Keywords:
pediatric high-grade gliomas - overall survival - extent of resection - temozolomide
Introduction
Brain tumors are the most common solid cancers in pediatric patients.[1] High-grade glioma (HGG) compromises approximately 15 to 20% of all central nervous
system (CNS) tumors in children.[2]
[3] The incidence of pediatric HGG is about 0.85 per 100,000.[4]
Although glioblastoma multiforme (WHO grade IV) is the most common primary brain tumor
in adults, anaplastic astrocytoma (WHO grade III) is more common than glioblastoma
in children.[5] Recent studies have established that pediatric HGGs (PHCGs) are biologically distinct
from their adult counterparts, with nearly half of these patients harboring somatic
mutations in histone genes.[6]
[7]
In spite of its relative rarity, HGG is a devastating disease in children with significant
morbidity and mortality.[1]
[8]
Cerebral hemispheres are the most commonplace of PHGG, however, it can arise from
other places including the brain stem (30%), the thalamus (13%), the spinal cord (3%),
and the cerebellum (5%).[9]
[10]
The standard care of adult glioblastoma multiforme (GBM) consists of maximal surgical
resection followed by concurrent and adjuvant chemoradiation.[11]
Although there is no such standard protocol for PHGG, a similar protocol is used by
most physicians across the world.[2] Despite such a multimodal approach to treatment, patient survival is already poor.[12]
In the present study, we aimed to present our 10-year experience in managing pediatric
patients with a high-grade glioma, focusing on the clinical presentation, prognostic
factors, clinical outcomes, and patients’ survival.
Materials and Methods
In this retrospective cohort study, we investigated pediatric patients with high-grade
glioma.
All patients with 18 years of age or younger who were diagnosed to have a high-grade
glioma and managed at our center between May 2009 and May 2018 were included ([Fig. 1]). Of a total of 323 patients with a high-grade glioma managed at our center, 53
cases were found to be in the pediatric age group. We excluded 12 patients with diffuse
intrinsic pontine glioma from the study. The medical records of all patients including
demographic features, clinical presentation, surgical and adjuvant therapies, and
follow-up data were reviewed. The present study was approved by the research committee
of the Kermanshah University of Medical Sciences and the medical ethics committee.
Informed written consent was obtained from all patients’ relatives before any intervention
and treatment and for publication of deidentified data.
Fig. 1 Flowchart of study population.
All patients underwent surgical resection followed by radiation therapy.
Postoperative radiation therapy was started within 4 to 6 weeks of surgery (54 grays,
at 1.8 grays per fraction daily over 6 weeks).
Twenty-five patients received concurrent temozolomide (TMZ 75 mg/m2) followed by 6 cycles of maintenance treatment (150–200 mg/m2 [day 1–5] every 4 weeks).
Based on the early postoperative magnetic resonance imaging (MRI), the extent of resection
was classified into three groups: the total resection group (resection more than 90%),
subtotal group (resection between 10 and 90%), and the biopsy group (resection < 10%).
The surgical sample of each patient was investigated by an experienced pathologist
meticulously.
The immunostaining was performed for the detection of Ki67 and p53. The labeling index
for p53 was defined as the percentage of immunostained cells per 200 cells in 5 fields.
The presence of Ki67 was determined using the percentage of positive cells per 1,000
cells.
Statistical Analysis
Overall survival defined as the time from diagnosis till death due to all causes or
the latest date of follow up. Kaplan–Meier method was used to estimate overall survival
(OS). The log-rank test was employed in order to test the survival patterns. The Cox
proportional hazards models were used to identify the predictors of mortality. Crude
and adjusted hazard ratios with 95% confidence intervals were calculated for each
variable. The level of statistical significance was set at 0.05. The preliminary assumptions
including proportional hazards assumption has been checked before the analysis; it
was not violated. All analyses were performed with Stata software version 14.2 (StataCorp,
College Station, Texas, United States).
Results
Non–Brain Stem High-Grade Glioma
A total of 41 children with the mean age of 9.12 ± 3.42 years were evaluated. There
were 21 boys (51.2%) and 20 girls (48.8%). The mean OS was 21.24 ± 10.16 months. Anaplastic
astrocytoma (grade III) and glioblastoma multiforme (grade IV) were diagnosed in 24
(58.5%) and 17 (41.5%) patients, respectively ([Table 1]).
Table 1
Basic characteristics of the children with non–brain stem high-grade glioma
|
Variable
|
Frequency
|
Percentage
|
|
Abbreviation: KPS, karnofsky performance scale.
|
|
Gender
|
|
Male
|
21
|
51.2
|
|
Female
|
20
|
48.8
|
|
KPS
|
|
|
|
>70
|
35
|
85.4
|
|
<70
|
6
|
14.6
|
|
Extent of surgical resection
|
|
|
|
Gross total resection
|
19
|
46.4
|
|
Partial resection
|
16
|
39.0
|
|
Biopsy
|
6
|
14.6
|
|
Temozolomide therapy
|
|
|
|
Yes
|
25
|
61.0
|
|
No
|
16
|
39.0
|
|
Tumor location
|
|
|
|
Frontal
|
10
|
24.4
|
|
Temporal
|
8
|
19.5
|
|
Parietal
|
6
|
14.6
|
|
Occipital
|
4
|
9.8
|
|
Insular
|
4
|
9.8
|
|
Cerebellar
|
3
|
7.3
|
|
Thalamus
|
6
|
14.6
|
|
Tumor size
|
|
|
|
<3 cm
|
9
|
22.0
|
|
3-5 cm
|
23
|
56.1
|
|
>5 cm
|
9
|
22.0
|
|
Grade of tumor
|
|
|
|
Grade III (anaplastic astrocytoma)
|
24
|
58.5
|
|
Grade IV (glioblastoma multiforme)
|
17
|
41.5
|
|
Outcome
|
|
|
|
Died
|
33
|
80.5
|
|
Survived
|
8
|
19.5
|
|
Ki67
|
|
|
|
<25%
|
18
|
43.9
|
|
>25%
|
16
|
39/0
|
|
Unavailable
|
7
|
17.1
|
|
P53
|
|
|
|
>50%
|
16
|
39.0
|
|
<50%
|
19
|
46.3
|
|
Unavailable
|
6
|
14.6
|
|
Age (mean ± standard deviation), year
|
9.12 ± 3.42
|
|
|
Overall survival (mean ± standard deviation), month
|
21.24 ± 10.16
|
|
Different presentations of the patients have been shown in [Table 2].
Table 2
Different presentations of the children with non–brain stem high-grade glioma
|
Number of patients
|
Presentations
|
|
24
|
Increased intracranial pressure
|
|
18
|
Seizures
|
|
9
|
Motor weakness
|
|
3
|
Visual failure
|
|
2
|
Ataxia
|
The most common site of the tumor was frontal (24.4%), followed by temporal (19.5%),
parietal (14.6%), and thalamus (14.6%) ([Table 1]).
Gross total resection was achieved in 19 patients (46.4%), 16 patients (39.0) underwent
partial resection and biopsy was performed in 6 patients (14.6%). Age, gender, tumor
location, and size of the tumor were not associated with survival ([Table 3]). Based on the univariate analyses extent of surgery (p = 0.001), karnofsky performance scale ( KPS) (p < 0.0001), TMZ therapy (p < 0.0001), grade of tumor (p < 0.0001), Ki67 (p = 0.001), and p53 (p = 0.017) were the predictors of the OS ([Table 3]).The extent of resection (p = 0.002, hazard ratio [HR] = 4.84), the grade of the tumor (p = 0.017, HR = 4.36), and TMZ therapy (p = 0.038, HR = 3.57) were the independent predictors of the OS according to the adjusted
Cox model ([Table 3]
[Fig. 2]
[3]
[4]). According to multivariate analyses, the probability of death in children who underwent
partial resection was more than those who underwent gross total tumor resection by
5 times (p < 0.005). The odds of death were higher in children who did not receive TMZ compared
to those receiving it by 3.57 times (p < 0.05). Patients with GBM had a higher chance of death in comparison with those
with anaplastic astrocytoma by 4.36 times (p < 0.0001).
Table 3
Predictive factors of overall survival in children with non–brain stem high-grade
glioma
|
Variable
|
Crude HR (95% confidence interval)
|
p-Value
|
Adjusted HR (95% confidence interval)
|
|
|
Abbreviations: HR, hazard ratio; KPS, Karnofsky Performance Scale.
|
|
Gender
|
Male
|
1
|
|
|
|
|
Female
|
0.74 (0.36, 1.52)
|
0.410
|
|
|
|
KPS
|
>70
|
1
|
|
1
|
|
|
<70
|
7.19 (2.43, 21.30)
|
<0.0001
|
2.18 (0.65, 7.31)
|
0.205
|
|
Extent of surgical resection
|
Gross total resection
|
1
|
|
1
|
|
|
Partial resection
|
3.92 (1.74, 8.83)
|
0.001
|
4.84 (1.76, 13.26)
|
0.002
|
|
Biopsy
|
8.88 (2.58, 30.54)
|
0.001
|
2.07 (0.49, 8.72)
|
0.322
|
|
Temozolomide
|
Yes
|
1
|
|
1
|
|
|
No
|
9.91 (3.58, 27.37)
|
<0.0001
|
3.57 (1.07, 11.84)
|
0.038
|
|
Tumor location
|
Frontal
|
1
|
|
|
|
|
Temporal
|
0.61 (0.18, 1.99)
|
0.413
|
|
|
|
Parietal
|
0.58 (0.19, 1.18)
|
0.354
|
|
|
|
Occipital
|
0.83 (0.25, 2.75)
|
0.765
|
|
|
|
Insular
|
0.85 (0.25, 2.93)
|
0.803
|
|
|
|
Cerebellar
|
1.27 (0.33, 4.78)
|
0.719
|
|
|
|
Thalamus
|
0.88 (0.26, 2.91)
|
0.845
|
|
|
|
Tumor size
|
<3 cm
|
1
|
|
|
|
|
3–5 cm
|
1.77 (0.66, 4.77)
|
0.257
|
|
|
|
>5 cm
|
2.37 (0.78, 0.72)
|
0.129
|
|
|
|
Grade of tumor
|
Grade III (anaplastic astrocytoma)
|
|
|
1
|
|
|
Grade IV (glioblastoma multiforme)
|
6.37 (2.79, 14.56)
|
<0.0001
|
4.36 (1.30, 14.62)
|
0.017
|
|
Ki67
|
<25%
|
1
|
|
1
|
|
|
>25%
|
7.42 (2.84, 19.41)
|
<0.0001
|
2.63 (0.74, 9.33)
|
0.135
|
|
Unavailable
|
1.29 (0.49, 3.37)
|
0.597
|
0.78 (0.07, 8.46)
|
0.843
|
|
P53
|
>50%
|
1
|
|
1
|
|
|
<50%
|
2.62 (1.19, 5.78)
|
0.017
|
0.98 (0.31, 3.09)
|
0.979
|
|
Unavailable
|
1.20 (0.43, 3.36)
|
0.723
|
2.33 (0.18, 30.50)
|
0.519
|
|
Age
|
1.03 (0.91, 1.15)
|
0.666
|
|
|
Fig. 2 Comparison of Kaplan–Meier estimates of survival between the patients with anaplastic
astrocytoma and Glioblastoma multiforme (p-value = <0.0001).
Fig. 3 Comparison of Kaplan–Meier estimates of survival between the patients with different
amount of tumor resection (p-value = <0.0001).
Fig. 4 Comparison of Kaplan–Meier estimates of survival between the patients received temozolomide
(TMZ) with children that did not receive TMZ (p-value = <0.0001).
Discussion
Our results showed that the extent of resection, the grade of the tumor, and temozolomide
therapy were the independent predictors of OS in children with NBSHGG.
PHGGs consist of a heterogeneous class of CNS neoplasms that affect pediatric patients
of different ages.[8]
[10] They can originate from different sites of CNS and have different histologic aspects.[11] While gliomas compromise 40 to 50% of all CNS tumors in children, supratentorial
HGGs represent only 6 to 12% of all primary brain neoplasms in children.[13]
[14] Depending on tumor grade, PHGGs are divided into either anaplastic astrocytoma (WHO
grade III) or glioblastoma multiforme (WHO grade IV). PHGGs are characterized by some
histopathological features, including nuclear atypia, hypercellularity, and vascular
proliferation.[2]
[15] Symptoms of PHGG are various and may be nonspecific. The most common presentations
are increased intracranial hypertension, seizure, long tract signs, and motor weakness.
Some studies have declared a higher incidence of seizure in PHGG compared with adult
HGG.[5]
[16] The incidence of seizure has been reported in about 30% of affected children.[17]
[18] The reported survival for PHGG ranges from 8 to 70 months. Several factors including
the extent of resection, tumor grade, age, and KPS have been suggested as the predictors
of outcome in PHGG.[1]
[9]
[10] Several studies have established the positive impact of maximal safe tumor resection
on the survival of children with HGG.[9]
[19] Nikitović et al in their study evaluated 15 children with glioblastoma. The median
OS of their cases was 13.5 months. They found that there was no relationship between
age, gender, type of radiotherapy, or tumor location with patient survival. However,
pediatric patients who underwent gross total resection have a longer survival.[6]
In another study, Das et al investigated 65 cases with glioblastoma. The median progression-free
survival and OS of their cases were 10 and 20 months, respectively. They reported
that the extent of surgical tumor resection was the strongest predictor of patients’
survival.[20]
Perkins et al, in a retrospective study, evaluated 24 pediatric patients with glioblastoma.
In their study, median OS was 13.5 months as well as the 2-year OS rate was 32%. There
was no relationship between OS and patients’ age, gender, tumor location, radiation
volume, radiation dose, or the use of chemotherapy. However, they found that patients
with gross total resection had a longer survival.[12] Using advanced techniques such as intraoperative navigation, intraoperative magnetic
resonance imaging, and intraoperative cortical mapping as well as using easily available
cheaper useful alternatives like ultrasonography may help to a larger extent of tumor
resection.[1]
[7]
Rapid analysis of the surgical sample for molecular hallmarks such as isocitrate dehydrogenase
(IDH) mutation can facilitate a molecular diagnosis within no time. This could provide
better intraoperative decision-making about extent of resection.[21]
[22]
Radiotherapy at the dose of 50 to 60 grays fractionated over 6 weeks is a significant
compartment of treatment in PHGG.[5] Radiotherapy usually is not used in children younger than 3 years due to its adverse
effect on the developing brain.[5]
[16]
The impact of concomitant and adjuvant TMZ therapy on improving survival in adults
with HGG has been demonstrated.[4]
[23] TMZ is a cytotoxic drug acts by alkylation at the O6 position of guanine.[24] TMZ can improve survival by diminishing methylguanine DNA methyltransferase (MGMT)
promoter expression in tumor cells.[25] Although our findings demonstrated that children who received TMZ had longer survival,
most studies demonstrated that TMZ therapy did not improve survival in PHGG.[24]
[25]
Limitations
There are several limitations to this study. Our study was a retrospective analysis
of a single-center experience. The small sample size and the lack of a complete molecular
profile for all the cases were other important limitations of our work. In spite of
the mentioned limitations, our study could be a valuable addition to our knowledge
about PHGG. Finally, we suggest multicenter prospective studies to investigate prognostic
factors in PHGG.
Conclusion
HGGs are uncommon pediatric tumors with an aggressive nature and a poor prognosis.
Our results showed that in NBSHGG cases, those with maximal safe tumor resection and
those that received temozolomide therapy as well as children with anaplastic astrocytoma
had higher survival.
