,
and treatment decisions.
N
euro-oncologists have long practiced personalized
medicine. In the absence of highly effective therapies,
treatment decisions for adults with malignant gliomas (including low-grade glioma) were made on the basis of histopathologic and clinical factors such as histologic type, patient
age and performance status, tumor size and location, and extent of surgical resection. Outcomes, however, even within a
relatively homogenous group such as those defıned by the
Radiation Therapy Oncology Group (RTOG) recursive partitioning analysis (RPA), remained variable, in part as a result
of the subjective nature of histologic grading or interpretation of extent of resection. The identifıcation of molecular
markers would facilitate stratifıcation of patients into clinically relevant categories. Currently there are three markers—
codeletion of 1p and 19q, methylation status of the O6methylguanine-DNA methyltransferase (MGMT) gene
promoter, and isocitrate dehydrogenase 1 and 2 mutations
(IDH1/2)—for which there are strong enough data that these
markers can be taken into consideration when making either
diagnostic or care decisions (Table 1).
1p/19q CODELETION IN ANAPLASTIC
OLIGODENDROGLIOMAS
Use in Clinical Decision Making
Prognostic. A positive prognostic marker in anaplastic glioma; may be an independent positive prognostic marker in
low-grade glioma.
Predictive. In newly diagnosed anaplastic oligodendroglioma
(AO), the presence of 1p/19q codeletion predicts longer survival when chemotherapy is added to radiation.
Diagnostic. Useful to support the diagnosis of oligodendroglioma.
Background
The observation in the early 1990s that many patients with
AO showed responses to chemotherapy, (mostly a combination of procarbazine, lomustine, and vincristine [PCV]) and
that response and improved outcomes were associated with
codeletion of chromosome arms 1p and 19q lead to two prospective phase III trials, RTOG 9402 and European Organisation for Research and Treatment of Cancer (EORTC)
26951.1 The initial results published after a follow-up of at
least 3 years demonstrated that compared with radiation
alone, patients with newly diagnosed AO obtained no overall
survival (OS) benefıt with the addition of neoadjuvant
(RTOG 9402) or adjuvant (EORTC 26951) PCV.2,3 Both
studies demonstrated a difference in the time to tumor progression strongly favoring patients who received chemotherapy. The use of PCV was associated with a signifıcant rate of
hematologic toxicity, with grade 3 or 4 toxicity reported in up
to 65% of patients. Because of the uncertain benefıt and
added toxicity of PCV, neither of these regimens became a
generally accepted standard of care.
From the Department of Neurology, Hospital Clínic/IDIBAPS, Barcelona, Spain.
Author’s disclosures of potential conflicts of interest are found at the end of this article.
Corresponding author: Myrna R. Rosenfeld, MD, PhD, Department of Neurology, Hospital Clínic /IDIBAPS, c/ Villarroel, 170, Barcelona, Spain 08036; email: [email protected].
© 2013 by American Society of Clinical Oncology.
108
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
MOLECULAR MARKERS IN NEURO-ONCOLOGY
The studies validated 1p/19q codeletion as a favorable
prognostic marker for OS. In RTOG 9402 patients with 1p/
19q codeletion had median OS of at least 7 years compared
with 2.8 years for those without the codeletion (p ⱕ 0.001).
Although neither study had a chemotherapy-only arm or a
delayed radiation arm, on the assumption that radiation at
recurrence after upfront PCV would be effective, these results led some practitioners to delay radiation until progression in patients with codeletions to avoid late cognitive
effects. The poor survivals of patients with nondeleted tumors led some to use the radiation and temozolomide (TMZ)
regimen used for treating GB.
Now with a median follow-up of 11.3 years, outcomes have
been further clarifıed. In RTOG 9402, the OS of patients with
1p/19q codeletions who received neoadjuvant PCV was 14.7
years compared with 7.3 years (p ⫽ 0.03) for those who received only radiation.4 The addition of PCV to radiation had
no impact on the OS of patients without 1p/19q codeletions
(2.6 vs. 2.7 years; p ⫽ 0.39). Similar but not quite statistically
signifıcant data were found for adjuvant PCV (p ⫽ 0.059) in
EORTC 26951.5
These studies suggest that neoadjuvant or adjuvant PCV is
highly effective for 1p/19q codeleted tumors. Yet not all
neuro-oncologists are adopting this strategy, with some opting to replace PCV with TMZ or continuing to defer radiation until progression. The reasoning behind these choices is
varied and includes the increased toxicity of the PCV regimen, the proven effıcacy of TMZ in GB, and studies that, although not formally designed to compare PCV with TMZ,
suggest equal effıcacy (studies also exist suggesting superiority of PCV).6-8
In both studies, the divergence of survival curves for patients with 1p/19q codeletion did not occur until after 5 years
suggesting that not all patients with codeletion benefıt
equally from PCV plus radiation. It also suggests that treatment at recurrence in some patients can be effective and that
delay of radiation may not be detrimental.
KEY POINTS
䡠 Codeletion of 1p/19q is a predictive marker in anaplastic
oligodendroglioma and can be used for treatment decisions.
䡠 MGMT promoter methylation status is a predictive marker
in newly diagnosed glioblastoma (GB); however, at this
time, it likely should be used only for treatment decisions
in elderly patients.
䡠 IDH1/2 mutation is a strong prognostic marker in all grades
of gliomas but does not currently have predictive value.
䡠 Triple-mutant glioma (intact 1p/19q, nonmethylated MGMT
promoter, and wild-type IDH1/2) identifies a distinct
subclass of glioma with poor outcome; this status can
guide treatment decisions, especially at recurrence.
䡠 The use of other markers for treatment decisions outside
of clinical trials is not empirically based.
Survival curves for patients without codeletion also diverged in favor of PCV and radiation. In RTOG 9402, 18% of
patients without codeletion were alive at 10 years although
median OS was 2.6 years, suggesting a benefıt of PCV and
radiation in a subgroup not defıned by 1p/19q codeletion.4
How do these studies direct treatment decisions for patients with anaplastic tumors without 1p/19q codeletions?
For this group, neither RTOG 9402 nor EORTC 26951 demonstrated a survival benefıt from PCV, and only EORTC
26951 showed a signifıcant increase in progression-free survival (PFS) with the addition of PCV (median, 1.2 vs. 0.7
years; p ⫽ 0.026). These poor outcomes are similar to those
seen in GB. Therefore, some practitioners treat GB in these
patients with radiation and TMZ even though there are no
clear data supporting this as effective. The Chemoradiation
and Adjuvant Temozolomide in Nondeleted Anaplastic Tumors (CATNON) study may provide direction. This study
has four arms: (1) radiation, (2) radiation and concurrent
TMZ, (3) radiation followed by TMZ, and (4) radiation with
concurrent and adjuvant TMZ. Eligibility is predicated on
the absence of the 1p/19q codeletion and not histology, and
the study is ongoing.
Recommendations
Patients with newly diagnosed 1p/19q codeleted anaplastic
tumors should not receive treatment with radiation only.
There are no prospective studies comparing responses to
PCV and TMZ in this population, and available studies are
not defınitive. Thus, evidenced-based practice directs the use
of PCV and radiation; TMZ can be considered if PCV is not
tolerated. The decision to delay radiation is fraught with diffıculty as data are scant; however, as noted above, for some
patients treatment at recurrence appears to be effective. This
author would consider delay of radiation in a young adult
with 1p/19q codeletion and with previously defıned good
prognostic factors such as gross total resection and high performance status.
Although oncologists consider PCV toxicities relatively
mild compared with those produced by other multiagent regimens, for some systemic cancers, the trade-off can be a durable remission. Patients with glioma rarely have durable
remissions, and one must weigh even moderate toxicity
against impact on quality of life and expected improvement
in OS. Some practitioners (who have been prescribing PCV
consistently even after TMZ became available) do not stringently recommend dietary restrictions to patients. The
avoidance of foods high in tyramine because of weak inhibition of monoamine oxidase by procarbazine results in patients receiving an extensive list of foods to eliminate, many
of which are staples of many diets. A fascinating review of
food restriction and procarbazine written more than 30 years
ago persuasively asserts that these food interactions are of
minor clinical signifıcance.9 Because food restriction can adversely affect quality of life and lead to unwanted weight loss,
if PCV is to be used more commonly, perhaps after 30 years it
is time to readdress this issue.
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
109
MYRNA R. ROSENFELD
TABLE 1. Clinical Value of Three Commonly Considered Molecular Markers in Glioma
1p/19q co-deletion
Clinical significance
MGMT promoter
methylation
Clinical significance
IDH1/2
Prognostic
Predictive
Diagnostic
Comment
Favorable: anaplastic glioma*
(possibly in low-grade
glioma)
Anaplastic glioma#
Yes, presence supports diagnosis of
oligodendroglioma, but absence
does not rule out this diagnosis
Clinical value demonstrated
in randomized controlled
trial
Patients with newly diagnosed anaplastic glioma with 1p/19q co-deletion have longer survivals if PCV is added to radiation.
Favorable: anaplastic and GB&;
unclear role in low-grade
No
No clear consensus on
optimal testing method
that has contributed to
differences between
studies examining
prognostic and predictive
value
Elderly patients with GB without MGMT promoter methylation should be considered for radiation only; those with MGMT promoter
methylation can likely receive TMZ alone.
Favorable: low-grade,
anaplastic and GB
Clinical significance
GB† and in particular elderly
GB patients‡
Not at this time
Yes, distinction with non-malignant
entities, pilocytic astrocytoma
and primary from secondary GB
Reproducible testing
method
A favorable prognostic marker that is not currently useful for directing treatment specific decisions. The uniformly unfavorable prognosis
of triple-negative tumors should be taken into consideration when developing treatment plans such as the use of aggressive but toxic
therapies at recurrence versus focusing efforts on quality of life.
Abbreviations: GB, glioblastoma; MGMT, O6-methylguanine-DNA methyltransferase; PCV, procarbazine, lomustine, vincristine; TMZ, temozolomide.
* For patients treated with radiation, chemotherapy or both2,3
# For the addition of chemotherapy (PCV) to radiation4,5
& For patients treated with radiation and alkylating chemotherapy12
† Of response to alkylating chemotherapy12
‡ For the use of temozolomide in patients ⱖ60 years with GB14,15
1p/19q CODELETION AS A DIAGNOSTIC MARKER
Background
The presence of 1p/19q codeletion can support a pathologic
diagnosis of oligodendroglioma, especially for cases in which
histology does not show many classic oligodendroglial features (e.g., perinuclear halos, rounded nuclei). However, up
to 20% of GB can have oligodendroglial features and of these,
5% to 25% will have 1p/19 codeletion. Thus, the presence of
1p/19q codeletion is not by itself suffıcient to diagnose oligodendroglioma.
Tumors that can mimic oligodendroglioma include dysembryoplastic neuroepithelial tumors (DNET), neurocytomas, clear cell ependymoma and small-cell variants of
anaplastic astrocytoma and GB.10 These tumors have varying
clinical outcomes and treatment approaches. Because they
do not exhibit 1p/19q codeletion, the detection of the deletion is a useful ancillary diagnostic tool.
The EORTC/National Cancer Institute of Canada (NCIC)
trial that demonstrated the benefıt of TMZ in newly diagnosed GB also confırmed the relationship between clinical
outcome and tumor levels of the DNA repair enzyme O6methylguanine-DNA methyltransferase (MGMT).11,12
MGMT levels were indirectly determined by analysis of
MGMT promoter methylation, an epigenetic change that silences gene transcription. MGMT promoter methylation was
an independent favorable prognostic factor (p ⬍ 0.001) and
predictive of response to TMZ and radiation. Patients with
methylated MGMT promoters who received TMZ in addition to radiation had median OS of 21.7 months compared
with 15.3 months for those receiving radiation alone (p ⫽
0.007). Patients with nonmethylated MGMT showed a small
but not quite signifıcant (p ⫽ 0.06) survival benefıt from
TMZ compared with radiation. Temozolomide concurrent
with radiation and 6 months of adjuvant TMZ became the
standard of care, but not just for patients with methylated
MGMT. In the absence of a therapeutic alternative to offer
patients with nonmethylated tumors and considering the
small benefıt that some of these patients received, outside of a
clinical trial TMZ is not withheld from patients on the basis
of MGMT methylation status.
Some practitioners use MGMT status as the basis for continuing adjuvant TMZ beyond the 6-month regimen used in
the EORTC/NCIC trial, assuming this would translate into
better outcomes. On the basis of data showing that prolonged
exposure to TMZ depletes cellular MGMT, others asked
whether TMZ could be used to overcome TMZ resistance in
MGMT PROMOTER METHYLATION STATUS
Use in Clinical Decision Making
Prognostic. Positive prognostic factor for improved outcome
in anaplastic gliomas and GB regardless of therapy type (radiation, chemotherapy, or both).
Predictive. Yes, of benefıt to alkylating chemotherapy in GB
including in older patients.
Diagnostic. Can aid in distinguishing tumor progression
from pseudoprogression.
110
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
MOLECULAR MARKERS IN NEURO-ONCOLOGY
nonmethylated tumors. The RTOG 0525 study of newly diagnosed GB addressed these scenarios. In this study, after
completion of radiation with concurrent TMZ, patients were
randomly assigned to undergo 12 cycles of standard 5-dayson, 23-days-off TMZ dosing or lower-dose TMZ on a 21days-on, 7-days-off schedule. The fınal results confırmed that
MGMT status was a positive prognostic factor for improved
outcome.13 However, comparison of the two dosing arms
showed no signifıcant difference in OS or median PFS. Patients in the 21/28 arm also had more grade 3 or 4 episodes of
lymphopenia and fatigue. Thus, neither longer exposure to
TMZ in patients with MGMT methylation nor the use of
(theoretically) MGMT-depleting dose-dense TMZ in nonmethylated tumors affected outcome. These phase III data do
not support a change in the standard 5/28 dosing regimen or
the use of TMZ beyond six adjuvant cycles.
Recently a predictive role for MGMT promoter methylation was demonstrated in two trials studying older patients
(ⱖ 60 years) with GB and anaplastic glioma. The NOA-08
study randomly assigned patients to receive radiation or
TMZ, and showed no difference in median OS between the
treatments.14 MGMT promoter methylation was associated
with prolonged OS (11.9 months for methylated compared
with 8.2 months for nonmethylated; p ⫽ 0.014) for the group
as a whole, and longer PFS if the patient received TMZ compared with radiation (8.4 vs. 4.6 months). In contrast, patients without MGMT promoter methylation had longer PFS
when receiving radiation (4.6 vs. 3.3 months). The NORDIC
trial randomly assigned older patients to undergo a standard
6 weeks of radiation, hypofractionated radiation delivered
over 2 weeks, or TMZ alone.15 Temozolomide and hypofractionated radiation each provided better OS than did standard
radiation. MGMT promoter methylation was associated with
signifıcantly better OS in patients who received TMZ (9.7 vs.
6.8 months; p ⫽ 0.03) but not in those who received radiation
(8.2 vs. 7.0 months; p ⫽ 0.88). Thus, in older patients with
high-grade glioma, MGMT promoter methylation is predictive of response to TMZ. Because older patients without
MGMT methylation do not seem to derive a benefıt from
TMZ, it is reasonable to consider radiation alone with hypofractionated dosing decreasing overall time needed for treatment.
Recommendations
Because the phase III EORTC/NCIC study included patients
70 years of age or younger, knowledge of MGMT status in
this age group is not needed currently to move forward to
initial standard postsurgical care for a patient with GB. Outside of a clinical trial or medical contraindication, patients
with newly diagnosed GB should be offered TMZ. The predictive value of MGMT status in older patients should be
taken into consideration when making treatment recommendations—in particular, the choice of short-course hypofractionated radiation for patients with nonmethylated
MGMT and TMZ without radiation for those with methylated MGMT.
MGMT Promoter Methylation Status as a Diagnostic
Marker
MGMT promoter methylation status can assist in distinguishing tumor progression from pseudoprogression that
occurs in 20% to 30% of patients receiving radiation and
TMZ.16 Several reports have shown that patients with GB
who have MGMT promoter methylation have an increased
rate of developing pseudoprogression compared with those
without methylation.17 In the appropriate clinical context
and in association with other data such as advanced MRI,
MGMT methylation status can be helpful to support a diagnosis of pseudoprogression.
IDH1/2 MUTATION
Use in Clinical Decision Making
Prognostic. A positive prognostic marker in low-grade, anaplastic glioma, and although rarely found, in GB.
Predictive. Not defıned at this time.
Diagnostic. Useful to distinguish glioma from other entities.
Background
Mutations of IDH1 and, less commonly, IDH2 are found in
more than 50% to 80% of low-grade and anaplastic gliomas
and secondary GB (those that progress from lower-grade astrocytomas) but only 5% to 10% of primary (de novo) GB.18,19
A positive prognostic value of IDH1 mutation is supported
by many studies. In one, patients with low-grade gliomas and
IDH1 mutation had 4.7-year median PFS and 42% fıve-year
OS compared with 1.4 years PFS and 14% fıve-year OS for
those with wild-type IDH.20 A study of patients with anaplastic astrocytoma and GB found that IDH1 mutation was the
strongest independent prognostic factor for good outcome.
In general, for patients with GB and IDH1 mutation, studies
report median OS ranging from 24 to 36 months compared
with 9 to 15 months for wild-type IDH1 GB.21,22
Mutation of IDH1 is strongly associated with younger age
(⬍ 50 years), TP53 mutation, 1p19q codeletion and MGMT
promoter methylation; 90% to 100% of 1p/19q codeleted
gliomas have IDH1 or -2 mutations. In contrast tumors with
wild-type IDH1 often do not have 1p/19q codeletion or TP53
mutation. A study of triple-negative (wild-type IDH and
TP53, 1p/19q intact) low-grade glioma demonstrated that
these tumors more commonly involved the insula, were large
in size, had a more infıltrative MRI pattern, and occurred in
older patients compared with tumors with IDH1 and TP53
mutation and 1p/19q codeletions.20 Patients with triplenegative tumors had a 54% 5-year OS compared with 91% for
those with IDH1 and TP53 mutations and methylated
MGMT.
Recommendations
Although the presence of IDH1/2 mutation provides prognostic information, this has not yet translated into a change
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
111
MYRNA R. ROSENFELD
in clinical practice, and a predictive role of the mutation has
not yet been defıned. Triple-negative tumors likely constitute
a unique entity with poor clinical outcomes, and this can be
taken into consideration but should not direct treatment decisions.
IDH1/2 Mutation as a Diagnostic Marker
The absence of IDH1/2 mutation in other entities makes it a
very useful diagnostic marker for distinguishing glioma—in
particular, low-grade glioma—from other low-grade tumors
such as pilocytic astrocytoma and from nontumor entities
such as gliosis, ischemia, or radiation-induced damage, and
between primary and secondary GB.23 Because 60% to 80% of
pilocytic astrocytomas show BRAF duplication or express a
BRAF fusion protein not found in other gliomas, the combination of IDH mutation analysis and BRAF testing is highly
useful in cases with indeterminate histologic features. This
can be particularly valuable when there is limited tissue availability.
CONCLUSION
Many factors determine the clinical value of a molecular
marker, including the reproducibility of the laboratory methods used for marker measurement and the stringency of the
data that determined the markers’ clinical value. The testing
method for some markers, such as IDH1/2 mutation, is relatively straightforward and reproducible. In contrast, multiple
techniques are used for measuring MGMT promoter methylation. A recent National Comprehensive Cancer Network
Task Force report on the clinical utility of molecular markers
used a level-of-evidence system.24 The highest level of evidence (1A) was given to a biomarker that had been evaluated
in at least one adequately powered and specifıcally designed
prospective controlled trial. For gliomas, only 1p/19q codeletion achieved level 1A and some may argue this because
both RTOG 9402 and EORTC 26951were amended after initiation to allow the study of the relationship of 1p/19q status
with outcomes. Furthermore, 1p/19q codeletion was not absolutely predictive of response to PCV and radiation, and
there was a subpopulation without codeletion that seemed to
benefıt from these treatments. In the absence of effective
alternatives or additional markers to defıne these subpopulations, are neuro-oncologists prepared to withhold chemotherapy from patients without 1p/19q codeletion? This is
similar to the scenario that has occurred in GB in the absence
of specifıc therapies for tumors without MGMT promoter
methylation.
Although there are clinical scenarios in which markers can
support diagnostic and treatment decisions, it is not readily
apparent that we are ready to make decisions solely on the
basis of these markers. For example, are the data so strong
that we are compelled to offer a 21-year-old with a 1p/19q
codeleted AO upfront radiation and PCV? Are we so sure
that we cannot delay radiation? Until unequivocal or further
supporting robust data are available, neuro-oncologists and
their patients will decide how to apply this new knowledge,
weighing the goals of increasing survival while maintaining
quality of life.
Disclosures of Potential Conflicts of Interest
Relationships are considered self-held and compensated unless otherwise noted. Relationships marked “L” indicate leadership positions. Relationships marked “I” are those held by an immediate
family member; those marked “B” are held by the author and an immediate family member. Relationships marked “U” are uncompensated.
Employment or Leadership Position: None. Consultant or Advisory Role: Myrna R. Rosenfeld, Merck. Stock Ownership: None. Honoraria: None.
Research Funding: Myrna R. Rosenfeld, Euroimmun. Expert Testimony: None. Other Remuneration: None.
References
1. Cairncross JG, Ueki K, Zlatescu MC, et al. Specifıc genetic predictors of
chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas. J Natl Cancer Inst. 1998;90:1473-1479.
2. Cairncross G, Berkey B, Shaw E, et al. Phase III trial of chemotherapy
plus radiotherapy compared with radiotherapy alone for pure and
mixed anaplastic oligodendroglioma: Intergroup Radiation Therapy
Oncology Group Trial 9402. J Clin Oncol. 2006;24:2707-2714.
3. van den Bent MJ, Carpentier AF, Brandes AA, et al. Adjuvant procarbazine, lomustine, and vincristine improves progression-free survival
but not overall survival in newly diagnosed anaplastic oligodendrogliomas and oligoastrocytomas: a randomized European Organisation for
Research and Treatment of Cancer phase III trial. J Clin Oncol. 2006;24:
2715-2722.
4. Cairncross G, Wang M, Shaw E, et al. Phase III trial of chemoradiotherapy
112
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
for anaplastic oligodendroglioma: long-term results of RTOG 9402.
J Clin Oncol. 2013;31:337-343.
5. van den Bent MJ, Brandes AA, Taphoorn MJ, et al. Adjuvant procarbazine, lomustine, and vincristine chemotherapy in newly diagnosed anaplastic oligodendroglioma: long-term follow-up of EORTC Brain
Tumor Group study 26951. J Clin Oncol. 2013;31:344-350.
6. Wick W, Hartmann C, Engel C, et al. NOA-04 randomized phase III
trial of sequential radiochemotherapy of anaplastic glioma with procarbazine, lomustine, and vincristine or temozolomide. J Clin Oncol. 2009;
27:5874-5880.
7. Brada M, Stenning S, Gabe R, et al. Temozolomide versus procarbazine,
lomustine, and vincristine in recurrent high-grade glioma. J Clin Oncol.
2010;28:4601-4608.
8. Lassman AB, Iwamoto FM, Cloughesy TF, et al. International retrospec-
MOLECULAR MARKERS IN NEURO-ONCOLOGY
9.
10.
11.
12.
tive study of over 1000 adults with anaplastic oligodendroglial tumors.
Neuro Oncol. 2011;13:649-659.
Maxwell MB. Reexamining the dietary restrictions with procarbazine
(an MAOI). Cancer Nurs. 1980;3:451-457.
Aldape K, Burger PC, Perry A. Clinicopathologic aspects of 1p/19q loss
and the diagnosis of oligodendroglioma. Arch Pathol Lab Med. 2007;
131:242-251.
Stupp R, Mason WP, van den Bent MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;
352:987-996.
Hegi ME, Diserens AC, Gorlia T, et al. MGMT gene silencing and benefıt
from temozolomide in glioblastoma. N Engl J Med. 2005;352:997-1003.
13. Gilbert MR, Wang M, Aldape KD, et al. RTOG 0525: a randomized
phase III trial comparing standard adjuvant temozolomide (TMZ) with
a dose-dense (DD) schedule in newly diagnosed glioblastoma (GBM).
J Clin Oncol. 2011;29(suppl), abstr 2006.
14. Wick W, Platten M, Meisner C, et al. Temozolomide chemotherapy alone
versus radiotherapy alone for malignant astrocytoma in the elderly: the
NOA-08 randomised, phase 3 trial. Lancet Oncol. 2012;13:707-715.
15. Malmstro¨m A, Grønberg BH, Marosi C, et al. Temozolomide versus
standard 6-week radiotherapy versus hypofractionated radiotherapy in
patients older than 60 years with glioblastoma: the Nordic randomised,
phase 3 trial. Lancet Oncol. 2012;13:916-926.
16. Taal W, Brandsma D, de Bruin HG, et al. Incidence of early pseudoprogression in a cohort of malignant glioma patients treated with chemoirradiation with temozolomide. Cancer. 2008;113:405-410.
17. Brandes AA, Franceschi E, Tosoni A, et al. MGMT promoter methylation status can predict the incidence and outcome of pseudoprogression after concomitant radiochemotherapy in newly diagnosed
glioblastoma patients. J Clin Oncol. 2008;26:2192-2197.
18. Yan H, Parsons DW, Jin G, et al. IDH1 and IDH2 mutations in gliomas.
N Engl J Med. 2009;360:765-773.
19. Parsons DW, Jones S, Zhang X, et al. An integrated genomic analysis of
human glioblastoma multiforme. Science. 2008;321:1807-1812.
20. Metellus P, Coulibaly B, Colin C, et al. Absence of IDH mutation identifıes a novel radiologic and molecular subtype of WHO grade II gliomas
with dismal prognosis. Acta Neuropathol. 2010;120:719-729.
21. Hartmann C, Hentschel B, Wick W, et al. Patients with IDH1 wild type
anaplastic astrocytomas exhibit worse prognosis than IDH1-mutated
glioblastomas, and IDH1 mutation status accounts for the unfavorable
prognostic effect of higher age: implications for classifıcation of gliomas.
Acta Neuropathol. 2010;120:707-718.
22. Weller M, Felsberg J, Hartmann C, et al. Molecular predictors of
progression-free and overall survival in patients with newly diagnosed
glioblastoma: a prospective translational study of the German Glioma
Network. J Clin Oncol. 2009;27:5743-5750.
23. Nobusawa S, Watanabe T, Kleihues P, Ohgaki H. IDH1 mutations as
molecular signature and predictive factor of secondary glioblastomas.
Clin Cancer Res. 2009;15:6002-6007.
24. Febbo PG, Ladanyi M, Aldape KD, et al. NCCN Task Force report: Evaluating the clinical utility of tumor markers in oncology. J Natl Compr
Canc Netw. 2011;9 suppl 5:S1-S32.
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
113
M. J. VAN DEN BENT
How to Use Molecular Markers When Caring for a Patient with
Brain Cancer: 1P/19Q as a Predictive and Prognostic Marker in
the Neuro-oncology Clinic
M. J. van den Bent, MD
OVERVIEW
Although the central role of 1p/19q codeletion in oligodendroglioma was established almost two decades ago, apart from clear
prognostic significance the implications for clinical care have been less clear. This has changed with the long-term follow-up analysis
of the EORTC and RTOG trials on procarbazine, lomustine, and vincristine (PCV) chemotherapy in anaplastic oligodendroglioma. These
have shown that 1p/19q loss in these tumors is predictive of overall survival benefit of the addition of PCV chemotherapy to radiotherapy.
I
n 1998 a National Institutes of Health study group gave the
following defınition of markers: “a characteristic that is
objectively measured and evaluated as an indicator of normal
biologic processes, pathogenic processes, or pharmacologic
responses to a therapeutic intervention.” Markers can have
indeed a variety of clinical signifıcances: they can be useful
when diagnosing a disease, when assessing prognosis in an
individual patient, when measuring response to treatment or
predicting outcome to treatment.
associated with a favorable treatment response) or whether
there will be no treatment effect at all or even an adverse effect in the absence of the marker (qualitative marker). HER2/
neu presence is an example of such a qualitative marker in
breast cancer; no effect is to be expected from herceptin in the
absence of HER2/neu expression. The current knowledge
suggests that 1p/19 codeletion is both a disease marker, a
prognostic marker, and a predicitive marker.
BIOMARKER DEVELOPMENT
DIFFERENT TYPES OF MARKERS
The role of markers can be manyfold, depending on the information they carry. Markers can be considered diagnostic,
when they give information about the presence or absence of
a disease. Examples are serum CA 15.3 in breast cancer and
prostate-specifıc antigen in prostrate cancer. Markers can be
considered prognostic, if they provide prognostic information irrespective of the treatment given. Although this will in
general not lead to an altered treatment, they can still support
decisions on intensity of treatment (e.g., the decision not to
treat in the face of poor prognostic markers) or they can be
used to stratify patients in randomized trials to ensure a wellbalanced patient population across treatment arms. An example here is serum lactate dehydrogenase (LDH) in nonHodgkin lymphoma (NHL), which is part of the prognostic
index for NHL.
Markers can also be predictive for outcome to a specifıc
treatment, which implies that if present the treatment is more
or less likely to be successful. Here, a distinction can be made
between qualitative and quantitative predictivity. These
terms refer to the question of whether the marker simply
predicts more treatment activity (assuming the marker is
Biomarker development is as cumbersome as the conduct of
randomized phase III trials, and as a rule of thumb similar
numbers of patients are required. All too often data are presented from small series in which a trend toward a better outcome in the presence of a certain biologic marker is found,
which result is as a rule decorated with a nice p value. Even in
the presence of these nice p values, positive and negative predictive values can be disappointing, and in the case of predictive markers, tests for interaction should also be statistically
signifıcant. Then, confırmatory tests in independent datasets
are required. It is important to realize that biomarkers investigated in uncontrolled series can only suggest prognostic information. From such projects one cannot determine whether a
marker is correlated to the outcome of a specifıc treatment. That
requires biomarker assessment in a well-controlled trial.
THE 1P/19Q TRANSLOCATION
1p/19 codeletion was the fırst genomic biomarker that heralded promise for widespread use in neuro-oncology. In
1994, Reifenberger and colleagues described the combined
loss of 1p/19q as typical for oligodendroglial tumors.1 In 1997
research on low-grade mixed oligoastrocytoma showed that
From the Department of Neuro-oncology, ErasmusMC-Cancer Institute, Rotterdam, Netherlands.
Author’s disclosures of potential conflicts of interest are found at the end of this article.
Corresponding author: M. J. van den Bent, Dept. of Neuro-oncology, ErasmusMC-Cancer Institute, Groene Hilledijk 301, Rotterdam, Netherlands; email: [email protected].
© 2013 by American Society of Clinical Oncology.
114
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
1P/19Q LOSS IN NEURO-ONCOLOGY
these tumors had either 1p/19q codeletion or TP53 mutations, suggesting these lesions could be used to distinguish
between the tumors from oligodendroglial and from astrocytic lineage.2 Subsequently, in another seminal paper it was
demonstrated that these 1p/19 codeleted tumors represented
the PCV-chemotherapy sensitive oligodendrogliomas.3 Despite that, non-1p/19q codeleted tumors still had a 25% response rate to PCV. Further studies showed that 1p/19q
codeleted tumors represent a specifıc subset of tumors, which
are more often frontally located, with a more indolent clinical
behavior, with more often a classical oligodendroglial morphologic appearance, and also responsive to temozolomide
chemotherapy and radiotherapy.4,5 Several papers focused
on radiologic features, although with somewhat different
conclusions, but ring enhancement is rare for a 1p/19q codeleted oligodendroglioma.3,6 It was then shown that this 1p/
19q codeletion is actually a t(1;19)(q10;p10) translocation
after which 1p and 19q are lost.7,8
Despite the quantitative predictive effect of 1p/19q loss for
response of recurrent oligodendroglial tumors to PCV chemotherapy, the 2006 reports on the large adjuvant PCV randomized studies failed to show a predictive value of 1p/19q
status, although these trials established a clear prognostic role
of 1p/19q status in anaplastic oligodendrogliomas.9,10 The
long-term follow-up data from both studies, however, show
that in the 1p/19q codeleted anaplastic oligodendroglial tumors, the addition of PCV to radiation therapy signifıcantly
improves overall survival with a hazard ratio (HR) reduction
of 0.55-0.60 (EORTC study: HR 0.56; 95% CI, 0.31-1.03;
RTOG study: HR 0.59; 95% CI, 0.37-0.95).11,12 It is noteworthy that each single trial would not have allowed the conclusion that 1p/19q deletion is predictive for outcome to adjuvant
PCV chemotherapy, but the combined data showing exactly the
same long-term effect is the convincing observation.
Is 1p/19q all there is, when it comes to predicting benefıt
from PCV chemotherapy? Probably not. Both trials suggest
that other molecular features may allow assessment of benefıt
to PCV. In particular, gene expression arrays and genomewide methylation studies have yielded interesting results.13,14
Other single gene candidate markers are IDH and MGMT
promoter methylation. Which technique offers the best
KEY POINTS
䡠 1p/19q codeletion reflects a t(1;19)(q10;p10) translocation.
䡠 Glial tumors with 1p/19q codeletion usually also have IDH
mutations and MGMT promoter methylation.
䡠 The chemotherapy sensitivity of 1p/19q codeleted anaplastic
oligodendroglial tumors is also reflected in improved
overall survival if chemotherapy is given immediately after
or before radiotherapy.
䡠 1p/19q status should be determined in all grade 2 and
grade 3 tumors, especially if an oligodendroglial component
is present.
䡠 The clinical significance of recently identified CIC and FUBP
mutations in 1p/19q codeleted tumors is unclear.
predictive power remains to be established. It is also clear that
the increased responsiveness of p/19q codeleted tumors is
not limited to PCV, but includes temozolomide and radiation therapy.15,16 Still, absence of 1p/19q loss does not imply a
patient will not benefıt from chemotherapy. It is at this point
in time, however, uncertain if a patient with a grade 3 tumor
will benefıt in a clinically signifıcant way from the addition of
adjuvant chemotherapy to radiation therapy in the absence
of the 1p/19q codeletion.
ASSESSMENT OF 1P/19Q CODELETION
With these combined data, the prognostic and predictive
roles of 1p/19q have been clearly established. Assessing 1p/
19q status may also help in case of diagnostic uncertainty
with oligodendroglial mimics (e.g., central neurocytoma).
Assessing 1p/19q status should now be considered in all
grade 2 and grade 3 tumors, especially in the presence of oligodendroglial features. A variety of techniques exist to assess
1p/19q status, and efforts have been made to correlate these
to each other (with in general good but certainly not complete correlations).17,18 Both loss of heterozygosity and FISH
are used frequently, for 1p/19q FISH an extensive guideline
exist.19 It is important to realize that the 1p36.6 probe frequently used to assess loss of 1p can also be deleted in glioblastoma that show the loss of only the tip of 1p.20 If LOH is
used for assessment of 1p/19q status, care should be taken
that several markers on the 1p arm are absent. Indeed, bot the
entire arm of 1pand 19q are lost. Grade 3 tumors not infrequently show loss of 19q without additional 1 loss. Laboratories conducting 1p/19q analysis should participate to quality
programs that include ring tests.
NOVEL DISCOVERIES
Next-generation sequencing (Illmina HiSeq platform) has
identifıed in 1p/19q codeleted oligodendroglial tumors novel
and probably inactivating mutations in the CIC and FUBP
genes. In a total of 34 1p/19q codeleted tumors 18 (53%) mutations were identifıed in the CIC gene located on 19q, and 5
(15%) in the FUBP gene located on 1p. No CIC mutations
were identifıed in tumors without 19p loss. Subsequent studies confırmed the high mutation rate of CIC in 1p/19q codeleted tumors but only rarely in tumors without the
deletion.21-23 Taken together, CIC mutations appear to occur
in 50% to 70% of 1p/19 codeleted tumors and FUBP mutations in 15%. CIC mutations occur almost invariably with
IDH mutations, and occur almost exclusively in 1p/19q
codeleted grade 2 and 3 oligodendroglial tumors. They are
exceedingly rare in other brain tumors. Cic is a downstream
component of receptor kinase pathways (RTK-RAS-RAFMAPK). Cic blocks transcription through binding to regulatory regions, and is negatively regulated by RTK signaling,
which blocks the function of cic through -MAPK mediated
phosphorylation. This results in the degradation of cic. FUBP
mutations may result in MYC activation. Although fascinating new developments, the clinical signifıcance of these mutations is currently unknown.
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
115
M. J. VAN DEN BENT
PRACTICAL APPROACH OF 1P/19Q TESTING
IN 2013 NEURO-ONCOLOGY
In general, the assessment of markers should lead to alterations in the management of patients. The clinical value of a
marker that is prognostic only is limited, as it will not change
management of the patient.
The current data suggest testing for 1p1/9q loss in all grade
2 and 3 diffuse glial tumors, especially when some oligodendroglial characteristics are present. 1p/19q status can help to
decide for treatment with chemotherapy alone, or for the decision to add chemotherapy to radiotherapy. Both these approaches are much more likely to bring patient benefıt in the
presence of 1p/19q codeletion.
Disclosures of Potential Conflicts of Interest
Relationships are considered self-held and compensated unless otherwise noted. Relationships marked “L” indicate leadership positions. Relationships marked “I” are those held by an immediate
family member; those marked “B” are held by the author and an immediate family member. Relationships marked “U” are uncompensated.
Employment or Leadership Position: None. Consultant or Advisory Role: Martin J. Van Den Bent, Abbott Laboratories; Antisense Pharma; Merck KGaA;
Merck Sharpe & Dohme; Roche; Siena Biotech. Stock Ownership: None. Honoraria: Martin J. Van Den Bent, Merck KGaA; Merck Sharpe & Dohme; Roche.
Research Funding: Martin J. Van Den Bent, Roche. Expert Testimony: None. Other Remuneration: Martin J. Van Den Bent, UpToDate.
References
1. Reifenberger J, Reifenberger G, Liu L, et al. Molecular genetic analysis of
oligodendroglial tumors shows preferential allelic deletions on 19q and
1p. Am J Pathol. 1994;145:1175-1190.
2. Maintz D, Fiedler K, Koopmann J, et al. Molecular genetic evidence for subtypes of oligoastrocytomas. J Neuropathol Exp Neurol. 1997;56:1098-1104.
3. Cairncross JG, Ueki K, Zlatescu MC, et al. Specifıc genetic predictors of
chemotherapeutic response and survival in patients with anaplastic oligodendrogliomas. J Natl Canc Inst. 1998;90:1473-1479.
4. Kouwenhoven MC, Kros JM, French PJ, et al. 1p/19q loss within oligodendroglioma is predictive for response to fırst line temozolomide but
not to salvage treatment. Eur J Cancer. 2006;42:2499-2503.
5. Kaloshi G, Benouaich-Amiel A, Diakite F, et al. Temozolomide for lowgrade gliomas: predictive impact of 1p/19q loss on response and outcome. Neurology. 2007;68:1831-1836.
6. Jenkinson MD, du Plessis DG, Smith TS, et al. Histological growth patterns and genotype in oligodendroglial tumours: correlation with MRI
features. Brain. 2006;129:1884-1891.
7. Jenkins RB, Blair H, Ballman KV, et al. A t(1;19)(q10;p10) mediates the
combined deletions of 1p and 19q and predicts a better prognosis of
patients with oligodendroglioma. Cancer Res. 2006;66:9852-9861.
8. Griffın CA, Burger P, Morsberger L, et al. Identifıcation of der(1;
19)(q10;p10) in fıve oligodendrogliomas suggests mechanism of concurrent 1p and 19q loss. J Neuropathol Exp Neurol. 2006;65:988-994.
9. Cairncross G, Berkey B, Shaw E, et al. Phase III trial of chemotherapy
plus radiotherapy compared with radiotherapy alone for pure and
mixed anaplastic oligodendroglioma: Intergroup Radiation Therapy
Oncology Group Trial 9402. J Clin Oncol. 2006;24:2707-2714.
10. van den Bent MJ, Carpentier AF, Brandes AA, et al. Adjuvant procarbazine, lomustine, and vincristine improves progression free survival
but not overall survival in newly diagnosed anaplastic oligodendrogliomas and oligoastrocytomas: a randomized European Organisation for
Research and Treatment of Cancer phase III trial. J Clin Oncol. 2006;24:
2715-2722.
11. van den Bent MJ, Brandes AA, Taphoorn MJ, et al. Adjuvant procarbazine, lomustine, and vincristine chemotherapy in newly diagnosed anaplastic oligodendroglioma: long-term follow-up of EORTC Brain Tumor
Group Study 26951. J Clin Oncol. 2013; 31:344-350. Epub 2012 Oct 15.
12. Cairncross G, Wang M, Shaw E, et al. Phase III trial of chemoradiotherapy
116
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
for anaplastic oligodendroglioma: long-term results of RTOG 9402.
J Clin Oncol. 2013;31:337-343. Epub 2012 Oct 15.
Erdem-Eraslan L, Gravendeel LA, de Rooi JJ, et al. Intrinsic molecular
subtypes of glioma are prognostic and predict benefıt from adjuvant
procarbazine, lomustine, and vincristine chemotherapy in combination
with other prognostic factors in anaplastic oligodendroglial brain tumors: a
report from EORTC study 26951. J Clin Oncol. 2013;31:328-336.
van den Bent MJ, Gravendeel LA, Gorlia T, et al. A hypermethylated phenotype in anaplastic oligodendroglial brain tumors is a better predictor of
survival than MGMT methylation in anaplastic oligodendroglioma: a report from EORTC study 26951. Clin Cancer Res. 2011;17:7148-7155.
Wick W, Weller M. NOA-04 randomized phase III study of sequential
radiochemotherapy of anaplastic glioma with PCV or temozolomide.
J Clin Oncol. 2008; 26 (suppl; abstr LBA2007).
Bauman GS, Ino Y, Yeki K, et al. Allelic loss of chromosome 1p and
radiotherapy plus chemotherapy in patients with oligodendrogliomas.
Int J Radiat Oncol Biol Phys. 2000;48:825-830.
Smith JS, Perry A, Borell TJ, et al. Alterations of chromosome arms 1p
and 19q as predictors of survival in oligodendrogliomas, astrocytomas,
and mixed oligoastrocytomas. J Clin Oncol. 2000;18:636-645.
Franco-Hernandez C, Martinez-Glez V, Torres-Martin M, et al. Identifıcation of genetic alterations by multiple ligation-dependent probe amplifıcation (MLPA) analysis in oligodendrogliomas. Neurocirugia (Astur). 2009;
20:117-123.
Woehrer A, Sander P, Haberler C, et al. FISH-based detection of 1p 19q
codeletion in oligodendroglial tumors: procedures and protocols for
neuropathological practice - a publication under the auspices of the Research Committee of the European Confederation of Neuropathological Societies (Euro-CNS). Clin Neuropathol. 2011;30:47-55.
Idbaih A, Kouwenhoven M, Jeuken J, et al. Chromosome 1p loss evaluation
in anaplastic oligodendrogliomas. Neuropathology. 2008; 28:440-443.
Bettegowda C, Agrawal N, Jiao Y, et al. Mutations in CIC and FUBP1
contribute to human oligodendroglioma. Science. 2011;333:1453-1455.
Yip S, Butterfıeld YS, Morozova O, et al. Concurrent CIC mutations,
IDH mutations, and 1p/19q loss distinguish oligodendrogliomas from
other cancers. J Pathol. 2012;226:7-16.
Jiao Y, Killela PJ, Reitman ZJ, et al. Frequent ATRX, CIC, and FUBP1
mutations refıne the classifıcation of malignant gliomas. Oncotarget.
2012;3:709-722.
CENTRAL NERVOUS SYSTEM TUMORS
Controversies in Antiangiogenesis
Therapy: Point/Counterpoint
CHAIR
John Frederick De Groot, MD
The University of Texas MD Anderson Cancer Center
Houston, TX
SPEAKERS
Tracy Batchelor, MD, MPH
Massachusetts General Hospital
Boston, MA
David A. Reardon, MD
Dana-Farber Cancer Institute
Boston, MA
ANTIANGIOGENIC THERAPY IN GLIOBLASTOMA: MECHANISMS OF ACTION, RESISTANCE, AND FUTURE CHALLENGES
Antiangiogenic Therapy for Glioblastoma: The Challenge of
Translating Response Rate into Efficacy
John de Groot, MD, David A. Reardon, MD, and Tracy T. Batchelor, MD
OVERVIEW
Glioblastoma are one of the mostly vascularized tumors and are histologically characterized by abundant endothelial cell proliferation.
Vascular endothelial growth factor (VEGF) is responsible for a degree of vascular proliferation and vessel permeability leading to
symptomatic cerebral edema. Initial excitement generated from the impressive radiographic response rates has waned due to concerns
of limited long-term efficacy and the promotion of a treatment-resistant phenotype. Reasons for the discrepancy between high
radiographic response rates and lack of survival benefit have led to a focus on identifying potential mechanisms of resistance to
antiangiogenic therapy. However, equally important is the need to focus on identification of basic mechanisms of action of this class
of drugs, determining the optimal biologic dose for each agent and identify the effect of antiangiogenic therapy on oxygen and drug
delivery to tumor to optimize drug combinations. Finally, alternatives to overall survival (OS) need to be pursued using the application
of validated parameters to reliably assess neurologic function and quality of life.
A
pathophysiological hallmark of glioblastoma is the expression of VEGF and other pro-angiogenic cytokines,
which, in turn, stimulate endothelial cell proliferation, migration, and survival.1,2 This leads to the formation of a
highly abnormal tumor vasculature characterized by hyperpermeable vessels, increased vessel diameter, and abnormally
thickened basement membranes. This abnormal vascular
network not only promotes tumor growth but may also impair the effıcacy of cytotoxic chemotherapy and radiation by
enhancing tumor hypoxia and compromising intratumoral
delivery of chemotherapy.
The utilization of angiogenesis inhibitors for the treatment
of glioblastoma has brought hope that blocking this central
component of gliomas will inhibit tumor growth and prolong
patient survival. Early, uncontrolled studies have been encouraging. Both the pan-VEGF receptor (VEGFR)-2 tyrosine
kinase inhibitor cediranib and the anti-VEGF-A antibody
bevacizumab demonstrated impressive radiographic response rates and prolongation of progression-free survival
(PFS) in single arm phase II clinical trials. With a strong tail
wind, these agents entered into phase III clinical trials for
recurrent (cediranib) and newly diagnosed (bevacizumab)
glioblastoma. Although these agents may improve symptoms
and allow for reduction in steroid use, to date, they have
not demonstrated an improvement in OS compared with
standard of care therapy. These data require reflection on potential mechanisms by which antiangiogenic therapy could
augment benefıt associated with current approaches, and
identify mechanisms that may limit the ability of these agents
to improve patient survival.
In this review and at our presentation at the 2013 American
Society of Clinical Oncology Annual Meeting, we will explore the potential biologic mechanisms of action and reasons why antiangiogenic therapy should be effective in the
treatment of glioblastoma. Point by point, we will describe
potential flaws in current approaches to inhibit angiogenesis
and possible mechanisms of resistance limiting effıcacy.
Finally, we will describe areas that require new insights as
well as approaches that may improve understanding and better exploit these therapies to meaningfully improve patient
outcome.
MECHANISMS OF ACTION
There are a number of hypothesized mechanisms of action
of antiangiogenic agents in the treatment of solid tumors.
These mechanisms are not mutually exclusive and may be
operative simultaneously or at different time points during
the course of antiangiogenic treatment. The fırst hypothesis is the classical view that antiangiogenic agents have a
direct effect on tumor endothelium resulting in endothelial
cell apoptosis, ultimately leading to a cytostatic effect on
new blood vessel growth.3 Consequently, there are effects
on vascular function, including vasoconstriction, decreased
From the Department of Neuro-Oncology, University of Texas MD Anderson Cancer Center, Houston, TX; Center for Neuro-Oncology, Dana-Farber Cancer Institute, Boston, MA; Stephen E. and
Catherine Pappas Center for Neuro-Oncology, Massachusetts General Hospital Cancer Center, Boston, MA.
Authors’ disclosures of potential conflicts of interest are found at the end of this article.
Corresponding author: John de Groot, MD, Department of Neuro-Oncology, University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Unit 431, Houston, TX, 77030; email:
[email protected].
© 2013 by American Society of Clinical Oncology.
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
e71
DE GROOT, REARDON, AND BATCHELOR
permeability, and decreased perfusion resulting in decreased
delivery of oxygen and nutrients to the tumor. The resulting
“tumor starvation” inhibits tumor growth. Moreover, when
used in combination therapy, VEGF pathway inhibitors
may sensitize glioma-associated endothelial cells to cytotoxic
therapy and counteract a surge in VEGF expression and endothelial precursor cell recruitment induced by genotoxic
stress from chemotherapy and radiation.4-6 A second hypothesis is that antiangiogenic agents achieve antitumor effects by vascular normalization.7 Preclinical and clinical
studies in a number of solid tumors, including glioblastoma,
have demonstrated that there is a transient period of “vascular normalization” following administration of anti-VEGF
therapies characterized by reduced vessel diameter and permeability, improved vessel perfusion, reduction in tumor
interstitial pressure, and improved tumor oxygenation.8-10
In aggregate, these changes enhance delivery and effıcacy
of cytotoxic chemotherapy during the “normalization window.” Vascular normalization might also improve the effıcacy of ionizing radiation.9 In mouse models of glioblastoma, VEGFR2 inhibition produced a transient period of
reduced tumor hypoxia (i.e., a vascular normalization window) during which radiation had a synergistic effect with
anti-VEGFR2 therapy. During this normalization window,
pericytes are recruited to blood vessels by activation of Ang1/Tie2 signaling.9 These vessels are more effıcient at oxygen
delivery, thus augmenting the effects of radiation. However,
the effect of anti-VEGF agents on chemotherapy delivery to
tumors is incompletely understood. In a small PET study in
patients with non-small cell lung cancer, there was reduced
perfusion and net influx rate of radiolabeled docetaxel within
5 hours after the administration of bevacizumab.10 However, other studies demonstrate that perfusion increases
in a subset of glioblastoma patients within 24 hours after
treatment with anti-VEGF therapy and those patients with
KEY POINTS
䡠 Antiangiogenic therapy rapidly reduces vascular
permeability and cerebral edema, which corresponds with
high radiographic response rates, symptom improvement,
and reduced corticosteroid requirement.
䡠 Despite multiple potential mechanisms by which
antiangiogenic therapy may augment the efficacy of
chemotherapy and radiation, benefits to patients may not
translate into improved survival.
䡠 Effect on drug delivery, vascular endothelial growth
factor-independent angiogenesis, transformation to a
mesenchymal phenotype, and a heterogeneous
microenvironment may contribute to a lack of therapeutic
efficacy.
䡠 Future efforts should address specific knowledge gaps
regarding mechanism of action, identification of patient
subsets more likely to benefit, informative biomarkers, the
value of alternative outcome measures, and the
development of therapies to improve efficacy.
e72
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
improved perfusion have longer OS.11,12 A third hypothesis is
that antiangiogenic agents may have activity against glioblastoma stem-like cells (GSCs). GSCs seem to be at least partially
responsible for resistance to genotoxic treatments through
activation of DNA damage checkpoint response and an increase in DNA repair capacity. In preclinical studies, these
GSCs exhibit upregulation of VEGF expression, form highly
angiogenic tumors in mice, and reside in aberrant perivascular stem cell niches supported by endothelial cells.13 In these
models it has been demonstrated that antiangiogenic agents
disrupt the perivascular-stem cell niche and contribute to
the death of GSCs.13 Treatment of GSCs with bevacizumab
or metronomic chemotherapy also suppresses their tumorigenicity in animal models.13,14 A fourth hypothesis is that
antiangiogenic agents disrupt pro-angiogenic signaling
from bone marrow– derived myeloid cells thereby suppressing tumor vascularization.15 Angiogenesis is dependent on
diverse cell populations in the tumor microenvironment, including inflammatory cells of myeloid lineage like mast cells,
dendritic cells, eosinophils, neutrophils, and macrophages—
all sources of VEGF and other proangiogenic factors. VEGF
and the closely related placental growth factor (PlGF) are potent myeloid cell chemokines that attract VEGF receptor 1
(VEGFR1)⫹ monocytes to tumors. VEGFR1⫹ cells appear
to be important in sustaining glioblastoma angiogenesis.15
Preclinical experiments demonstrate that elimination of
VEGFR1 signaling in bone marrow– derived myeloid cells
decreases tumor angiogenesis and growth. Thus, the antitumor effects of antiangiogenic agents may be at least partially mediated by disruption of VEGFR1 signaling and
reduction of tumor-infıltrating VEGFR1⫹ monocytes. Resistance to antiangiogenic agents may be conferred by eventual infıltration of additional bone marrow–derived macrophages that promote aggressive mesenchymal features and
increased stem cell marker expression.16
VASOGENIC CEREBRAL EDEMA
VEGF, originally termed vascular permeability factor, increases the permeability of tumor vessels leading to vasogenic
brain edema. Primary and metastatic brain tumors are often
associated with vasogenic edema, which contributes to neurological morbidity. Vasogenic brain edema is a primary reason why a large proportion of patients with brain tumor
require treatment, often long term, with corticosteroids.
Clinical and radiographic studies of patients with glioblastoma treated with anti-VEGF agents demonstrate both antiedema and steroid-sparing effects.8,17 In addition, preclinical
studies of cediranib, a pan-VEGF receptor tyrosine kinase inhibitor, demonstrated potent antiedema effects in three different orthotopic murine glioma models.18 In these studies,
cediranib signifıcantly alleviated edema through rapid normalization of tumor vasculature. Of note, survival time was
prolonged (p ⬍ 0.05) with no change in tumor growth indicating that a potent antiedema effect may prolong survival
in some orthotopic glioma models.18 Thus, a salutary benefıt
of vascular normalization in patients with brain tumors is
ANTIANGIOGENIC THERAPY IN GLIOBLASTOMA: MECHANISMS OF ACTION, RESISTANCE, AND FUTURE CHALLENGES
reduction in cerebral edema and sparing of steroid-related
complications.
PSEUDOPROGRESSION
It is not uncommon to observe increased contrastenhancement and surrounding T2/FLAIR hyperintensity
within the radiation treatment fıeld on MRI scans in patients
with glioblastoma obtained within a few months of chemoradiation completion. While these radiographic fındings
raise the possibility of tumor progression, these changes may
also reflect the cytotoxic effect of chemoradiation on the tumor and the tumor microenvironment, typically referred to
as tumor “pseudoprogression.” The incidence of tumor pseudoprogression ranges from 28% to 66% in all patients with
glioblastoma undergoing chemoradiation and typically occurs within 3 months after completion of concurrent radiation and temozolomide.19 The radiographic fındings
typically consist of an increased area of contrast enhancement and enlargement of noncontrast T2/FLAIR hyperintense signal surrounding the enhancement. Approximately
one-third of patients are symptomatic from tumor pseudoprogression and may require treatment with corticosteroids.
Bevacizumab appears to be effıcacious in the treatment of
radiation-related brain necrosis but has not been adequately
studied or established as an effective therapy for symptomatic tumor pseudoprogression.20 There is some evidence
to suggest that concurrent treatment of newly diagnosed
patients with glioblastoma with chemoradiation and an inhibitor of VEGF may reduce the incidence of pseudoprogression.21 Reduction in the incidence of pseudoprogression may
be a secondary benefıt of the vascular normalizing effects
of anti-VEGF treatment, which may in turn spare patients
from neurological symptoms and steroid usage. This benefıt
may be of particular value among patients with large, unresectable tumors in order to permit better tolerance of chemoradiation.
RESISTANCE TO ANTIANGIOGENIC THERAPY
The radiographic response rate of antiangiogenic therapy in
glioblastoma is very high, but the duration of response on
average lasts approximately 4 months. “Vascular response,”
or “pseudoresponse,” referring to a reduction in contrast enhancement in the absence of tumor growth inhibition, can
occur rapidly, even within 24 hours of antiangiogenic therapy administration. The rapid responses and short duration
of benefıt from these therapies have prompted the suggestion
that there is little intrinsic antitumor activity of these agents
and that the main benefıt may be due to indirect effects attributable to reduced cerebral edema. In light of recent preliminary data failing to show an improvement in OS for
newly diagnosed patients treated with bevacizumab together
with the standard of care, all glioblastoma may be intrinsically resistant to these agents. This is not to say that antiangiogenic therapy might not be a useful tool for the treatment
of glioblastoma, but at this point, evidence for the ability of
these agents to augment survival is lacking. Here we provide
a counterpoint to the discussion above and describe several
potential mechanisms of resistance that may limit the effıcacy
of antiangiogenic therapy in glioblastoma.
THE BALANCE BETWEEN NORMALIZATION AND
EXCESSIVE VESSEL PRUNING
As described above, antiangiogenic therapy-mediated vascular normalization may normalize an otherwise chaotic tumor
vasculature to improve oxygenation, drug delivery, and radiation therapy effıcacy. As eloquently described by Jain and
colleagues this process requires a balance of pro and antiangiogenic factors.7 However, an overabundance of antiangiogenic factors, as may occur with excess antiangiogenic
therapy, may tip the balance toward a devascularized and hypoxic environment. Several animal models and human biopsy studies have identifıed increases in tumor hypoxia
following prolonged antiangiogenic therapy administration.22,23 As opposed to the benefıcial effects of oxygenation,
tumor hypoxia is a well-known mediator of resistance to chemotherapy and radiation and may contribute to resistance
mechanisms described below.
ANTIANGIOGENIC THERAPY MAY LIMIT DRUG
DELIVERY
There appears to be a complex interaction between the potential benefıt provided by vascular normalization and drug
delivery. Although some reports have demonstrated an improvement in radiation sensitivity following vascular normalization,7 these benefıts may be limited to the period
immediately following antiangiogenic drug dosing. Time
from initiation of drug administration as well as dose of antiangiogenic may limit the potential benefıt provided by the
vascular normalization window. For example, low-dose antiangiogenic therapy resulted in improved temozolomide delivery in an orthotopic glioma xenograft model, but highdose antiangiogenic therapy decreased delivery compared to
untreated controls.24 Translation of high- and low-dose antiangiogenic therapy to the human GBM experience has not
been explored and poses a challenge when designing clinical
trials to evaluate combination therapies. Furthermore, recent
data suggest that antiangiogenic therapy decreases delivery
of small molecules such as erlotinib to glioma xenograft tumors (personal communication, J Sarkaria). Implications of
these data to the neuro-oncology community are enormous.
Utilizing maximum tolerated dose as opposed to the optimal
biologic dose of antiangiogenic therapy is the approach currently used for patients. It is unclear if this is a “high” or “low”
dose and thus the effect of dose on drug delivery to patients
remains unknown. There are numerous completed and ongoing clinical trials combining antiangiogenic therapy with
chemotherapy and molecularly targeted agents. These studies have failed to demonstrate improved outcomes compared
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
e73
DE GROOT, REARDON, AND BATCHELOR
to the antiangiogenic agent alone.25 At this time, it is not
known if this is due to a lack of synergy between the combinations or if there is antagonism resulting from a reduction
in drug delivery following antiangiogenic dosing. Future
combination approaches should incorporate techniques to
ensure that intratumoral drug delivery is not impeded by
concurrent use of antiangiogenic therapy.
TUMOR VESSELS INSENSITIVE TO ANTIANGIOGENIC
THERAPY
Although small vessel pruning is one of the most prominent
and oft cited benefıts of antiangiogenic therapy, only certain
blood vessels within the tumor microenvironment are likely
to be affected. Recent studies suggest that only mother vessels
and glomeruloid microvascular proliferations contain
VEGFR2-expressing endothelial cells and are susceptible to
anti-VEGF therapies. Conversely, feeding arteries and draining veins are less sensitive to anti-VEGF strategies due to
lower expression of VEGF receptors.26 The histologic prominence of glomeruloid vascular proliferation and the rapid
reduction in contrast enhancement on MRI suggest that
VEGF plays a dominant role in permeable blood vessels in
glioblastoma. Although the vast majority of patients treated
with antiangiogenic therapies have some reduction in contrast enhancement, it is not clear to what degree these imaging changes correlate with a reduction in tumor vascularity.
Anti-VEGF therapy may not eliminate blood flow to tumor
via larger non-VEGF-dependent vascular structures.
MYELOID CELL-MEDIATED RESISTANCE
Although initially associated with reduced recruitment of
monocytes and macrophages, antiangiogenic therapy eventually promotes the recruitment of multiple myeloid cell
types. These cells are associated with diverse aspects of an
evasive tumor phenotype. It is known that M2-skewed tumor
associated macrophages contribute to angiogenesis, assist
with immune evasion, and promote tumor invasion.30 It is
important to note that macrophages account for only a fraction of bone marrow– derived cells recruited to resistant tumors. Elimination of macrophages in a glioma stem cell
xenograft model using a colony stimulating factor 1 receptor
inhibitor did not improve the effıcacy of antiangiogenic therapy alone and promoted resistance consistent with described
mechanisms of resistance common to many antiangiogenic
therapies (unpublished data, J. de Groot). In fact, attraction
of macrophages to tumor may eliminate regions of hypoxia
and improve sensitivity to therapy31 highlighting the challenges of inhibiting just one myeloid cell type in the complex
tumor microenvironment. Equally or more important, neutrophils, Tie-2 expressing monocytes, and myeloid-derived
suppressor cells all contribute to a highly resistant tumor
through complex microenvironment interactions between
tumor, endothelial cells and the microenvironment. Several
studies have indicated the importance of targeting neutrophils in antiangiogenic therapy resistance. In animal models,
both the JAK/STAT inhibitor AZD148032 and an inhibitor
of Bv833 in combination with antiangiogenic therapy reduced infıltration of neutrophils and led to signifıcant prolongation of animal survival and reduced tumor size.
VEGF-INDEPENDENT ANGIOGENESIS
Many groups have demonstrated that inhibition of VEGF-A
in glioma tumor models promotes both tumor and stromalmediated VEGF-A-independent angiogenesis. A plethora of
pro-angiogenic factors have been shown to increase including VEGF-C, VEGF-D, PlGF, PDGF, basic fıbroblast growth
factor, and others.27 It is unclear how many human glioblastomas switch to VEGF-A-independent angiogenesis. Interestingly, only a fraction of tumors progressing during
treatment with an antiangiogenic agent develop new contrast
enhancement and frequently this contrast enhancement is
atypical. MRI images display punctate and scattered enhancement as opposed to the typical ring-enhancement of
glioblastoma. Analysis of tumor cells biopsied from enhancing areas of human glioblastoma progressing on bevacizumab revealed a distinct set of upregulated genes, an
increase in tumor cell proliferation but no changes in tumor
vascularity compared to their pretreated samples28 suggesting that bevacizumab may not reduce tumor vasculature in
all tumors. Importantly, tumor cells requiring oxygen and
nutrients may grow and migrate along (co-opt) normal blood
vessels within the brain.29 This juxtaposition of tumor cells
along the normal vasculature may obviate the tumor’s need
for neoangiogenesis for survival and may provide a route for
tumor infıltration to distant areas of the brain protected by
the blood brain barrier, as described in more detail below.
e74
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
AUTOPHAGY-MEDIATED RESISTANCE
In addition to direct effects on tumor vasculature, antiangiogenic therapy has been shown to minimally induce glioma
cell apoptosis in animal models.18,34 These indirect effects are
small, but glioblastoma may utilize autophagy to prevent cell
death. Glioblastoma are inherently hypoxic, which may be
further exacerbated by antiangiogenic therapy. Glioma cells
may utilize autophagy to confer a survival advantage when
exposed to this hypoxic microenvironment.35,36 Additionally, resistance to chemotherapy may in part be mediated by
autophagy37 leading to a more resistant phenotype consistent
with the absence of effıcacy of standard therapies. Attempts
to overcome pro-survival autophagic pathway activation will
require a more detailed understanding of target protein for
inhibition as well as the overall consequences of autophagy
inhibition.
MESENCHYMAL AND INVASIVE TRANSFORMATION
Finally, there is the concern that antiangiogenic therapy
promotes tumor invasion. In animal models, invasion of
glioma cells during antiangiogenic therapy is thought to occur through vessel co-option, which has been demonstrated
in human and animal histologic studies.23,29 Attempts to
ANTIANGIOGENIC THERAPY IN GLIOBLASTOMA: MECHANISMS OF ACTION, RESISTANCE, AND FUTURE CHALLENGES
overcome this invasive phenotype in preclinical models have
been variably successful.34,38 However, the effect of this phenotypic shift in human disease remains controversial.39 At
the heart of this debate is the fact that glioblastomas are
highly invasive in the absence of antiangiogenic therapy and
all glioblastomas ultimately progress on standard and experimental therapy. At a minimum, inhibition of vascular permeability with anti-VEGF therapies leads to a decoupling
between contrast enhancement visualized on MRI and its expected (nonenhancing) tumor growth/infıltration. This phenomenon occurs in a third or fewer cases and results in
clinical progression in the absence of worsening enhancing
tumor burden.
For at least a decade, the proto-oncogene MET has been
implicated in resistance to antiangiogenic therapy.40 Some
groups have identifıed activation of this pathway in glioblastoma,41 whereas others have not,16,42 possibly reflecting the
complexity of the disease, differences in animal models, and
heterogeneous tumor cell responses to therapy. Additionally,
the spectrum of molecular alterations associated with resistance to antiangiogenic therapy includes the induction of
mesenchymal genes,16 which are known to promote invasion
in glioblastoma, a proneural to mesenchymal phenotype,
epithelial-mesenchymal transition, and metastasis in other
solid tumors. An increase in stem cells16 and extracellular
matrix proteins28,34 may also reflect selection for highly resistant cells, which may contribute to therapeutic resistance.
Resistance of antiangiogenic therapy-treated tumors remains a formidable problem and will require new approaches
to elucidate and overcome these complex and likely interwoven contributing mechanisms.
CHALLENGES FOR THE FUTURE
With the exception of temozolomide, evaluation of a wide
array of therapeutics for patients with glioblastoma over the
past few decades has demonstrated, at best, short-term benefıt for a small minority of patients. In contrast, most patients
with glioblastoma derive some benefıt from bevacizumab,
and to varying degrees, other VEGF/VEGFR-targeting therapeutics. The durability of clinical benefıt for most patients is
limited. These results suggest that mechanisms of resistance
represent a major challenge for antiangiogenic therapies.
Furthermore, many fundamental questions regarding the
application of these agents for patients with glioblastoma remain unanswered. Answers to these questions may hold the
key to overcoming current limitations and future challenges.
Prolifıc blood vessel formation is a characteristic feature
of glioblastoma, and several pathophysiologic mechanisms,
including vascular co-option, angiogenesis, vasculogenesis,
vascular mimicry, and glioblastoma-endothelial cell transdifferentiation, contribute to this process. Paradoxically,
glioblastoma vasculature is dysfunctional due to ultrastructural, functional, and biochemical abnormalities leading
to poor blood flow and increased vessel permeability, which
in turn contribute to elevated interstitial pressure and tissue hypoxia/acidosis. Furthermore, glioblastoma tumors are
intrinsically infıltrative and capable of adapting to hypoxic/
acidotic microenvironments. In these contexts, it is unclear
how truly dependent glioblastoma tumors are on angiogenesis and, beyond that, how dependent they are on VEGF.
From this perspective, a survival benefıt associated with
VEGF inhibitor therapy would be unexpected and many
would predict intrinsic resistance to be the norm for glioblastoma tumors. Nonetheless, cumulative clinical data suggests
that patients with glioblastoma derive benefıt from VEGF/
VEGFR inhibitors. A major clinical challenge is therefore
how to effectively prove and measure these clinical benefıts
(net clinical benefıt).
Several fundamental questions regarding antiangiogenic
agents for glioblastoma remain unanswered. A basic unknown is their underlying mechanism of action and, in
particular, whether they elicit bona fıde antitumor activity.
The antipermeability capability of VEGF/VEGFR inhibitors,
with their potential to decrease tumor-associated edema and
mass effect, is undisputed. The critical question however remains, what is happening to the underlying tumor cells? Are
they being starved and shifted toward apoptosis and diminished proliferation as indicated by some preclinical studies?43
Are glioblastoma stem cells preferentially targeted? Or does
blocking VEGF diminish immunosuppression that may in
turn enhance antitumor immune responses?44
DETERMINING THE OPTIMAL BIOLOGIC DOSE
As discussed previously, one proposal is that judicious dosing of antiangiogenic agents may achieve a “sweet spot” to
effectively normalize aberrant tumor vasculature. Specifıcally, optimally dosed antiangiogenic agents may prune abnormal tumor vasculature in a limited manner while
avoiding vessel obliteration associated with more aggressive
dosing. The former process can enhance intratumoral blood
flow and improve effıcacy of coadministered therapeutics,7
while the latter process may detrimentally effect delivery and
increase hypoxia/acidosis in the tumor microenvironment.
Advanced MRI studies among recurrent patients with glioblastoma following cediranib administration support the
vascular normalization mechanism.8 Nonetheless, if the normalization mechanism is truly relevant for glioblastoma, a
critical question then becomes dosing schedules. Few studies
have included alternative dosing schedules of antiangiogenic
agents, but none has formally tested less intensive compared
with more intensive dosing. A recent report demonstrates
that bevacizumab decreases radiolabeled docetaxel delivery
among patients with lung cancer.10 Whether antiangiogenics
can limit intratumoral delivery of concurrently administered
therapeutics among patients with glioblastoma remains to be
studied. Before initiating expensive and time-consuming
clinical trials, efforts are needed to determine the optimal
biologic dose that improves drug delivery and oxygenation.
Further mechanistic-based studies of antiangiogenic agents
and their effect on delivery are clearly needed for patients
with malignant glioma.
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
e75
DE GROOT, REARDON, AND BATCHELOR
IDENTIFICATION OF PATIENTS THAT BENEFIT
Another unanswered question is whether informative biomarkers can predict outcome for antiangiogenic therapy.
This defıciency remains a shortcoming of antiangiogenic
agents across oncology,45 although recent studies suggest
that circulating levels of the short isoforms of VEGF-A may
correlate with outcome.46 Validated biomarkers could potentially serve two key roles. First, they may identify patients
more likely to benefıt, for whom antiangiogenic agents
should be prioritized, or conversely, those patients who are
less likely to respond, for whom alternative agents should
be considered. Second, biomarkers offer the potential for
early determination of progressive tumor among initially
responding patients. The latter role offers particular value
given current limitations associated with routinely used MRI
response assessment methods as discussed below. A wide
array of potential biomarkers have demonstrated encouraging early data among patients with glioblastoma, but none
have been validated, including advanced MRI applications,
PET strategies, measurement of plasma markers or circulating endothelial cells, and the measurement of hypoxia
markers from archival tumor samples.45 One intuitive and
easily assessed biomarker appears to be early radiographic
response.47
Along these lines, another important question is whether
antiangiogenic agents may effect subsets of patients with glioblastoma differently. Glioblastoma tumors are currently defıned by a spectrum of genetic abnormalities and gene
microarray expression studies identify three to four distinct
subclasses. Linked expression of mesenchymal and angiogenic genes has been associated with poor prognosis. Future
studies may therefore consider patient stratifıcation or enrichment based on genetic characterization.
ESTABLISHING THE CLINICAL BENEFIT OF
ANTIANGIOGENIC THERAPY
Benefıt from antiangiogenic therapy varies widely across
cancers. Among responsive tumors, antiangiogenics consistently achieve greater PFS increments and less prominent, or
in some cases, lack of OS benefıt. A similar pattern has
emerged for glioblastoma. Specifıcally, among patients with
recurrent glioblastoma, the BRAIN study as well as several
additional trials, reported PFS-6 rates of approximately 40%,
compared to 10% to 15% achieved among historic cohorts.17,25 In contrast, median OS on the BRAIN study was
eight to 10 months compared to six and seven months reported in historic data. Preliminary results of the AVAglio
study, a placebo-controlled, randomized phase III evaluation of
bevacizumab for patients with newly diagnosed glioblastoma
confırmed similar fındings, although mature follow-up is
lacking and the effect of cross-over by control patients to receive bevacizumab at progression is unclear.48 Specifıcally,
median investigator-reported PFS for the bevacizumab and
placebo patients were 10.6 and 6.2 months, respectively (p ⬍
0.0001; hazard ratio ⫽ 0.64), while OS at 1-year was 72%
compared with 66%, respectively (p ⫽ 0.052).
e76
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
These data raise a critical unanswered question regarding
the value of various metrics of outcome. Specifıcally, is there
a meaningful value of prolonging PFS independent of OS?
Although prolonging OS remains the gold standard, PFS is
more controversial, particularly if not clearly validated to
predict OS. For most cancers, it is not unreasonable to question whether therapies that prolong PFS are of substantive
merit if OS is not prolonged. However, the effect of underlying tumor progression is not the same for all cancers. For
example, while progressive enlargement of a metastatic lung
nodule or a new hepatic lesion may bode for poor overall outcome, in general these fındings usually harbor minimal acute
effect for patients with solid tumor. In contrast, progression
of inherently infıltrative and destructive central nervous system tumors like glioblastoma typically translates into worsened/new neurologic defıcits and chronic corticosteroid
dependence, both of which erode quality of life.
Thus, a critical consideration in assessing the value of prolonging PFS, is the effective application of validated parameters to reliably assess neurologic function and quality of life.
In a single-center, retrospective series, patients with recurrent glioblastoma who received bevacizumab had a nearly
twofold higher independent living score, compared to patients treated without bevacizumab.49 Similarly, preliminary
AVAglio study results demonstrate that bevacizumab recipients reported higher scores on fıve predetermined quality of
life parameters and maintained a Karnofsky Performance
Status of 70 or higher compared to placebo controls. Another
important correlate of overall well-being among neurooncology patients is dependence on chronic corticosteroid
dosing to reduce symptoms from tumor-associated cerebral
edema and mass effect. The sequelae of chronic corticosteroid dependence on quality of life are substantial including
incapacitating proximal myopathy, osteoporosis and pathologic fracture, marked truncal weight gain, hypertension, and
hyperglycemia. Of note, both the BRAIN and AVAglio studies reported reduced corticosteroid dosing.17,48
A wide array of metrics to assess quality of life are available
for future glioblastoma studies including measures of overall
performance, standardized neurocognitive evaluations and
validated self-reporting quality of life and symptom inventories adapted for neuro-oncology patients. In addition, an objective scale of neurologic function for neuro-oncology
patients is being developed and will integrate into the RANO
criteria (personal communication, D. Reardon).
The utility of assessing neurologic function and other readouts reflecting quality of life, is further heightened by the conundrum frequently encountered by clinicians attempting to
radiologically assess response among patients with malignant glioma undergoing antiangiogenic therapy. Traditional
MRI response assessment based on enhancing tumor frequently provides an unreliable surrogate of underlying
tumor activity in this setting and was one of the important
bases for the delineation of the RANO criteria.50 Nonetheless, a signifıcant challenge for our neuroradiology colleagues
is further refınement of advanced imaging techniques to
ANTIANGIOGENIC THERAPY IN GLIOBLASTOMA: MECHANISMS OF ACTION, RESISTANCE, AND FUTURE CHALLENGES
more reliably assess underlying tumor activity following
antiangiogenic therapy.
NEW SALVAGE THERAPIES NEEDED IN THE ERA OF
ANTIANGIOGENIC THERAPY
Finally, given that all patients with malignant glioma ultimately progress following bevacizumab, a major unanswered
question is defıning effective therapies for patients with
bevacizumab-resistant disease. Outcome for such patients is
dismal with no effective therapy currently defıned. As discussed previously, better understanding of factors underlying acquired resistance to VEGF/VEGFR therapies may
provide critical insight to design effective therapies for these
patients. In the meantime, one approach frequently used in
daily practice is bevacizumab continuation beyond initial
progression, usually in combination with a new or different
chemotherapeutic. A recently reported phase III study validated the benefıt of this approach among patients with colorectal cancer.51 A recent retrospective analysis among
patients with recurrent glioblastoma demonstrated that bevacizumab continuation beyond initial progression modestly
improved PFS and OS over nonbevacizumab therapies.52 A
prospective study to evaluate bevacizumab continuation beyond initial progression is being planned for patients with
glioblastoma. Importantly, innovative therapeutics focusing
on novel malignant glioma targets remain critically needed.
CONCLUSION
Targeting angiogenesis has been a prominent focus of clinical research for glioblastoma over the past fıve years. Although some answers have been obtained, the cumulative
knowledge gained from myriad conducted clinical trials and
associated efforts have failed to resolve many fundamental
questions. New data suggesting that antiangiogenic therapy
may not prolong overall patient survival when added to the
standard of care, raises multiple questions regarding the true
benefıt of these drugs for our patients. Future studies should
prioritize addressing specifıc knowledge gaps regarding
mechanism of action, identifıcation of patient subsets more
likely to derive durable benefıt, informative biomarkers, the
value of alternative outcome endpoints, and effective therapy
to delay or overcome resistance.
Disclosures of Potential Conflicts of Interest
Relationships are considered self-held and compensated unless otherwise noted. Relationships marked “L” indicate leadership positions. Relationships marked “I” are those held by an immediate
family member; those marked “B” are held by the author and an immediate family member. Relationships marked “U” are uncompensated.
Employment or Leadership Position: None. Consultant or Advisory Role: Tracy Batchelor, Genentech; Kirin. John F. De Groot, Genentech; VBL
Therapeutics. David A. Reardon, EMD Serono; Genentech; Merck/Schering. Stock Ownership: None. Honoraria: Tracy Batchelor, Advance Medical; Champions
Biotechnology; Merck; Oakstone Publishing; Research to Practice; UpToDate. John F. De Groot, Merck. Research Funding: Tracy Batchelor, AstraZeneca;
Millennium; Pfizer. John F. De Groot, AstraZeneca; EMD-Serono; Sanofi. Expert Testimony: None. Other Remuneration: None.
References
1. Chi AS, Sorensen AG, Jain RK, et al. Angiogenesis as a therapeutic target
in malignant gliomas. Oncologist. 2009;14:621-636.
2. Brastianos PK, Batchelor TT. VEGF inhibitors in brain tumors. Clin
Adv Hematol Oncol. 7:753-. 2009;60:768.
3. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med.
1971;285:1182-1186.
4. Duda DG, Jain RK, Willett CG. Antiangiogenics: the potential role of
integrating this novel treatment modality with chemoradiation for solid
cancers. J Clin Oncol. 2007;25:4033-4042.
5. Shaked Y, Kerbel RS. Antiangiogenic strategies on defense: on the possibility of blocking rebounds by the tumor vasculature after chemotherapy. Cancer Res. 2007;67:7055-7058.
6. Gorski DH, Beckett MA, Jaskowiak NT, et al. Blockage of the vascular
endothelial growth factor stress response increases the antitumor effects
of ionizing radiation. Cancer Res. 1999;59:3374-3378.
7. Jain RK. Normalization of tumor vasculature: an emerging concept in
antiangiogenic therapy. Science. 2005;307:58-62.
8. Batchelor TT, Sorensen AG, di Tomaso E, et al. AZD2171, a panVEGF receptor tyrosine kinase inhibitor, normalizes tumor vasculature and alleviates edema in glioblastoma patients. Cancer Cell. 2007;11:
83-95.
9. Winkler F, Kozin SV, Tong RT, et al. Kinetics of vascular normalization
by VEGFR2 blockade governs brain tumor response to radiation: role of
oxygenation, angiopoietin-1, and matrix metalloproteinases. Cancer
Cell. 2004;6:553-563.
10. Van der Veldt AA, Lubberink M, Bahce I, et al. Rapid decrease in delivery of chemotherapy to tumors after anti-VEGF therapy: implications
for scheduling of anti-angiogenic drugs. Cancer Cell. 2012;21:82-91.
11. Sorensen AG, Emblem KE, Polaskova P, et al. Increased survival of glioblastoma patients who respond to antiangiogenic therapy with elevated blood perfusion. Cancer Res. 2012;72:402-407.
12. Gerstner ER, Emblem KE, Chi AS, et al. Effects of cediranib, a VEGF
signaling inhibitor, in combination with chemoradiation on tumor
blood flow and survival in newly diagnosed glioblastoma. J Clin Oncol.
2012;30 (suppl; abstr 2009).
13. Calabrese C, Poppleton H, Kocak M, et al. A perivascular niche for brain
tumor stem cells. Cancer Cell. 2007;11:69-82.
14. Folkins C, Man S, Xu P, et al. Anticancer therapies combining antiangiogenic and tumor cell cytotoxic effects reduce the tumor stem-like cell
fraction in glioma xenograft tumors. Cancer Res. 2007;67:3560-3564.
15. de Groot JF, Piao Y, Tran H, et al. Myeloid biomarkers associated with
asco.org/edbook | 2013 ASCO EDUCATIONAL BOOK
e77
DE GROOT, REARDON, AND BATCHELOR
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
e78
glioblastoma response to anti-VEGF therapy with aflibercept. Clin Cancer Res. 2011;17:4872-4881.
Piao Y, Liang J, Holmes L, et al. Glioblastoma resistance to anti-VEGF
therapy is associated with myeloid cell infıltration, stem cell accumulation, and a mesenchymal phenotype. Neuro Oncol. 2012;14:1379-1392.
Friedman HS, Prados MD, Wen PY, et al. Bevacizumab alone and in
combination with irinotecan in recurrent glioblastoma. J Clin Oncol.
2009;27:4733-4740.
Kamoun WS, Ley CD, Farrar CT, et al. Edema control by cediranib, a
vascular endothelial growth factor receptor-targeted kinase inhibitor,
prolongs survival despite persistent brain tumor growth in mice. J Clin
Oncol. 2009;27:2542-2552.
Fink J, Born D, Chamberlain MC. Pseudoprogression: relevance with
respect to treatment of high-grade gliomas. Curr Treat Options Oncol.
2011;12:240-252.
Levin VA, Bidaut L, Hou P, et al. Randomized double-blind placebocontrolled trial of bevacizumab therapy for radiation necrosis of the
central nervous system. Int J Radiat Oncol Biol Phys. 2011;79:1487-1495.
Pinho M, Polaskova P, Jennings D, et al. Impact of adjuvant anti-VEGF
therapy on treatment-related pseudoprogression in patients with newly
diagnosed glioblastoma receiving chemoradiation with or without antiVEGF therapy. J Clin Oncol. 2012;30 (suppl; abstr 2025).
Du R, Lu KV, Petritsch C, et al. HIF1alpha induces the recruitment of
bone marrow-derived vascular modulatory cells to regulate tumor angiogenesis and invasion. Cancer Cell. 2008;13:206-220.
de Groot JF, Fuller G, Kumar AJ, et al. Tumor invasion after treatment
of glioblastoma with bevacizumab: radiographic and pathologic correlation in humans and mice. Neuro Oncol. 2010;12:233-242.
Zhou Q, Gallo JM. Differential effect of sunitinib on the distribution of
temozolomide in an orthotopic glioma model. Neuro Oncol. 2009;11:
301-310.
Reardon DA, Turner S, Peters KB, et al. A review of VEGF/VEGFRtargeted therapeutics for recurrent glioblastoma. J Natl Compr Canc
Netw. 2011;9:414-427.
Sitohy B, Nagy JA, Jaminet SC, et al. Tumor-surrogate blood vessel subtypes exhibit differential susceptibility to anti-VEGF therapy. Cancer
Res. 2011;71:7021-7028.
Bergers G, Hanahan D. Modes of resistance to anti-angiogenic therapy.
Nat Rev Cancer. 2008;8:592-603.
DeLay M, Jahangiri A, Carbonell WS, et al. Microarray analysis verifıes
two distinct phenotypes of glioblastomas resistant to antiangiogenic
therapy. Clin Cancer Res. 2012;18:2930-2942.
Holash J, Davis S, Papadopoulos N, et al. VEGF-Trap: a VEGF blocker
with potent antitumor effects. Proc Natl Acad Sci U S A. 2002;99:1139311398.
Murdoch C, Muthana M, Coffelt SB, et al. The role of myeloid cells in
the promotion of tumour angiogenesis. Nat Rev Cancer. 2008;8:618631.
Sasajima J, Mizukami Y, Sugiyama Y, et al. Transplanting normal vascular proangiogenic cells to tumor-bearing mice triggers vascular remodeling and reduces hypoxia in tumors. Cancer Res. 2010;70:62836292.
de Groot J, Liang J, Kong LY, et al. Modulating antiangiogenic resistance
by inhibiting the signal transducer and activator of transcription 3 pathway in glioblastoma. Oncotarget. 2012;3:1036-1048.
Shojaei F, Wu X, Qu X, et al. G-CSF-initiated myeloid cell mobilization
and angiogenesis mediate tumor refractoriness to anti-VEGF therapy in
mouse models. Proc Natl Acad Sci U S A. 2009;106:6742-6747.
2013 ASCO EDUCATIONAL BOOK | asco.org/edbook
34. Lucio-Eterovic AK, Piao Y, de Groot JF. Mediators of glioblastoma resistance and invasion during antivascular endothelial growth factor
therapy. Clin Cancer Res. 2009;15:4589-4599.
35. Lisanti MP, Martinez-Outschoorn UE, Chiavarina B, et al. Understanding the “lethal” drivers of tumor-stroma co-evolution: emerging role(s)
for hypoxia, oxidative stress and autophagy/mitophagy in the tumor
micro-environment. Cancer Biol Ther. 2010;10:537-542.
36. Hu YL, DeLay M, Jahangiri A, et al. Hypoxia-induced autophagy promotes tumor cell survival and adaptation to antiangiogenic treatment in
glioblastoma. Cancer Res. 2012;72:1773-1783.
37. White E. Deconvoluting the context-dependent role for autophagy in
cancer. Nat Rev Cancer. 2012;12:401-410.
38. Lu KV, Chang JP, Parachoniak CA, et al. VEGF inhibits tumor cell invasion and mesenchymal transition through a MET/VEGFR2 complex.
Cancer Cell. 2012;22:21-35.
39. Chamberlain MC, Lassman AB, Iwamoto FM. Patterns of relapse and
prognosis after bevacizumab failure in recurrent glioblastoma. Neurology. 2010;74:1239-1241.
40. Pennacchietti S, Michieli P, Galluzzo M, et al. Hypoxia promotes invasive growth by transcriptional activation of the met protooncogene.
Cancer Cell. 2003;3:347-361.
41. Jahangiri A, De Lay M, Miller LM, et al. Gene expression profıle identifıes tyrosine kinase c-Met as a targetable mediator of anti-angiogenic
therapy resistance. Clin Cancer Res. Epub 2013 Mar.
42. Keunen O, Johansson M, Oudin A, et al. Anti-VEGF treatment reduces
blood supply and increases tumor cell invasion in glioblastoma. Proc
Natl Acad Sci U S A. 2011;108:3749-3754.
43. Kunkel P, Ulbricht U, Bohlen P, et al. Inhibition of glioma angiogenesis
and growth in vivo by systemic treatment with a monoclonal antibody
against vascular endothelial growth factor receptor-2. Cancer Res. 2001;
61:6624-6628.
44. Johnson BF, Clay TM, Hobeika AC, et al. Vascular endothelial growth
factor and immunosuppression in cancer: Current knowledge and potential for new therapy. Expert Opin. Biol Ther. 2007;7:449-460.
45. Jain RK, Duda DG, Willett CG, et al. Biomarkers of response and resistance to antiangiogenic therapy. Nat Rev Clin Oncol. 2009;6:327-338.
46. Lambrechts D, Lenz HJ, de Haas S, et al. Markers of response for the
antiangiogenic agent bevacizumab. J Clin Oncol. 2013;31:1219-1230.
47. Prados M, Cloughesy T, Samant M, et al. Response as a predictor of
survival in patients with recurrent glioblastoma treated with bevacizumab. Neuro-oncology. 2011;13:143-151.
48. Chinot O, Wick W, Mason W, et al. Phase III trial of bevacizumab added
to standard radiotherapy and temozolomide for newly diagnosed glioblastoma: fınal progression-free survival and interim overall survival results in AVAglio, in Yung AW (ed): Society for Neuro-Oncology.
Wahsington, D.C., 2012.
49. Nagpal S, Harsh G, Recht L. Bevacizumab improves quality of life in
patients with recurrent glioblastoma. Chemother Res Pract. 2011:
602812.
50. Wen PY, Macdonald DR, Reardon DA, et al. Updated response assessment criteria for high-grade gliomas: response assessment in neurooncology working group. J Clin Oncol. 2010;28:1963-1972.
51. Bennouna J, Sastre J, Arnold D, et al. Continuation of bevacizumab after
fırst progression in metastatic colorectal cancer (ML18147): a randomised phase 3 trial. Lancet Oncol. 2013;14:29-37.
52. Reardon DA, Herndon JE, 2nd, Peters KB, et al. Bevacizumab continuation beyond initial bevacizumab progression among recurrent glioblastoma patients. Br J Cancer. 2012;107:1481-1487.