Results
Clinical presentation in diffuse low-grade glioma
Epileptic seizures are the most common presentation, i.e.,72%–89% and range from simple-to-complex
seizures with or without secondary generalization.[[1]],[[2]] Majority of the presenting seizures are resistant to medical management[[3]] and affect the patients' quality of life and cognitive capacity and may cause other
complications.[[4]],[[5]] The seizures at presentation in dLGG patients define the postoperative potential
of seizure continuation and are relevant to prognostication.[[6]] In neurologically intact patients, presentation with seizures is associated with
better prognosis.[[7]],[[8]],[[9]],[[10]] In the largest retrospective series of patients published by Pallud et al. 2013,
seizures at diagnosis along with complete resection of tumor were found to be independent
predictors of, malignant progression-free survival (PFS) and overall survival (OS)
of patients.[[10]]
Multiple factors predispose dLGG patients to the risk of epileptic activity. In such
patients, seizures at presentation are commonly seen in patients younger than 60 years.[[11]] The tumors' tendency for highly epileptogenic areas of mesiotemporal and insular
cortex in addition to frontal, temporal, and parietal lobes explains most of the presentation.
Higher frequencies of seizures have been reported with tumors in proximity to the
central sulcus.[[12]] Similarly, oligodendroglioma and oligoastrocytoma which involve the cortex in majority
of the cases have a higher susceptibility to epileptic activity compared to astrocytoma,
found predominantly in the white matter.[[13]] Preventing seizures is an important aspect of the multidisciplinary approach in
managing dLGG.
dLGGs represent 15% of gliomas and have the highest incidence in the age group of
35–44 years with increased prevalence in Caucasian males.[[14]] In cases where seizures are not the initial presentation, progressive tumor growth
and infiltration of the adjacent parenchyma leads to neurocognitive changes, bringing
the patient to attention due to alteration in mental status (3-30%)[[15]],[[16]],[[17]] and rarely focal neurological deficits (2%–30%)[[15]],[[16]],[[18]] or even raised intracranial pressure.[[15]],[[16]] In up to 10% of the cases, the diagnosis is incidental, and the tumor is discovered
as a consequence of imaging for head trauma or other unrelated neurological conditions.[[19]]
Imaging of diffuse low-grade glioma
Magnetic Resonance (MR) is the standard of care in dLGG imaging. Though they are by
nature diffuse and ill defined, the lesions may be deceptively well circumscribed.
dLGG are hyper-intense on T2W images and hypo-intense on T1W images.[[20]] Fluid-attenuated inversion recovery (FLAIR) tends to show a larger area of signal
abnormality than standard T2W images. The extent of the lesion on FLAIR images tends
to better correlate with the adequacy of resection margins and is also the most useful
in the follow-up of surgical patients to detect recurrence postresection.[[21]] Contrast enhancement is generally absent but faint and patchy enhancement can be
identified up to 60% of the dLGGs.[[22]] Among the contrast-enhancing dLGGs, majority are identified as oligodendroglioma.[[23]]
The newer MR sequences and techniques have been widely applied to dLGGs in attempts
to predict tumor grade and biological behavior.[[24]] Magnetic resonance spectroscopy shows a relative preservation of N-acetyl aspartate
with an elevated choline peak and without lipid/lactate peaks.
There is however significant overlap between dLGGs and other neoplastic and nonneoplastic
lesions.[[25]]
Tumor histological sampling for examination and genotyping remains the gold standard
for diagnosing dLGGs.[[26]] Apparent diffusion coefficient (ADC) measurements have a high accuracy in differentiating
high-grade from LGGs, but however suffer from lack of standardized criteria and methodology.[[27]] Diffusion tensor imaging (DTI) has been used both in surgical planning[[28]] and in trying to determine tumor grade using diffusivity parameters.[[29]],[[30]] MR perfusion studies reveal near-normal relative cerebral blood volume (rCBV) in
LGGs. Changes in rCBV have been used to monitor changes in biological behavior of
these lesions.[[31]] Susceptibility-weighted imaging is exquisitely sensitive to the presence of intra-tumoral
hemorrhages, the presence of which tends to mitigate against the lesion being a dLGG.
Progression of the dLGG to a higher grade lesion is associated with changes in the
MR spectrum, perfusion parameters, as well as ADC values, and these studies are useful
in follow-up over time.[[20]]
On CT, the lesion is usually iso-dense to white matter on noncontrast images and does
not demonstrate any significant postcontrast enhancement. Calcifications can be identified
in the lesions in up to 20% of the cases, and the calcifications are particularly
indicative of oligodendrogliomas.[[23]] Hemorrhage is rarely if ever present.
Natural history of diffuse low-grade glioma
Though dLGG is generally considered a slow-growing benign natured tumor, clinical
studies suggest that these are progressive neoplastic lesions, and over 70% will transform
to an anaplastic glioma variant or a secondary glioblastoma within a decade. However,
there is considerable variation and unpredictability in the growth potential, and
the studies of the Montpellier group allow consideration of the patients in three
stages of disease: (i) a presymptomatic stage of unknown duration evidenced by the
tumors discovered incidentally,[[32]] (ii) a symptomatic period of about 7 years after the initial presentation (usually
a seizure) during which the patient maybe fully functional if the seizures are controlled,[[33]] and (iii) a transformational stage of 2–3 years of more rapid and clinically overt
tumor progression associated with induction of anaplasia.[[34]] Large series of dLGG assessed on serial MRI scans prior to treatment show that
they grow continuously at a rate of about 4 mm/year.[[32]] Steady growth at varying rates can be seen in incidentally discovered tumors and
in symptomatic ones by serial imaging using appropriate analytic software to allow
objective measurement of velocity of tumor diameter expansion. There is no such thing
as a static or stable dLGG, and there is an inverse relation between growth rate and
survival.
Besides increase in tumor diameter, adverse prognostic features at presentation are
patients' age over 40 years, nonseizure symptoms, neurological deficits, larger tumor
size particularly in excess of 6 cm diameter, tumor crossing the midline, and dominance
of astrocytic cellular features in tumor histology.[[34]]
Of the three histological varieties of dLGG, oligodendrogliomas have a better prognostic
outlook than astrocytomas with oligoastrocytomas in between though the latter diagnosis
is being altered by application of molecular testing for gene mutation status.
Molecular features in low-grade gliomas
There have been significant advances in the knowledge about the molecular biology
of gliomas along with the development of a series of biomarkers allowing improved
diagnosis, prognosis, and prediction of response to treatment. The most valuable of
these in practice are the mutation status of the genes regulating the Krebs cycle
enzymes, isocitrate dehydrogenase 1 and 2 (IDH1/2), whole-arm co-deletion of chromosome
arms 1p and 19q, alterations in tumor protein 53 (TP53), alpha thalassemia/mental
retardation syndrome X-linked (ATRX), and telomerase reverse transcriptase (TERT).
Mutations of IDH1 are encountered more often than IDH2 and occur in 65%–80% of gliomas.
Co-deletion of 1p and 19q is characteristic of oligodendroglial tumors. Point mutations
in IDH1/2 and co-deletion of 1p19q delineate subsets of dLGG with distinct biology
and clinical behavior.
IDH1 mutation at codon 132 is noted in >70% of WHO Grade II and III gliomas,[[35]] whereas IDH2 mutation is noticed in up to 6% of them.[[36]] Deletion of 1p36 is reported in up to 18% of astrocytomas, and 73% of oligodendrogliomas,
while the deletion of 19q13.3 is reported in 38% of astrocytomas and 73% of oligodendrogliomas.[[36]] Their co-deletion is noted in up to 11% of astrocytomas and 64% of oligodendrogliomas.
TP53 has been reported as genetic hallmark of low-grade astrocytomas, present in more
than 60% of the cases[[37]] and occurring only rarely in oligodendrogliomas. 1p19q and IDH 1and 2 mutations
are associated with prolonged survival and better response to chemotherapy (see below).
ATRX inactivation is also associated with astrocytic tumors, particularly IDH mutated
gliomas (86%) and may represent a subset of astrocytomas with improved treatment outcomes.[[38]] Mutations of the TERT promoter is one of the most common molecular markers in gliomas
and is to be found in more than 90% of IDH mutant and 1p/19q co-deleted oligodendrogliomas.[[39]]
Classification: Integrating histologic and genetic parameters
In the previous edition (2007) of the WHO Classification of Tumors of the Central
Nervous System, microscopic features were used for cytogenetic typing and malignancy
grading. The new edition of 2016 is a major conceptual change from the past, in including
molecular biomarkers to the traditional histopathologic features in the categorization
of individual dLGGs. These are now grouped together irrespective of origin from astrocytes
or oligodendrocytes as they share identifiable genetic mutations and prognostic features.
The diagnosis of gliomas by integration of histologic and genotypic features in the
new classification greatly improves diagnostic accuracy by reducing the interobserver
variation of traditional histology but is dependent on the availability of genomic
assays. Tumors not assignable to one of the designated tumor categories or not being
subjected to molecular testing are labeled not otherwise specified [[Figure 1]].
Figure 1: Layered diagnosis of low-grade glioma according to the WHO classification
The Cancer Genome Atlas has performed an analysis of 293 untreated dLGG and has been
able to define three molecular diagnostic classes more accurately than histological
class.[[40]] The first type was characterized by IDH mutations and 1p/19q co-deletion and showed
a strong association with oligodendroglioma histology. Other findings in this type
were activating mutations of TERT promoter in 96% of samples. The patients had favorable
clinical outcomes with a median survival of 8 years. The second type had IDH mutations
but without 1p/19q co-deletion or TERT promoter mutations. Rather, inactivation of
ATRX and mutation of TP53 (86% and 94%, respectively) was found in most of this type,
and these findings represent a strong association with astrocytoma. The patients had
a median survival of 6.3 years. The third type did not have an IDH mutation (wild-type
IDH) and had a spectrum of genetic alterations and disease outcomes more typical of
primary glioblastoma.
Diagnosis and management of diffuse low-grade glioma
The declaration of a dLGG by a seizure followed by the disclosure of the suspicious
lesion by MRI scanning is often a matter of grave concern to the patient, his/her
family, and his/her medical advisor. In the absence of any other physical manifestation
of intracranial disease and particularly when the lesion is in an eloquent or deep-seated
region of the hemisphere, the tendency is to prescribe anticonvulsants and plan to
follow-up with serial scans. Definitive therapy is deferred until tumor progression
is demonstrated on MRI scan or if clinical features change such as seizure intractability
or emergence of headache or neurological deficit. However, the accumulating evidence
and experience argue for a more active interventional management at onset and the
following are offered in support of such a course:
-
The growth rate of dLGG is variable between patients but does continue at a steady
rate of approximately 4 mm/year,[[32]] eventually accelerating in the 6 months leading up to the inevitable anaplastic
transformation
-
Visual inspection of serial MRI scans of a slow growing glioma for change in tumor
size may misleadingly underestimate volume expansion and lead to unnecessary delay
and loss of survival benefit from tumor resection and/or chemoradiation due to malignant
transformation
-
MRI is not diagnostic for a dLGG. Up to 30% of nonenhancing intrinsic brain tumors
seen on postcontrast MRI may eventually be found to be glioblastoma[[41]]
-
Patients with dLGG undergoing “upfront” early maximal safe resection have a survival
advantage over those undergoing a biopsy followed by watchful waiting.[[42]] This advantage is sustained even after adjustment for the status of IDH mutation
and 1p/19q co-deletion (Jakola 2017)
-
Even when located in putatively eloquent regions of the cerebrum, dLGG rarely present
with neurological deficits. The slow tumor expansion allows a plastic reorganization
of neural function which is displaced away from the tumor to adjacent brain.[[43]] The implication of this is that the tumor itself is nonfunctional tissue and even
seemingly inoperable lesions can be resected, if the access to and the functional
boundaries of, the tumor are defined by image guidance and electrostimulation brain
mapping
-
Tissue for diagnosis may come from a gross total resection (GTR) or debulking.[[44]],[[45]] Larger tissue sample is preferred in order to study the heterogeneity within the
tumor. In cases where a stereotactic biopsy with preoperative or intraoperative imaging
is planned, the areas with higher contrast enhancement are targeted to sample the
highest possible grade. The accuracy of such biopsies is reported at 51%–83% for the
highest possible grade.[[22]]
Surgical resection and biopsy
The role of initial surgical resection versus biopsy has remained controversial in
the management of patients with dLGGs without significant neurological deficits or
mass effect that may demand early intervention. The infiltrative nature of dLGGs in
the majority of cases and eloquent locations makes GTR challenging due to significant
risk of neurological morbidity. However, biopsy, even assisted with stereotactic targeting,
may not identify the highest possible grade in up to 50% of the cases because of tumor
heterogeneity.[[46]],[[47]] Moreover, biopsy is not risk free and indeed may have mortality/morbidity risk
equating those of modern series of surgical resections[[48]] Gross or near GTR is associated with better seizure control and higher PFS and
OS in addition to a lower risk of malignant transformation.[[10]] Chang et al. in 2008 retrospectively reviewed medical charts of 332 patients who
underwent initial surgical resection of dLGGs at a single center.[[13]] Two hundred and sixty-nine (81.02%) patients had seizure episodes preoperatively,
of which 132 (49%) had pharmacoresistant epilepsy. Postoperatively, 67% remained seizure
free, 17% had rare seizure episodes, 8% showed improvement in the seizure activity,
whereas only 9% were noted to have no improvement of the seizure episodes over a follow-up
period of 1 year.[[13]]
GTR was associated with better seizure control compared to subtotal resection (STR)
or biopsy alone (odds ratio: 16, P = 0.0064). Tumor progression was identified in
cases that showed seizure relapse following initial postoperative seizure control
(P = 0.001). Sanai and Berger in 2008 reviewed the literature from 1990 to 2008 and
found ten relevant studies describing surgical resection in dLGG patients.[[49]] The review confirmed a higher OS in patients with near-GTR. Duffau in 2008 reported
a personal consecutive series of 51 patients who underwent surgical resection for
insular WHO Grade II gliomas under cortico-subcortical stimulation.[[50]] Fifty patients underwent diagnostic studies for their presentation with seizure
episodes, among which 45% of the patients revealed normal neurologic examination findings
preoperatively. In the immediate postoperative period, 59% of the patients showed
worsening of their condition, whereas 3-month postoperative follow-up showed 96% of
the total patients returning to their baseline or an improved state of health. An
82% survival rate was reported at over a median follow-up period of 4 years.
Majchrzak et al. in 2012 prospectively analyzed the extent of resection (EOR) and
assessed final outcomes in 68 patients who underwent surgical resection of hemispheric
dLGGs.[[51]] The tumor volume and EOR were assessed by FLAIR MRI images to compare the pre-
and post-operative findings, whereas dynamic susceptibility contrast perfusion MRI
was used to assess rCBV.
They achieved GTR (>95%) in 31% patients (n = 21) (85%–95%), STR in 19% (n = 13),
and partial (<85%) resection (PTR) in 50% of the patients (n = 34). The EOR was statistically
lower for tumors in eloquent areas or close proximity. In Cox proportional hazard
analyses, the OS was predicted by age at presentation (hazard ratio [HR] 1.12, P =
0.032), EOR (HR 0.96, P = 0.025), and rCBV (HR 7.39, P = 0.002) at 5% level of significance.
The estimated 5-year OS was 100% for EOR >80%. Over a median follow-up period of 34
months, 6% mortality was observed. Similarly, the PFS was predicted by preoperative
tumor volume (HR 1.01, P = 0.005), postoperative tumor volume (HR 1.01, P = 0.008),
the EOR (HR 5.17, P = 0.001), percentage of resection (HR 0.98, P = 0.004), and by
the rCBV (HR 1.70, P = 0.033) at 5% level of significance. Permanent neurologic sequelae
were noticed in 6 (9%) patients which showed no statistical dependence on the EOR.[[51]]
Jakola et al. in 2012 compared the OS in 47 patients who underwent biopsy and watchful
waiting to 75 patients that underwent early surgical resection of dLGGs at two different
centers with parallel cohorts.[[42]] The cohort undergoing biopsy and observation had a median survival of 5.9 years
while median survival was not reached in the group, with the center favoring early
resection. In a follow-up analysis published recently, the authors extended the follow-up
to report that the OS was significantly worse in the center advocating watchful waiting
with a median survival of 5.8 years compared to 14.4 years for the center that preferred
early surgical resection (P < 0.01).[[52]]
A randomized controlled trial to test the efficacy of surgery for dLGG is not a practical
or ethical consideration with the weight of observational studies and reviews supporting
early and maximum safe resection “up front.” Seizure control is greatly improved by
tumor surgery,[[13]] and if the zone of T2-FLAIR hyperintensity is eliminated on the postresection MR
scan, survival is improved and malignant transformation delayed.[[53]] However, it must be recognized that resection cannot cure a dLGG, and minimizing
operative morbidity must be the surgeon's priority. Functional regions adjacent to
the tumor must be preserved by accurate identification.
This may be planned preoperatively by integrating functional MRI (FMRI) and DTI tractography
information into the surgical plan. Intraoperatively, neuronavigation systems incorporating
anatomical and tractographic images can guide the surgery. Adding ultrasonography
and, where possible, intraoperative MRI or CT updates the preoperatively acquired
images. Electrostimulation for brain mapping is the most useful method of delineating
the boundaries of resection in the cortex and subjacent subcortical white matter.
This requires the patient to be awake and cooperative, and this method is being used
increasingly to maximize safe resections. De Witt et al. carried out a meta-analysis
of 90 reports on 8091 patients.[[54]] They compared surgical resection of gliomas with and without intraoperative stimulation
mapping (ISM). They demonstrated a radiologically proven GTR rate of 75% (95% confidence
interval [CI], 66% to 82%) with the use of ISM compared to 58% (95% CI 48%–69%) without
using ISM.[[54]] Of the newer technologies, only ISM has been proven to improve EOR and safety of
the procedure. The rate of new neurological deficits was 3.4% compared to 8.2% with
and without the use of ISM, respectively.[[54]]
Neuronavigation has not shown promise in increasing the EOR or increasing its safety.[[55]]
Adjuvant chemotherapy and radiation
The more recent reported experience has supported the use of chemotherapeutics for
managing dLGGs as initial therapy and as adjunct to postresection residual disease
or progressive recurrent tumors. Hoang-Xuan et al. administered temozolomide (TMZ)
to sixty adult patients with biopsy-proven and radiologically progressive dLGGs.[[56]] The tumors responded radiologically in 31% of the cases, the disease remained stable
in 61% of the patients, whereas 8% of the dLGGs showed radiologic progression.
The tumors' maximum response was noticed at a median follow-up of 12 months, and chromosome
1p deletion was significantly associated with the tumor response (P < 0.004). Ricard
et al. in 2007 analyzed clinical information of 107 adult patients who underwent TMZ
chemotherapy for biopsy-proven dLGG with clinical or radiologic progressive disease.[[57]] The tumors' genetic profiles were obtained in majority of the cases for 1p, 19q
deletion, and p53 overexpression. At a median follow-up of 2 years, 65 (60.7%) patients
achieved a partial or minor response on imaging, 35 (32.7%) patients remained stable,
and 7 (6.5%) patients showed progression of the disease. Sixty-eight (63.5%) patients
showed clinical improvement, 34 (31.8%) patients remained stable, whereas 5 (4.7%)
showed deteriorating clinical condition. The tumors with 1p-19q codeletion showed
a significantly higher objective response to the therapy compared to the ones without
this codeletion (73 vs. 50%, P = 0.03). Seizure activity in the patients was better
controlled with chemotherapy even in the absence of radiological improvement.
Radiation therapy (RT) has been the adjuvant modality of convention in dLGG, but conflicting
information has emerged regarding the timing and the dose. In 2005, the EORTC 22845
randomized trial reported a post-dLGG resection comparison of OS and PFS in 157 patients
who underwent early RT and 157 patients in the control group who were watched radiologically
without any intervention.[[58]] The trial showed a median PFS of 5.3 years in the early RT-group and 3.4 years
in the control group (HR 0.59, P < 0.0001), whereas the OS showed no significant difference
for either group (HR 0.97, P = 0.872). One-year follow-up showed better seizure control
in the group undergoing early RT (P = 0.0329). Shaw et al. compared survival in 203
adult patients with dLGGs who underwent low- and high-dose RT.[[59]] One hundred and one of these patients underwent low-dose RT, while 102 underwent
high-dose RT, after GTR, STRs, or biopsy. At a median follow-up period of 6.43 years,
83 (41%) died, whereas low- or high-dose RT showed no significant difference in their
impact on survival. Tumor's histologic features, size, and patients age at presentation
were the significant prognostic factors recognized by multivariate analysis, and significantly
higher survival was associated with younger age at presentation, in addition to tumor
histology (oligodendroglioma) or oligo-rich histological components in the biopsy.
Seven (5.8%) patients showed Grade 3–5 radiation neurotoxicity. Douw et al. looked
at the cognitive function over a period of 12 years in dLGG survivors who underwent
RT.[[60]] RT had significant impact on the attention (P = 0.003), executive functions (P
= 0.03), and information processing speed (P = 0.05) when compared to the patients
who did not undergo RT.
The risk of delayed neurotoxicity following RT and the lack of survival advantage
of early treatment or dose escalation have led to a tendency to withhold radiation
as the first adjuvant option despite the recent advances in the planning and delivery
of conformal RT. In tumors without IDH mutation or codeletion or those clinically
at high risk of progression due to age, residual disease after surgery, and astrocytic
histology, early RT may be beneficial, particularly those patients progressing after
chemotherapy.[[61]]
In 2012, the RTOG 9802 showed that administering chemotherapy to dLGG patients, in
addition to RT, improved the PFS over a median follow-up of 5.9 years.[[62]] Adding chemotherapy to RT did not improve the OS in these patients, whereas on
post hoc analysis for 2-year survivors, the combination of chemotherapy with RT showed
better survival, which can be explained by delayed effectiveness. Based on the results
of this trial, the current trend is to prescribe combined postoperative chemoradiation
when the decision for early or late adjuvant therapy in made in a patient with dLGG.
Follow-up plan
Assessment of treatment response and disease progression poses a challenge in cases
of dLGGs, because of steady growth rate and limited contrast enhancement.[[63]] The Response Assessment in Neuro- Oncology (RANO) Working Group proposes criteria
for defining the assessment of dLGG post resection and adjuvant therapies using MRI
(T2/FLAIR sequence). According to RANO, the posttherapy status of the disease can
be categorized in five groups; (1) Complete response, (2) partial response, (3) minor
response, (4) stable disease, and (5) progression. The literature identifies no clinical
trials in order to recommend an optimal follow-up imaging frequency posttreatment.
The National Comprehensive Cancer Network[[64]] recommends imaging surveillance via MRI 3 or 6 monthly postoperative for the initial
5 years followed by yearly imaging later on. Rather than depending on the measurement
tools of MRI scanners for subjective assessment of tumor dimensions, it is recommended
that mean tumor diameter be calculated from semi-automated tumor contouring derived
from multiple FLAIR slices to allow improved detection of subtle changes.[[65]]
Follow-up after surgery of dLGG should be geared to assessing the progress of any
residual after sub-total resection or recurrence after GTR to preempt the progression
to malignancy by adjuvant therapy and/or repeat resection.
