From our study, the mean annual incidence of pediatric brain tumors receiving specialized care
in our center is 11.5 cases. Children accounted for 27.0% of all brain tumors treated in our center
during this period. Our hospital is a 1000- bed tertiary health facility; the largest multispecialty
hospital in the southeast region of our country endowed with a robust workforce in paediatric
neurosurgery. The annual incidence appears low; however, it may reflect constraints in accessing
care by patients generally and pediatric brain tumor patients specifically. Some of the constraints
have been studied extensively [23, 30]. At 9.5±2.1 years, the mean age of occurrence of pediatric
brain tumors is similar to that from a previous study performed by the lead author of this paper
and other researchers from another region of our country [30], as well as a recent retrospective
study from a center in our subregion with a smaller sample size [22]. In Morocco, a North African
country, a large epidemiological study found a mean age of 9.3years [11]. A slight male
preponderance found in our study is similar to some previous studies both from our country and
elsewhere [4, 22, 23, 30]. Headache was the most common presenting features among
supratentorial tumors while unsteadiness of gait occurred more frequently among infratentorial
tumors. 55.3% of tumors in our index series occurred in the supratentorial compartment.
Although the finding above is supported by another report from our sub region [4], it differed from
reports from other regions in our country [1, 23,30] as well as previous studies in other parts of
the world [4,11, 33]. This may suggest some geographical variability in the topographical
distribution of brain tumors which has been reported by some authors to be influenced by age [3,
15]. Craniopharyngioma (n = 21) is the most common tumor among histopathologically confirmed
cases in our series followed by medulloblastoma TABLE 2. However, when astrocytomas ( n = 12)
brain stem gliomas (n = 11) and oligodendrogliomas (n = 3) are included, gliomas (n = 26) will
represent the most common subgroup. Our previous study found medulloblastoma and
astrocytomas to be the most common tumors, while craniopharyngiomas were the third most
common tumor subtype [30]. An epidemiological study from Morocco, a country in the North of
Africa, reported medulloblastoma as the most common tumor followed by low grade astrocytoma,
while craniopharyngioma was the fifth most common tumor subtype [11]. We found a significant
age group bias in the occurrence of brain tumors among our cohorts (F = 5.7, DF1 = 2,
DF2 = 12,P = 0.029) TABLE 1. Low grade Astrocytoma and DNETS were seen in patients below 2
years of age. Some previous studies have also found a preponderance of astrocytomas in under –2
children [6, 8]. Ependymomas were predominant among 3 to 5-year age group, while
medulloblastomas were most common among age group 6 to 8 years respectively. Previous studies
have also reported age group variability in the occurrence of ependymomas and medulloblastomas
[5, 17, 20], as well as craniopharyngiomas which from our series is the most prevalent tumor
between ages 9 and 14 years. A previous study found a high incidence of craniopharyngiomas
among black children in the United States of America [29]. Our study seems to highlight this racial
and geographical predisposition to pediatric brain tumors or may reflect a changing trend more so
with the predominance of craniopharyngiomas in our series (22.6%) when compared to previous
studies from other centers in our country [22, 23, 30], and elsewhere, some of which reported a
higher frequency of occurrence of astrocytomas and medulloblastomas [3, 9, 10].Low grade
gliomas were predominant among adolescents between 15 and 16 years. We found a significant
association between hemoglobin genotype and the three most common histological variants of
brain tumors using Fishers exact test (P = 0.033). For instance, medulloblastomas occurred more
commonly among patients with Hb AA while craniopharyngiomas occurred mostly in Hb AS
patients. Moreover, both patients with Hb SS in our series had low grade astrocytomas. Our
findings are supported by a previous study which reported 3 cases of astrocytoma among 47
sicklers with cancer [27].Moreover, the distribution of Hb genotype among our study participants
( Hb AA = 83.9%, Hb AS = 13.8, Hb SS = 2.3% is similar to the result reported by Umoh et al from
another center within our sub region in which he found Hb AA = 78.7, Hb AS, 19.6, HbSS–1.5%)
31. Although our sample size is small, the difference in the hemoglobin genotype distribution
between our sample cohorts and the observed selective pediatric brain tumor occurrence
among individuals with specific hemoglobin genotypes are suggestive of possible hemoglobin
genotype profiling of pediatric brain tumor genotypes and phenotypes and will need to be
further evaluated through a large multicenter and perhaps global study. To our knowledge,
our study remains to date the largest prospective study reporting an association between
hemoglobin genotype and the predisposition to specific brain tumor phenotypes. We do not know
the extent to which our observations represent the salient relationship between hemoglobin
genotype and brain tumor profiles in children. A multicenter study is necessary to further verify
this association especially within sub-Saharan countries which alongside the Indian subcontinent
currently harbor over 90% of the global burden of sickle cell disease [24]. Such a study may
generate scientific evidence to suggest a review of the current risk factor associations of PBT’s.
We applied microneurosurgical resection to most cases (n = 69). Total microneurosurgical
resection (TMNR) alone was the treatment modality in 52.5% of surgically treated cases. Of all
microsurgically resected tumors, 43 cases were WHO Grade I, 3 were Grade II, 2 were Grade III
and 21 were Grade IV tumors. Surgery provides a direct and effective treatment path for brain
tumors and the extent of resection correlates with overall outcome [12, 19, 34]. From our series,
those who received TMNR had the best survival profile Table 3 and also the best school
performance Table 4. However, majority of brain tumors in paediatric patients grow near
functionally eloquent sites presenting a significant challenge to preserving them during surgery
[32]. Intraoperative neuronavigation and brain function mapping facilitate tumor resection and
help in preserving eloquent brain loci during surgery. 24 We currently do not have these parameters
in our unit and hence rely on microneurosurgical techniques. We believe the absence of
neuronavigation and intraoperative brain mapping may be a significant limiting factor to
performance of TMNR and may invariably play some role in lowering the survival profile of our
patients.
The role of radiotherapy in the treatment of pediatric brain tumors is well recognized [16, 25].
However, from our series, patients treated with radiotherapy had poorer overall survival (Table 3)
as well as school performance. Poor school outcomes with radiotherapy may partly reflect the
reality that more high-grade tumors with intrinsic propensity to poorer outcomes are treated with
this modality including brainstem gliomas. However, the long- term sequelae of radiation therapy
among paediatric brain tumor survivors especially its harmful effect on brain vasculature and
neurocognitive functions have been previously reported [2, 21]. Despite these sequelae,
radiotherapy together with cerebrospinal fluid diversion remain the mainstay for treatment of brain
stem gliomas which constitute 10.6% of pediatric brain tumors in our series[33]. Radiation
treatment for brainstem gliomas in our series was associated with 40% and 0%, 1 -Year and 5-
Year survival rates respectively which compares favorably with a report from a recent Asian
Study[33]. We also treated serum marker positive ICGCT’s with radiotherapy and the 1 year and
5-year survival rates for ICGCT from our series is 60% and 40% respectively which compares
with some previously reported outcomes for non germinomatous germ cell tumors [13, 33].
From our index study, marker positive ICGCT constitutes 6% of all paediatric brain
tumors. The use of generous radiation dose for tumor bed and ventricular space with or without
craniospinal irradiation and chemotherapy is considered a vital component of the standard
treatment protocol for germ cell tumors [18, 33]. Treatment complications and 30-day mortality
rates varied with tumor phenotype. Medulloblastoma was associated with the highest rate of post-
op complications and 30-day mortality TABLE 1, while Ependymomas, craniopharyngiomas and
low-grade gliomas were associated with lower rates.
Patients overall survival (OS) also varied with tumor phenotype as well as treatment modality
TABLE 1. Craniopharyngiomas, meningiomas, low grade gliomas, and ependymomas
were associated with the best survival, while brainstem gliomas and medulloblastomas had
the worst survival. Patients who received TMNR had the best survival of any treatment subgroup,
while surgically treated patients survived better than those treated with radiation (ꭓ2 = 8.9,
P = 0.017) TABLE 4, Fig 1 and 2. Although the overall 1-year and 5-year survival rates of 67%
and 52% respectively from our current study compare poorly to those from centers in the
developed world [14, 26], they however represent an improvement when compared to our previous
work as well as a recent study from our sub region [22, 30]. In our previous study, the 1-year and
5-Year survival rates were 53% and 47% respectively [30]. We believe this improvement is the
result of multiple factors, including the establishment of a multidisciplinary oncology center which
has an active radiotherapy unit equipped with a linear accelerator as well as increase in the number
of neurosurgeons with experience in paediatric neurosurgery. Furthermore, an increase in
awareness and improved referral practices may partly account for the overall positive survival shift
in this study, which more so is significantly more experienced among those treated with TMNR.
From our study, PBT’s negatively affect school attendance and performance, table 5
and the treatment modalities are also associated with variable influence on school performance.
Surgical resection especially total microneurosurgical tumor resection (TMNR) is associated with
a better school enrolment profile and better school performance compared to SMNR and
radiotherapy, X2 = 4.94, P = 0.015. Although, lack of existing local data on the school enrolment,
drop-out and completion profiles in the general pediatric population imposes a salient challenge
on the interpretation of our findings. However, our results are quite significant because PBT
survivors are often excluded from studies evaluating childhood cancer survivors, due in part to
concerns on the impact of cognitive limitations on the validity of assessment [26. 35]. Therefore
having data on the impact of pediatric brain tumor and its treatment of school enrolment and
performance profile in our setting will provide a useful resource for comparisons when the general
data becomes available and may actually drive efforts to obtain such resource. To our knowledge,
this is the first study evaluating school performance among survivors of PBT in our region and
therefore provides the reference for further studies on this subject. More clinical studies and
collaborations are needed between centers treating PBT’s within the sub-Saharan region so as to
further elucidate the variations in presentation, treatment practices, as well as evaluate relevant
potentially significant associations including hemoglobin genotype. This will help us to summate
the outcome of pediatric brain tumors within the region.