Abstract
Treatment of childhood brain cancer has been associated with long-term cognitive morbidity in children. In the present study, the cognitive status of children with brain tumors was examined prior to any treatment to single out the role of tumor and tumor-related factors in cognitive deficits. Eighty-three children with newly diagnosed brain tumors (mean age, 8.6 years; range, 7 months to 16.6 years; median, 9.4 years) completed an extensive battery of age-related tests to assess cognitive function before any therapeutic intervention. Magnetic resonance imaging (MRI) was used to determine tumor site and volume and tumor-related factors. Performance under test was compared with symptom duration, neurological status, epilepsy, and MRI. Cognitive difficulties are detected at diagnosis in as many as 50% of patients for some cognitive domains; 6% of patients present with true-diagnosed mental retardation. The location of the tumor is the principal determinant of cognitive deficits, with major impairment in children with cortical tumors. Symptom duration and the presence of epilepsy are significantly associated with neuropsychological disabilities, while neuroradiological tumor-related variables do not correlate clearly with neurocognitive performance. The knowledge of the pre-existing cognitive deficits is critical to evaluate the results of treatment, providing a baseline for assessing the true impact of therapy in determining cognitive decline. In addition, the study suggests that some clinical variables require careful monitoring, because they could be specifically implicated in the neuropsychological outcome; the efforts to reduce the impact of these factors could ameliorate long-term prognosis.
Keywords: brain tumor, children, neurocognitive deficit
The gradual improvement in the effective therapy of childhood central nervous system (CNS) tumors has resulted in a substantial increase in cure rate and survival. However neuropsychological deficits and sometimes severe disabilities have been frequently observed among long-term survivors.1–4 Neuropsychological effects may differ as a function of tumor localization,5 surgical procedure,6–8 and medical therapies.1,2,9,10 Radiotherapy (RT) is known to be the major determinant of cognitive sequelae,11–13 which mainly affect younger children and are progressive.9 The importance of tumor-related neuropsychological damage is not fully understood, because early studies were conducted mainly after treatment had been completed. In addition, neuropsychological testing was, in some cases, limited to intelligence quotient (IQ) measurements, thus obscuring the possible presence of specific deficits of cognitive functions. In a previous study of children with acute lymphoblastic leukaemia (ALL) treated with cranial irradiation and chemotherapy, we found cognitive difficulties that could be detected only by appropriate tests.14
Because most studies lack any evaluation of mental functioning before treatment, it is difficult to exclude all of the possible tumor-related negative effects, such as tumor site, tumor size, and intracranial hypertension. Similarly, these studies cannot be used to address the potentially beneficial effect of tumor removal and the role of surgery techniques on cognitive organization.
The influence of tumors at diagnosis has more recently been evaluated in adults,5,15 whereas only few reports have studied the neuropsychological functioning in children before treatment.16,17 Obviously, the results obtained in adults cannot easily be applied to children, because tumor characteristics, incidence, and, subsequently, therapeutic procedure are quite different. Moreover, CNS maturation is an ongoing process during childhood, and focal lesions could impact the cognitive organization in quite different ways.
Early detection of cognitive deficit could allow well-timed rehabilitative intervention, with greater chances of improving the quality of life or at least of avoiding or minimizing any cognitive deficit acquired in the course of cancer treatment.
In the present article, we report the results of an extensive evaluation of cognitive functions at the time of diagnosis in children with CNS primary brain tumors.
Methods
Patients
During the period 1998–2004, 102 children were referred to our hospital for CNS tumors. The inclusion criterion for the study was the presence of unifocal primary intracranial mass lesions. Patients with multiple tumor sites (2 patients) and severe neurological conditions (5 patients) were excluded. Ninety-five children were found to be eligible for the study. Six patients were excluded for the presence of pre-existing mental deficits (4 had learning disorders, and 2 had psychiatric diagnoses, according to the criteria of Diagnostic and Statistic Manual of Mental Disorders-IV-Text Revision). For 6 patients, we did not obtain informed consent. Neuropsychological tests were thus administrated to 83 of 95 children.
Early signs and symptoms of the tumor, duration of clinical manifestations before diagnosis, and neurological examination findings were investigated.
Because epilepsy is considered in itself to be a risk factor for cognitive functioning,18,19 presence and severity of epilepsy were considered as separate categories. For classification of clinical variables see Table 1.
Table 1.
Distribution of patients by tumor site
| Total no. of patients | |
|---|---|
| Supratentorial hemispheric | 38 |
| Frontal | 7 |
| Parietal | 2 |
| Temporal | 25 |
| Temporo-parietal | 2 |
| Parieto-occipital | 2 |
| Right hemisphere | 18 |
| Left hemisphere | 20 |
| Lateral | 13 |
| Mesial | 25 |
| Supratentorial midline | 21 |
| Hypothalamic | 7 |
| Thalamic | 3 |
| Third ventricle | 3 |
| Pineal | 4 |
| Pituitary gland | 3 |
| Corpus callosum | 1 |
| Posterior cranial fossa | 24 |
| Cerebellum | 19 |
| Vermal (9), Right hemisphere(4), Left hemisphere (4), Vermal and hemisphere (2) | |
| Ponto-cerebellar angle | 1 |
| Fourth ventricle | 3 |
| Brainstem | 1 |
Tumor characteristics and tumor-related factors
The diagnosis of tumor presence and tumor site was based on brain MRI. The whole population was divided into 3 principal categories according to tumor site: supratentorial hemispheric (SH), supratentorial midline (SM), and posterior cranial fossa (PCF).
In addition to the 3 main tumor sites, further analysis of tumor location was performed in SH and PCF according to the following criteria: in SH tumors, the differences between right- versus left-side and lateral versus mesial tumors were examined; in PCF tumors, cerebellar tumor, vermis versus hemispheric, and right versus left location were analyzed (Table 1).
In 62 (74.6%) of 83 patients, MRI scans were acquired in our department and reviewed for a complete analysis of tumor characteristics and tumor-related factors by one of us, who was blinded to clinical history. Tumor volume, perilesional edema, compartment/localized intracranial hypertension, and hydrocephalus were analyzed.
Tumor volume was evaluated according to the criteria of Gobel et al.20 The distinction between the tumor itself and the adjacent edema was based on comparison between contrast-enhanced T1-weighted and unenhanced T2-weighted images in tumors with contrast enhancement. In nonenhancing tumors, the distinction has been based on the different T2-weighted images (fluid attenuation inversion recovery [FLAIR], turbo spin echo [TSE], and gradient echo [GE]) and on diffusion-weighted images.
The median values for the 3 sites were calculated and considered for statistical analysis.
Neuropsychological assessments
Table 2 lists the neuropsychologic tests performed.
Table 2.
Neuropsycological domains and test of measurement
| Intelligence |
| GMDSa |
| Wechsler intelligence scales |
| Language |
| The hearing and speech scale of GMDSa |
| Verbal scale of Wechsler intelligence scales |
| Visual motor integration |
| The hand and eye coordination of GMDSa |
| Beery visual-motor integration test |
| Rey figure: copying |
| Verbal memory |
| Short and long term verbal memory |
| Visual memory |
| Rey figure: memory |
| Working memory |
| Phonological and spatial working memory |
| Attention |
| Visual-spatial analysis and attention: speedy and accuracy (bells test-r) |
| Freedom from distractibility by Wechsler scales |
| Executive functions |
| Word fluency: phonemic and semantic |
| Planning task (tower of London) |
GMDS indicates Griffiths mental developmental scales.
aTests performed in children aged less than 48 months are reported.
Statistical analysis
The results of neuropsychological tests, except for intelligence scales and for language and memory measures, were normalized and converted into z scores (mean value, 0; standard deviation, 1).
In the entire population, we analysed cognitive performance as a function of the following:
The patient's characteristics (sex and age).
Clinical variables (symptom duration, grading of neurological impairment, and presence and severity of epilepsy).
Tumor site. This variable was examined taking into consideration the 3 principal locations (SH, SM, and PCF) and the specific location in the SH and PCF.
The complete description of the aforementioned variables is showed in Table 3.
Table 3.
Clinical characteristics of study patients, by tumor location
| All patients | Tumor location |
|||
|---|---|---|---|---|
| (n = 83) | SH (n = 38) | SM (n = 21) | PCF (n = 24) | |
| Sex | ||||
| Female | 30 (36.1) | 14 (36.8) | 7 (33.3) | 9 (37.5) |
| Male | 53 (63.9) | 24 (63.2) | 14 (66.7) | 15 (62.5) |
| Histology | ||||
| Astrocytoma (grade I) | 24 | 13 | 2 | 9 |
| Astrocytoma (grade II) | 9 | 8 | 1 | 0 |
| Astrocytoma (grade III) | 4 | 3 | 0 | 1 |
| Glioblastoma | 4 | 2 | 2 | 0 |
| Medulloblastoma | 12 | 0 | 0 | 12 |
| Ependymoma | 1 | 0 | 0 | 1 |
| Germinoma | 6 | 0 | 6 | 0 |
| Craniopharyngioma | 6 | 0 | 6 | 0 |
| Ganglioglioma | 5 | 5 | 0 | 0 |
| Other histotypesa | 12 | 7 | 4 | 1 |
| Age at diagnosis, months | ||||
| Mean ± SD | 103.10 ± 48.6 | 99.47 ± 53.8 | 121.95 ± 42.2 | 92.33 ± 41.6 |
| Range | 7–200 | 7–200 | 39–191 | 18–178 |
| Age <48 months, proportion (%) | 11/83 (13.2) | 9/38 (23.7) | 1/21 (4.8) | 1/24 (4.2) |
| Symptom duration, mean days ± SD | 465.40 ± 769.5 | 739.16 ± 937.2 | 307.95 ± 484.8 | 169.71 ± 506 |
| Symptom durationb | ||||
| 1 | 39 (47) | 12 (31.6) | 9 (42.9) | 18 (75) |
| 2 | 44 (53) | 26 (68.4) | 12 (57.1) | 6 (25) |
| Neurological examinationc | ||||
| 0 | 47 (56.6) | 31 (81.6) | 10 (47.6) | 6 (25) |
| 1 | 27 (32.5) | 5 (13.1) | 10 (47.6) | 12 (50) |
| 2 | 9 (10.8) | 2 (5.3) | 1 (4.8) | 6 (25) |
| Epilepsyd | ||||
| 0 | 46 (55.4) | 5 (13.2) | 17 (80.9) | 24 (100) |
| 1 | 16 (19.3) | 13 (34.2) | 3 (14.3) | 0 |
| 2 | 13 (15.7) | 13 (34.2) | 0 | 0 |
| 3 | 8 (9.6) | 7 (18.4) | 1 (4.8) | 0 |
Data are no. (%) of patients, unless otherwise indicated. PCF indicates posterior cranial fossa, SD indicates standard deviation; SH indicates supratentorial hemispheric, SM indicates supratentorial midline.aMeningioma, dysembrioplastic neuroepithelial tumor, teratoid-rhabdoid tumor, primitive neuroepithelial tumor, hamartoma, and pinealocytoma.
bSymptom duration: the time elapsing between initial signs and tumor detection was investigated. The whole population was divided into two groups on the basis of symptom duration, using the median value of 100 days to discriminate between early and late diagnosis. 1 = ≤100 days; 2 = >100 days.
cNeurological examination: classified according to the severity as follows: 0, absent; 1, mild, not interfering with daily life; and 2, major neurological signs, disabling daily life.
dEpilepsy: classified as: 0, absent; 1, mild (less than monthly seizures); 2, moderate (weekly seizures); and 3, severe (daily seizures, generally refractory to pharmacological treatment).
In 62 of the 83 patients whose scans were reviewed we also analyzed cognitive scores as a function of the following:
Tumor characteristics (volume and histotype).
Tumor-related factors (perilesional edema, compartment/localized hypertension, and hydrocephalus).
The complete description of the radiological mentioned variables is shown in Table 4.
Table 4.
Radiological characteristics in 62 of 83 patients and in the 3 groups, by tumor location
| All patients (n = 62) | Patients with SH (n = 30) | Patients with SM (n = 17) | Patients with PCF (n = 15) | |
|---|---|---|---|---|
| Tumor volume, mm3 | ||||
| Mean ± SD | 19.15 ± 26.68 | 14.41 ± 23.95 | 21.58 ± 36.61 | 25.90 ± 16.68 |
| Median value | 9.2 | 5.5 | 8.8 | 25.1 |
| Perilesional edemaa | ||||
| 0 | 42 (67.7) | 21 (70) | 13 (76.5) | 8 (53.3) |
| 1 | 19 (30.7) | 9 (30) | 4 (23.5) | 6 (40) |
| 2 | 1 (1.6) | 0 | 0 | 1 (6.7) |
| Compartment/localized hypertensionb | ||||
| 0 | 51 (82.3) | 27 (90) | 15 (88.2) | 9 (60) |
| 1 | 2 (3.2) | 1 (3.3) | 0 | 1 (6.7) |
| 2 | 9 (14.5) | 2 (6.7) | 2 (11.8) | 5 (33.3) |
| Hydrocephalusc | ||||
| 0 | 45 (72.6) | 28 (93.3) | 9 (52.9) | 8 (53.3) |
| 1 | 17 (27.4) | 2 (6.7) | 8 (47.1) | 7 (46.7) |
aPerilesional edema: defined using a 3-point scale: 0, absent; 1, minimal; and 2, marked.
bCompartment/localized hypertension was graded on the basis of effacement of cerebrospinal fluid (CSF) spaces in the intracranial compartment harboring the mass lesion: 0, absent; 1, present; and 2, resulting in brain herniation.
cHydrocephalus: classified as follows: 0, absent; 1, minimal; and 2, marked (when supratentorial subarachnoid convexity spaces were completely effaced).
Group performances were compared using multivariate analysis of variance (MANOVA), and the Tuckey test was used for post-hoc multiple comparison procedures. Data were analyzed using SPSS statistical software, version 16 (SPSS).
Results
The ratio of male to female patients was 1.76:1. The mean patient age (± standard deviation [SD]) was 103.1 ± 48.6 months (median age, 112 months); 11 children (13.2%) were aged <4 years. In 38 patients (45.8%), tumors were located in SH; 21 children (25.3%) had SM masses, and 24 (28.9%) had PCF tumors. Data on clinical characteristics, by tumor site, in the full population and in the 3 different location groups are shown in Table 3.
Symptom duration varied from 5 days to 10 years (mean ± SD, 465.40 ± 769.5 days; median duration, 100 days). In children with cortical tumors (involving the cerebral cortex and the subcortical white matter), symptom duration was significantly higher than in other groups, with 68.4% of patients with clinical onset exceeding the median value of 100 days before diagnosis.
Table 3 shows data about tumor grade and tumor histotype in the full population and in the 3 groups, by site. Data on the radiological review performed in 62 patients are summarized in Table 4.
Median values of volume varied widely in the 3 different sites. PCF tumors were usually larger than SH and SM tumors (median values, 25.1 mL, 5.5 mL, and 8.8 mL, respectively). Table 5 shows the neuropsychological results in all patients and in the 3 groups, by location.
Table 5.
Summary of neuropsychological scores across tumor location groups and in the entire sample
| Neuropsychological domain | Full population | SH | SM | PCF | Multivariate ANOVA P value | Post-hoc comparison |
|---|---|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Mean (SD) | Mean (SD) | |||
| TIQ (score) | 99.3 ± 16.1 | 95.26 ± 19.1 | 105.52 ± 11.7 | 100.25 ± 12.2 | .058 | SH vs SM |
| VIQ (score) | 104.08 ± 16.2 | 98.85 ± 19.2 | 112.11 ± 10.5 | 104.96 ± 12.7 | .015 | SH vs SM |
| PIQ (score) | 96.32 ± 16.3 | 94.03 ± 19.3 | 100.36 ± 14.1 | 96.26 ± 12.8 | .410 | |
| Language | 104.08 ± 16.2 | 98.85 ± 19.2 | 112.11 ± 10.5 | 104.96 ± 12.7 | .015 | SH vs SM |
| Visual-Motor Integration Test (z-score) | −0.57 ± 1.2 | −1.22 ± 1 | −0.13 ± 1.3 | −0.19 ± 1.1 | .003 | SH vs SM/PCF |
| Coping Rey's Complex Figure (z-score) | −0.56 ± 2.2 | −0.94 ± 2.8 | −0.44 ± 1.6 | 0.01 ± 1.4 | .408 | |
| Verbal Short Term Memory (score) | 34.35 ± 11.1 | 32.65 ± 11.8 | 33.76 ± 10.8 | 37.66 ± 10.1 | .339 | |
| Verbal Long Term Memory (score) | 7.56 ± 3.6 | 7.19 ± 3.5 | 7.18 ± 3.7 | 8.54 ± 3.5 | .420 | |
| Visual Memory (z-score) | −0.8 ± 1.9 | −0.79 ± 1.8 | −1.22 ± 2.2 | −0.26 ± 1.8 | .399 | |
| Phonological Working Memory (score) | 4.7 ± 1 | 4.38 ± 1.1 | 5.35 ± 0.9 | 4.52 ± 0.8 | .030 | SH vs SM |
| Spatial Working Memory (score) | 4.2 ± 0.9 | 4.26 ± 0.8 | 4.46 ± 0.9 | 3.85 ± 1 | .138 | |
| Visual Attention Speedy (z-score) | −0.73 ± 1.1 | −0.79 ± 0.9 | −0.69 ± 1.5 | −0.68 ± 1 | .937 | |
| Visual Attention Accuracy (z-score) | −0.91 ± 1.9 | −0.92 ± 1.7 | −1.29 ± 2.7 | −0.6 ± 1.4 | .541 | |
| Freedom from Distractibility (z-score) | −0.44 ± 0.8 | −0.71 ± 0.7 | −0.07 ± 0.6 | −0.39 ± 0.9 | .028 | SH vs SM |
| Phonemic Fluency (z-score) | −0.38 ± 1.1 | −0.55 ± 1.2 | −0.06 ± 1.1 | −0.5 ± 0.7 | .320 | |
| Semantic Fluency (z-score) | −0.7 ± 1.2 | −0.99 ± 1.2 | −0.77 ± 1.3 | −0.12 ± 0.9 | .097 | |
| Planning Ability (z-score) | −1 ± 2.2 | −2.02 ± 2.7 | −0.13 ± 1.4 | −0.61 ± 1.3 | .041 | SH vs SM |
PCF indicates posterior cranial fossa, PIQ indicates performance intelligence quotient, SD indicates standard deviation, SH indicates supratentorial hemispheric, SM indicates supratentorial midline. TIQ indicates total intelligence quotient, VIQ indicates verbal intelligence quotient. Value in IQ points, normal = 100 ± 15; z-scores, normal = 0 ± 1.
Ten (12%) of 83 patients had total IQ (TIQ) scores lower than 1 SD below the normal level (range, 72–85); 5 patients (6%) of 83 had TIQ scores lower than 2 SD below the normal level (range, 54–68). All patients with TIQs of <70 had a tumor sited in the temporo-mesial structures.
Comparison among children with different tumor sites showed a statistical significance for TIQ (P = .058), verbal IQ (P = .015), visual-motor integration (P = .003), phonological working memory (P = .030), freedom from distractibility (P = .028), and planning tasks (P = .041). Children with cortical tumors showed the worst performance in all tests. Post-hoc analysis evidenced a significant difference between patients with cortical locations and children with SM tumors.
Patients with SM lesions displayed the lowest scores in tasks exploring the accuracy of visuospatial analysis and visual memory, with a z-score lower than −1.
We found no clear-cut neuropsychological profile trend when the scores of children with cerebellar tumors were compared with those of patients with PCF masses not involving the cerebellum. Conversely, in tumors involving cerebellar hemispheres, children with right-side tumors were significantly more impaired in phonological working memory (P = .053), visual attention speed (P = .049), and visual-motor integration (P = .033) than were children with tumors in the left cerebellum.
In SH tumors, measures of phonological working memory were lower in cases involving left hemispheric tumor than in those with right hemisphere involvement (P = .033).
Statistical analysis revealed no significant difference between patient's age, gender, neurological impairment, and cognitive variables. Patients with symptom duration >100 days were significantly impaired in language tests (P = .056).
MANOVA demonstrated that TIQ was significantly different in patients with diverse grading of epilepsy (P = .026); post-hoc analysis revealed that significance progressively increased as a function of epilepsy grade: grade 0 versus grade 3, P = .021; grade 1 versus grade 3, P = .037; grade 2 versus grade 3, P = .043. Moreover, phonological working memory and freedom from distractibility were significantly impaired in children with epilepsy grade 3 compared with nonepileptic children (P = .055 and .006, respectively). Even performance in the test exploring planning abilities was affected by the presence of epilepsy (P = .016), with the poorest scores found in children with severe epilepsy.
Statistical analysis failed in finding significances between tumor volume, histotype, presence and/or amount of perilesional edema, hypertension, hydrocephalus, and neurocognitive performance.
Discussion
Only a small percentage of children affected by brain tumor can be cured by surgery alone. Post-surgical treatment, which includes chemotherapy and in most cases RT, is usually based on the extent of resection, histology, and presence of metastatic lesions at diagnosis. The potentially devastating effects of RT on growing individuals are well known. Indeed, as cure rates increased, it soon became apparent that irradiated children have a higher chance of cognitive disability.
We performed a detailed analysis of cognitive organization before surgery, RT, and other treatments in children with recently diagnosed cerebral tumors.
In our study, it appears that mean IQs lie within the range of normality, thus confirming the results of a previous study.16 However, it is worth noting that almost 12% of children exhibited IQ values lower than 85, and 6% of population displayed a true mental retardation, with mean IQ values of 59. Such values of IQ are indicative of mild to moderate mental retardation, which corresponds to a seriously disabling condition, with impairment of some basic functions (such as motor and language attainment) and severe limitations of learning capabilities.
Tumor site is the major determinant of cognitive deficits, while tumor dimension and tumor-related factors seem to exert less effect. In our population, the presence of low IQs appeared to be mainly related to SH locations. Previous reports on children with brain tumors observed at different stages of tumor treatment pointed to greater IQ impairment in those with hemispheric tumors.1,11 The present study shows that these data are present at diagnosis and are specific of certain anatomical sites—namely, the mesial areas of temporal lobes. In fact, all children with mental retardation documented at diagnosis had cortical tumors involving the mesial regions of the temporal lobes.
Tumor volume may seem to be a less crucial factor than site for cognitive deficits. Notably, some large tumors may have a more limited impact on neuropsychological functioning than small tumors located in critical areas (Fig. 1). The presence and/or amount of perilesional edema, compartment/localized hypertension, and hydrocephalus do not seem to be significantly correlated with cognitive data. This finding is to some extent unexpected, in view of the well-known neuropsychological deterioration depending on these conditions that has been widely demonstrated in hydrocephalus.21,22 However, it is known that the cognitive deficits in hydrocephalus are multiply determined, because neuropsychological impairments have been described mainly in persons with severe and chronic conditions.
Fig. 1.
Examples of a lack of correspondence between tumor size/mass-effect and neurocognitive status (A–C) Pre-operative MRI of a large astrocytoma (grade I) of the left cerebellar hemisphere. (A and B) Axial and sagittal post-contrast T1-weighted images. (C) Coronal T2-weighted images. Note the large cystic tumor, with contrast-enhancing walls, severe distortion of the cerebellum and brainstem, as well as secondary tonsillar herniation in the foramen magnum (due to infratentorial hypertension). Despite the obvious mass effect, the tonsillar herniation, and the co-existence of hydrocephalus, the patient did not have cognitive deficits. (D–F) Pre-operative computed tomography (CT) and MRI of small ganglioglioma of the left mesial temporal pole. (D) Unenhanced CT scan; (E and F) Axial and coronal T2-weighted images. The tumor is largely calcified, without edema; the lesion is located in the temporal uncus, with minimal mass effect. Despite this minimal focal mass effect, the patient was experiencing severe mental retardation and autistic behaviour.
In previous posttreatment studies, several risk factors were identified as associated with cognitive sequelae in survivors, including younger age, longer time since treatment, female sex, hydrocephalus, irradiation, RT dose, and the volume of brain irradiated.1,13 This study shows that symptom duration and epilepsy are the main clinical factors affecting cognitive status at diagnosis, whereas a child's age, sex, and neurological examination findings appear to be less prominent. Young age and female sex did not clearly appear as main risk factors for cognitive deficits at diagnosis in our population, most probably because they are related to the higher vulnerability of the growing female brain to neurotoxic effects of treatment, as evidenced in ALL studies.23
It has been shown that irradiation adversely affects the neural and synaptic processes during brain maturation13 and repair mechanisms following CNS lesions.
Brain plasticity is higher in the earlier ages and decreases during development. These data, extensively observed in laboratory studies24 and in early CNS focal lesions,25 theoretically mean that the more precocious the lesion, the more efficiently the healthy regions can take over the functions of the damaged areas.
Many studies demonstrate the detrimental effect of epilepsy on brain plasticity, so that performance in cognitive tests is possibly markedly reduced26 and declines with time in epileptic children.27 In our sample, the presence of seizures plays a role in cognitive organization, because neurocognitive functioning is significantly less efficient in children with epilepsy. Moreover, a direct correlation was found between the severity of epilepsy and the degree of cognitive dysfunction, with maximum cognitive impairment in children who presented with grade III epileptic syndromes. Because these children receive ≥1 antiepileptic drug, which are known to interfere with cognitive performances, the effect of epilepsy is most likely to be read as the sum of seizures and pharmacological agents.
Our evidence that specific and long-lasting symptoms could represent a risk factor for mental development may suggest that an early, radical surgical approach could protect some children from cognitive decline.
In cortical tumors neurological signs are long lasting and epilepsy is quite frequent, so a complex interaction of clinical variables and specific sites may put these children at a higher risk for cognitive impairment already at the time of diagnosis. Moreover, cortical tumors are usually slow-growing low-grade tumors, so the evaluation of the possible treatment procedure is often conditioned by the risk of major neurological impairment after surgery. However, the careful monitoring of clinical manifestations and the possibility of cognitive deterioration, mainly due to epilepsy, has to be taken into account from the outset before making any decision regarding a different treatment approach, even for the purpose of mitigating long-term cognitive and neurobehavioral sequelae.
Evaluation of different cognitive functions has shown that cognitive deficits could be present even in children with normal IQs, and that neuropsychological profile is often congruent with tumor location.
Patients with SM neoplasms had mainly defective memory abilities; a powerful correlation between linguistic measures and cortical left-side tumors was demonstrated. Moreover, visual-motor integration and planning capabilities are defective in patients with cortical tumors. These deficits were common and often not linked to motor impairment, because they could also appear in patients with mild or no neurological signs. The limited relationship between the neurological findings and neuropsychological scores in children studied before treatment is consistent with reports from previous studies16 and suggests the importance of conducting both neurological and neuropsychological examinations.
Data on PCF tumors are recurrent in literature reports, because these tumors occur more frequently in the pediatric population than in the adult population. Cognitive deficits and behavioral disturbances have been described in children surgically treated for cerebellar tumor. In our population, a detailed cognitive evaluation before treatment revealed that selective deficits could depend on the cerebellar localization of the tumor, indicating that the tumor itself affects cognitive functioning.28 The cerebellum is included in a cerebro-cerebellar network subserving several cognitive and emotional regulatory functions whose role is crucial during development. Planning ability—that is, the capacity to program efficiently the strategies needed to resolve a multilevel task—represents the main cognitive impairment in the children of our study with cerebellar tumors.29 The role of the cerebellum in the so-called frontal functions is widely demonstrated and interpreted as depending on the dysfunction of circuits connecting the cerebellum with the associative areas of the frontal cortex.30
The present study suggests the importance of a detailed neuropsychological evaluation at the time of diagnosis as a baseline to clarify the specific role of the tumor, surgery, and medical procedures in the possible cognitive decline of children with CNS tumors. Our data demonstrate that some children with brain tumors are severely affected at diagnosis and may require adequate psychological and educational support. These interventions should be started at an early stage and become part of the multi-disciplinary integrated treatment of children with brain tumors.
Conflict of interest statement. None declared.
Funding
This study was supported by grants from Fondazione per l'Oncologia Pediatrica and Associazione Ali di Scorta Onlus.
References
- 1.Mulhern RK, Butler RW. Neurocognitive sequelae of childhood cancers and their treatment. Pediatr Rehabil. 2004;7(1):1–14. doi: 10.1080/13638490310001655528. [DOI] [PubMed] [Google Scholar]
- 2.Butler RW, Haser JK. Neurocognitive effects of treatment for childhood cancer: mental retardation and developmental disabilities. Res Rev. 2006;12:184–191. doi: 10.1002/mrdd.20110. [DOI] [PubMed] [Google Scholar]
- 3.Wolfe-Christensen C, Mullins LL, Scott JG, McNall-Knapp RY. Persistent psychological problem in children who develop posterior fossa syndrome after medulloblastoma resection. Pediatr Blood Cancer. 2007;49:723–726. doi: 10.1002/pbc.21084. doi:10.1002/pbc.21084. [DOI] [PubMed] [Google Scholar]
- 4.Kull KR, Fatih Okcu M, Potter B, et al. Screening for neurocognitive impairment in pediatric cancer long-term survivors. J Clin Oncol. 2008;26:4138–4143. doi: 10.1200/JCO.2008.16.8864. doi:10.1200/JCO.2008.16.8864. [DOI] [PubMed] [Google Scholar]
- 5.Tucha O, Smely C, Preier M, Lange KW. Cognitive deficits before treatment among patients with brain tumors. Neurosurgery. 2000;47(2):324–334. doi: 10.1097/00006123-200008000-00011. doi:10.1097/00006123-200008000-00011. [DOI] [PubMed] [Google Scholar]
- 6.Garcìà Conde M, Martin Viota L, Febles Garcìa P, Ravina Cabrera J, López Almaraz R. Severe cerebellar mutism after posterior fossa tumor resection. An Pediatr (Barc) 2007;66(1):75–79. doi: 10.1157/13097364. doi:10.1157/13097364. [DOI] [PubMed] [Google Scholar]
- 7.Gelabert-Gonzáles M, Fernàndez-Villa J. Mutism after posterior fossa surgery: review of the literature. Clin Neurol Neurosurg. 2001;103:111–114. doi: 10.1016/s0303-8467(01)00125-1. doi:10.1016/S0303-8467(01)00125-1. [DOI] [PubMed] [Google Scholar]
- 8.Levisohn L, Cronin-Golomb A, Schmahmann JD. Neuropsychological consequences of cerebellar tumour resection in children. Brain. 2000;123:1041–1050. doi: 10.1093/brain/123.5.1041. doi:10.1093/brain/123.5.1041. [DOI] [PubMed] [Google Scholar]
- 9.Michaud LJ, Ried SR, Mcmahon MA. Rehabilitation of the child with cancer. In: Pizzo PA, Poplack DG, editors. Principles and Practice of Pediatric Oncology. 4th ed. Philadelphia: Lippincott Williams and Wilkins; 2002. pp. 1351–1364. [Google Scholar]
- 10.Bière ME, Scott JG, McNall-Knapp RY, Adams RL. Cognitive outcome in pediatric brain tumor survivors: delayed attention deficit at long-term follow-up. Pediatr Blood Cancer. 2008;50:337–340. doi: 10.1002/pbc.21223. doi:10.1002/pbc.21223. [DOI] [PubMed] [Google Scholar]
- 11.Reimers TS, Ehrenfels S, Mortensen EL, et al. Cognitive deficits in long-term survivors of childhood brain tumors: identification of predictive factors. Med Pediatr Oncol. 2003;40:26–34. doi: 10.1002/mpo.10211. doi:10.1002/mpo.10211. [DOI] [PubMed] [Google Scholar]
- 12.Mulhern RK, Phipps S, White H. Neuropsychological outcomes. In: Wallace H, Green D, editors. Late Effects of Childhood Cancer. London: Arnold; 2004. pp. 18–36. [Google Scholar]
- 13.Knab B, Connell PP. Radiotherapy for pediatric brain tumors: when and how. Expert Rev Anticancer Ther. 2007;7(suppl 12):S69–S77. doi: 10.1586/14737140.7.12s.S69. doi:10.1586/14737140.7.12s.S69. [DOI] [PubMed] [Google Scholar]
- 14.Iuvone L, Mariotti P, Colosimo C, Guzzetta F, Ruggiero A, Riccardi R. Long term cognitive outcome, brain CT scan and MRI abnormalities in children cured for ALL. Cancer. 2002;95:2562–2570. doi: 10.1002/cncr.10999. doi:10.1002/cncr.10999. [DOI] [PubMed] [Google Scholar]
- 15.Taphoorn MJ, Klein M. Cognitive deficits in adult patients with brain tumours. Lancet Neurol. 2004;3(3):159–168. doi: 10.1016/S1474-4422(04)00680-5. doi:10.1016/S1474-4422(04)00680-5. [DOI] [PubMed] [Google Scholar]
- 16.Brookshire B, Copeland DR, Moore BD, Ater J. Pretreatment neuropsychological status and associated factors in children with primary brain tumors. Neurosurgery. 1990;27(6):887–891. doi: 10.1097/00006123-199012000-00005. doi:10.1097/00006123-199012000-00005. [DOI] [PubMed] [Google Scholar]
- 17.Lazareff JA, Castro-Sierra E. Preoperative and postoperative analysis of visual and auditory memory in children with cerebellar tumors. Child's Nerv Syst. 1996;12:81–86. doi: 10.1007/BF00819501. doi:10.1007/BF00819501. [DOI] [PubMed] [Google Scholar]
- 18.Carreño M, Donaire A, Sánchez-Carpintero R. Cognitive disorders associated with epilepsy: diagnosis and treatment. Neurologist. 2008;14(6):S26–S34. doi: 10.1097/01.nrl.0000340789.15295.8f. doi:10.1097/01.nrl.0000340789.15295.8f suppl 1. [DOI] [PubMed] [Google Scholar]
- 19.Hamiwka LD, Wirrell EC. Comorbidities in pediatric epilepsy: beyond “just” treating the seizures. J Child Neurol. 2009;24(6):734–742. doi: 10.1177/0883073808329527. doi:10.1177/0883073808329527. [DOI] [PubMed] [Google Scholar]
- 20.Gobel V, Jurgens H, Etspuler G, et al. Prognostic significance of tumor volume in localized Ewing's sarcoma of bone in children and adolescents. J. Cancer Res Clin Oncol. 1987;113(2):187–191. doi: 10.1007/BF00391442. doi:10.1007/BF00391442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Erickson K, Baron IS, Fantie BD. Neuropsychological functioning in early hydrocephalus: review from a developmental perspective. Child Neuropsychology. 2002;7(4):199–229. doi: 10.1076/chin.7.4.199.8737. [DOI] [PubMed] [Google Scholar]
- 22.Matarò M, Junqué C, Poca MA, Sahuquillo J. Neuropsychological findings in congenital and acquired childhood hydrocephalus. Neuropsychol Rev. 2001;11(4):169–178. doi: 10.1023/a:1012904907249. doi:10.1023/A:1012904907249. [DOI] [PubMed] [Google Scholar]
- 23.Jain N, Brouwers P, Fatih Okcu M, Cirino PT, Krull KR. Sex-specific attention problems in long-term survivors of pediatric acute lymphoblastic leukemia. Cancer. 2009;115(18):4238–4245. doi: 10.1002/cncr.24464. [DOI] [PubMed] [Google Scholar]
- 24.Hoffman KL, Logothetis NK. Cortical mechanisms of sensory learning and object recognition. Philos Trans R Soc Lond B Biol Sci. 2009;364(1515):321–329. doi: 10.1098/rstb.2008.0271. doi:10.1098/rstb.2008.0271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Guzzetta A, Pecini C, Biagi L, et al. Language organisation in left perinatal stroke. Neuropediatrics. 2008;39(3):157–163. doi: 10.1055/s-0028-1085465. doi:10.1055/s-0028-1085465. [DOI] [PubMed] [Google Scholar]
- 26.Vargha-Khadem F, Isaacs E, Muter V. A review of cognitive outcome after lesions unilateral sustained during childhood. J Child Neurol. 1994;9(suppl 2):67–73. [PubMed] [Google Scholar]
- 27.Ballantyne AO, Spilkin AM, Hesselink J, Trauner DA. Plasticity in the developing brain: intellectual, language and academic functions in children with ischaemic perinatal stroke. Brain. 2008;131:2975–2985. doi: 10.1093/brain/awn176. doi:10.1093/brain/awn176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Steinlin M. Cerebellar disorders in childhood: cognitive problems. Cerebellum. 2008;7(4):607–610. doi: 10.1007/s12311-008-0083-3. doi:10.1007/s12311-008-0083-3. [DOI] [PubMed] [Google Scholar]
- 29.Godefroy O, Jeannerod M, Allain P, Le Gall D. Frontal lobe, executive functions and cognitive control. Rev Neurol (Paris) 2008;164(suppl 3):S119–S127. doi: 10.1016/S0035-3787(08)73302-2. [DOI] [PubMed] [Google Scholar]
- 30.Strick PL, Dum RP, Fiez JA. Cerebellum and nonmotor function. Annu Rev Neurosci. 2009;32:413–434. doi: 10.1146/annurev.neuro.31.060407.125606. doi:10.1146/annurev.neuro.31.060407.125606. [DOI] [PubMed] [Google Scholar]