- Email: [email protected]
Late neurocognitive sequelae in survivors of brain tumours in childhood
Reviews Late neurocognitive sequelae in survivors of brain tumours in childhood Raymond K Mulhern, Thomas E Merchant, Amar Gajjar, Wilburn E Reddick, and Larry E Kun
As survival among children treated for cancer continues to improve, more attention is being focussed on the late effects of cancer treatment. In children treated for brain tumours, chronic neurocognitive effects are especially challenging. Deficits in cognitive development have been described most thoroughly among children treated for posterior-fossa tumours, specifically medulloblastomas and ependymomas, which account for about 30% of all newly diagnosed cases of brain tumours in children. Most children who have survived brain tumours have required surgical resection and focal or craniospinal radiotherapy (irradiation of the entire subarachnoid volume of the brain and spine), with or without systemic chemotherapy. Historically, intelligence quotient (IQ) scores have provided a benchmark against which to measure changes in cognitive development after treatment. Observed declines in IQ are most likely a result of failure to learn at a rate that is appropriate for the age of the child, rather than from a loss of previously acquired knowledge. The rate of IQ decline is associated with a several risk factors, including younger age at time of treatment, longer time since treatment, female sex, as well as clinical variables such as hydrocephalus, use of radiotherapy and radiotherapy dose, and the volume of the brain that received treatment. Loss of cerebral white matter and failure to develop white matter at a rate appropriate to the developmental stage of the child could partly account for changes in IQ score. Technical advances in radiotherapy hold promise for lowering the frequency of neurocognitive sequelae. Further efforts to limit neurocognitive sequelae have included design of clinical trials to test the effectiveness of cognitive, behavioural, and pharmacological interventions. Lancet Oncol 2004; 5: 399–408
In 1969, HJG Bloom was one of the first physicians to recognise the threat that brain tumours and their treatment present to the quality of life of children through a study of 82 patients with medulloblastoma (figure 1) treated at the Royal Marsden Hospital, London, UK.1 These early observations showed the high risk of dementia associated with use of cranial or craniospinal radiotherapy for children younger than 2 years. Since then, studies that have assessed various components of quality of life (eg, sensory, motor, social, emotional, and cognitive components) have become increasingly sophisticated, and many compare patients with healthy children in the general population or with healthy controls matched for age and sex. These studies have
Oncology Vol 5 July 2004
Figure 1. Sagittal T1-weighted MRI showing planned cranial radiation dosimetry for contemporary treatment of medulloblastoma. Note the postsurgical tumour void anterior to the cerebellum and posterior to the brainstem. Dosimetry: magenta, 55·8 Gy; light blue, 36·0 Gy; and green, 36·0 Gy. Boundaries are progressively lower doses of radiotherapy. Conformal planning of dosimetry could help reduce neurocognitive sequelae by limiting irradiation of healthy structures in the brain.
confirmed that children treated for brain tumours are at increased risk of a poor quality of life. For example, a large cross-sectional survey2 that included 342 survivors of brain tumours in childhood and 479 sibling controls showed that former patients were 10·8 times less likely to be employed and 28·8 times less likely to be able to drive a car than their siblings. In terms of mental function, survivors were 16·2 times more likely to be incompetent; a greater likelihood of incompetence was found among survivors who were younger at time of treatment, who received cranial or craniospinal radiotherapy, or who had supratentorial tumours. RKM is Chief of the Division of Behavioral Medicine; TEM is Chief of the Division of Radiation Oncology; AG is Director of Neurooncology in the Department of Hematology/Oncology; WER is Director of Signal Processing, Division of Diagnostic Imaging; and LEK is Chair, Department of Radiological Sciences. All authors are at St Jude Children’s Research Hospital, Memphis, TN, USA. Correspondence: Dr Raymond Mulhern, Division of Behavioral Medicine, St Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, TN 38105-2794, USA. Tel: +1 901 495 3580. Fax: +1 901 495 3121. Email: [email protected]
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
399
Review
Neurocognition and brain tumours
Here, we review the neurocognitive sequelae of posterior-fossa brain tumours of childhood and their treatment. Discussion is restricted to late effects on neurocognition, or cognitive problems that arise 2 or more
years after completion of treatment, as opposed to acute or subacute complications of treatment. Emphasis on late effects encompasses a range of neurocognitive problems that are therefore expected to be chronic, if not progressive, in
Table 1. Studies of intellectual development in survivors of medulloblastoma and ependymoma Sample
n
Study design
Treatment
Results
Ref
Medulloblastoma
50
Longitudinal
Radiotherapy (n=50) CSI: 35–40 Gy PF: 51–59 Gy
Mean loss of 2·2 IQ points per year after radiotherapy Greater IQ deficits with younger age at radiotherapy Greater IQ deficits among female patients
11
Medulloblastoma
44
Longitudinal
Radiotherapy (n=44) CSI: 23·4–48·0 Gy PF: 49–55 Gy Chemotherapy (n=23)
Mean loss of 2·6 IQ points per year after radiotherapy Rate of learning 50–60% of normal Greater IQ deficits with younger age at radiotherapy
12
Medulloblastoma
42
Cross-sectional
Radiotherapy (n=42) CSI: 49–54 Gy PF: 23·4–36·0 Gy Chemotherapy (n=29)
Greater IQ deficits with younger age at radiotherapy Greater IQ deficits with increasing time from radiotherapy
13
Medulloblastoma
43
Longitudinal
Radiotherapy (n=43) CSI: 23·4 Gy PF: 32·4 Gy
Mean IQ loss of 4·3 points per year after radiotherapy Greater IQ deficits with younger age at radiotherapy
14
Medulloblastoma
18
Cross-sectional
Radiotherapy (n=18) CSI: 23·4–36·0 Gy PF: 49–54 Gy Chemotherapy (n=9)
Mean IQ lower in patients with medulloblastoma than age-matched patients with low-grade astrocytoma
15
Low-grade astrocytoma Medulloblastoma
18 22
Cross-sectional
Medulloblastoma
25
Cross-sectional
Medulloblastoma
19
Cross-sectional
Ependymoma
12
Medulloblastoma
59
Ependymoma
37
Very young (<4 years) PF
27
Longitudinal
Very young (<4 years) medulloblastoma
19
Longitudinal
Very young (<4 years) medulloblastoma
37
Cross-sectional
Very young (<4 years) medulloblastoma
10
Longitudinal
Surgery only Radiotherapy (n=22) PF: 54 Gy (n=22) CSI: 36·0 Gy (n=13) CSI: 23·4 Gy (n=9) Radiotherapy (n=25) (dose not specified) Chemotherapy (n=7) Radiotherapy (n=19) CSI: 25–35 Gy (n=19) PF: 55 Gy (n=19) Chemotherapy (n=17)
Greater IQ deficits with younger age at radiotherapy
17
Greater IQ deficits with increased time from radiotherapy IQ score lower in patients with medulloblastoma than in patients with ependymoma Patients with medulloblastoma given 25 Gy (CSI) had better IQ scores than patients given 35 Gy (CSI)
Radiotherapy (n=12) Chemotherapy (n=4) PF: 25 Gy (n=11) Cross-sectional
15 16
Greater IQ deficits with younger age at radiotherapy Greater IQ deficits in patients given 36 Gy (CSI)
18
18
Radiotherapy (CSI, PR; n=59)
10% of patients had IQ of greater than 90
19
Radiotherapy (PF, n=28)
60% IQ >90 in ependymoma
19
Radiotherapy (n=7) CSI: 30–40 Gy PF: 40–50 Gy Chemotherapy (n=20) Radiotherapy (n=19) CSI: 35·2 Gy PF: 53·4 Gy Chemotherapy (n=19) Radiotherapy (n=37) CSI, PF doses varied Chemotherapy (n=15) Radiotherapy (n=10) CSI: 18 Gy PF: 54 Gy Chemotherapy (n=10)
Greater decline in IQ with radiotherapy
20
Mean IQ loss of 3·9 points per year after radiotherapy
21
8 of 16 long-term survivors in special schools
22
No change in baseline IQ score at 3 years after radiotherapy
23
CSI, craniospinal irradiation; PF, posterior fossa.
400
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Review
Neurocognition and brain tumours
long-term survivors. Other types of late effects (such as endocrinopathies) are reviewed elsewhere.3 First, we briefly review the epidemiology and contemporary approaches for the treatment (including conformal radiotherapy) of two histologically distinct tumours of the posterior fossa—ependymoma and medulloblastoma. The core and secondary neurocognitive symptoms associated with treatment of these tumours are presented along with known or suspected factors that place subsets of children at increased risk of cognitive deficits, such as area of radiotherapy dose and the dose given to some brain structures. Figure 1 shows the planning of radiotherapy for a postoperative patient with medulloblastoma. We also discuss changes that occur in the developing brain that can be detected by neuroimaging and that could explain the process of neurocognitive decline. Finally, we describe three methods of treatment that try to avoid or limit the augmentation of potentially unavoidable neurocognitive deficits: threedimensional conformal radiotherapy, cognitive or behavioural remediation, and pharmacotherapy.
Epidemiology and contemporary treatment There are two recognised peaks in the incidence of brain tumours in children and adolescents (<18 years). The first peak, with a value of 2·2–2·5 cases per 100 000 children per year, occurs the first decade of life; there is a slight male predominance at this age. This first incidence peak accounts for most cases of CNS embryonal tumours and ependymomas located in the posterior fossa. A second, much larger peak occurs in late adolescence and early adulthood, and characteristically accounts for most cases of glial tumours that occur mainly in the supratentorial compartment.4–6 A multidisciplinary team that involves specialists in neurosurgery, diagnostic imaging, radiation oncology, neuro-oncology, rehabilitation medicine, and behavioural medicine is the cornerstone for delivery of contemporary therapy at most academic neuro-oncology centres. Therapy for children diagnosed with medulloblastoma has been refined over the past 30 years. Early studies emphasised strategies to increase the rate of cure, whereas present therapeutic protocols in North America emphasise a riskbased approach to lower neurotoxicity. In patients staged with clinically standard-risk disease (defined as little or no gross evidence of tumour as shown by postoperative MRI and no evidence of metastatic disease), emphasis on a decrease in the dose of radiation to the neuraxis—including full-cranial radiation—has been a central strategy to lower the frequency of long-term neurocognitive deficits observed in earlier studies. In patients at high risk (defined as patients with >1·5 cm2 residual disease after surgery, or presence of metastatic disease), high-dose radiotherapy to the neuraxis is still the standard of care.7 Several strategies have been used to improve outcome for high-risk patients, but early results from a study that used high-dose adjuvant chemotherapy with stem-cell support seem to be most promising.8 At present, although new therapeutic approaches are being developed for high-risk patients, many could experience the more serious consequences of use of high-dose radiotherapy to the whole brain.
Oncology Vol 5 July 2004
Complete surgical resection followed by radiotherapy is the standard of care for intracranial ependymoma. Advanced neurosurgical techniques have facilitated the procedure of gross total resection of this tumour, with acceptable morbidity. New techniques of conformal radiation have enabled precise delivery of radiotherapy with less damage to the surrounding brain tissue. Early results from a study that used three-dimensional conformal radiotherapy are encouraging, especially because preliminary neurocognitive data show no adverse effect of use of this technique.9 Longterm neurocognitive follow-up is needed to validate the use of conformal radiotherapy as the standard of care for patients with intracranial ependymoma. Treatment approaches for infants diagnosed with medulloblastoma and ependymoma have previously been based on the delaying of radiotherapy to the infant brain to preserve neurocognitive function. Several studies have used an extended chemotherapy regimen after surgical resection to maintain disease control.10 For infants with localised disease, the emphasis on contemporary protocols has now shifted to use of early focal radiotherapy in an attempt to prevent local tumour recurrence. Early results from studies that have used this treatment strategy seem promising. For infants with disseminated disease, outcome remains poor because use of radiotherapy to the neuraxis is avoided.
Late effects of tumour and treatment The most common method used to document neurocognitive effects in children treated for brain tumours has been IQ tests with a normative mean score of 100 and standard deviations of 15 or 16. IQ tests are standardised on large numbers of the general population, so scores can be corrected for the age of the person who is to complete the test. The IQ score of a healthy person is thus not expected to change substantially over time. Table 1 summarises findings from 12 studies of IQ outcomes in children treated for medulloblastoma or ependymoma.11–22 Despite substantial variability in the methods used in the studies, several conclusions can be made. First, longitudinal studies consistently show significant declines in IQ over time in patients treated given craniospinal radiotherapy for medulloblastoma. Second, cross-sectional comparative studies have shown less severe deficits in IQ in those for ependymoma by use of postoperative irradiation confined to the posterior fossa alone, or with patients treated for lowgrade astrocytoma by use of surgery alone, than in those treated for medulloblastoma with craniospinal radiotherapy. Third, increasing age at time of treatment increases the risk of a decline in IQ. Studies that focused on very young patients (typically patients younger than 3 or 4 years old at time of diagnosis) showed that changes in IQ can have devastating effects; one series21 on salvage irradiation showed a median IQ score of 62 at 5 years after treatment. Although a decline in IQ has mainly been attributed to the effects of radiotherapy, other factors—such as hydrocephalus and posterior-fossa syndrome—are also known to have affect cognitive development.23 Furthermore, very young children have lower cognitive scores before radiotherapy treatment than their older counterparts,
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
401
Review
Neurocognition and brain tumours
Table 2. Main cognitive deficits among survivors of medulloblastoma Sample
n
Treatment
Main deficits
Findings
Ref
Medulloblastoma
42
Radiotherapy (n=42) CSI: 49–54 Gy PF: 23·4–36·0 Gy Chemotherapy (n=29)
Verbal memory Visual attention
Greater deficits with increasing time from radiotherapy Greater deficits with younger age at radiotherapy
13
Medulloblastoma
22
Radiotherapy (n=22) PF: 54 Gy (n=22) CSI: 36·0 Gy (n=13) CSI: 23·4 Gy (n=9)
Attention
Greater deficits in younger patients (<8 years at diagnosis) Greater deficits in patients given 36 Gy (CSI)
16
Radiotherapy (n=27) CSI: 30–40 Gy PF: 40–50 Gy Chemotherapy (n=20)
Attention
Overall decline greater in patients given radiotherapy than patients given no radiotherapy Overall decline Overall decline greater in patients given radiotherapy than patients given no radiotherapy
20
Very young (<4 years) 27 PF
Verbal memory Spatial memory
CSI, craniospinal irradiation; PF, posterior fossa.
implying that very young children are more vulnerable to the effects of the tumour, surgery, and perioperative factors.21 Another important point is that, because of assumptions of the normal distribution of IQ in the general population, 50% of patients would be predicted to have an IQ score of less than 100. A decline in IQ has frequently been associated with a commensurate decline in areas of basic academic achievement,14–16,20 but these neurocognitive outcomes are not functionally specific enough to help elucidate the neural networks that may be involved, or to help the process of rehabilitation. Increasingly, knowledge-based measures of IQ and academic achievement are viewed as the distant indicators of core cognitive functions such as attention, speed of processing, and working memory that provide the foundations for the ability to learn efficiently and retain information.13,16,20 This view is reinforced by evidence that normal, age-related improvements in processing speed and working memory account for nearly half of the age-related improvements that occur in intelligence.24 Table 2 summarises studies that have emphasised assessment of these core deficits in patients who survived medulloblastoma.13,16,20 In general, these studies confirm that deficits in the core areas of attention and memory are common,
Figure 2. External view of segmented cerebral cortex (left). Interior view of healthy white matter (green) segmented from the cerebral hemispheres with lateral ventricles shown in the segmented white matter to improve orientation (right).
402
and—similarly to IQ—that the more severe core deficits are associated with radiation treatment at a younger age, increased time from radiotherapy, and higher doses of radiotherapy. However, whether these core deficits truly cause declines in IQ and academic achievement has not yet been shown in children who survive brain tumours.
Neurodevelopmental consequences The pathophysiology of late CNS damage induced by radiation is not fully understood, especially with regard to the vulnerability of white matter to injury. Some hypotheses attribute primary mechanisms of CNS damage to the death of neuronal cells, oligodendrocytes, or endothelial cells and the subsequent microvascular damage. Ultimately, this damage seems to be a primary or secondary effect associated with administration of treatment to the CNS.25 Secondary processes, such as damage to the myelin membrane as a result of oxidative stress after radiotherapy, have also been proposed as putative mechanisms of CNS damage.26 Furthermore, study of the adverse effects of irradiation on the microenvironment of neural precursor cells has helped to integrate the findings from these competing hypotheses.27 At present, research on imaging techniques has progressed beyond diagnosis and surveillance to begin an assessment of neuropathological and neurodevelopmental correlates. One mechanism thought to account partly for the observed rate of IQ decline in survivors is loss of cerebral white matter or failure to develop white matter at a rate appropriate for the developmental stage of the survivor. This mechanism has begun to receive more attention because of newly developed methods for quantification of volumes of brain tissue (figure 2).28 Studies that have quantified toxic effects on white matter and investigated the association between neurotoxicity and cognitive deficits in children have mainly focused on survivors of medulloblastoma. For example, in a comparison of patients treated for medulloblastoma with age-matched controls who had received surgery alone for treatment of lowgrade tumours of the posterior fossa, survivors of medulloblastoma had a significantly smaller volume of cerebral white matter, a substantially greater volume of
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Neurocognition and brain tumours
Review
Volume (%)
attention, memory, and academic achievement. The researchers produced a developmental model in which academic achievement could be predicted by the volume of cerebral 54·0 54·0 54·0 white matter, ability to show attention, 36·0 18·0 23·4 and IQ score. These factors explained about 60% of the variance observed in tasks that measured reading and 100 spelling, and almost 80% of the 90 variance in mathematical tasks. The main cognitive change associated with 80 lower volume of cerebral white matter 70 was a decreased ability to pay attention, which was also associated 60 with a decrease in IQ score and 50 academic achievement. The relation between attention and cerebral white 40 matter in the same group of patients 30 was investigated in more detail by use of a self-administered computer test of 20 Total Brain DVH visual attention. After controlling 10 statistically for the effects of age at diagnosis and time elapsed from 0 treatment, the study showed a 0 10 20 30 40 50 60 70 significant association between functDose (Gy) ions that require attention and Figure 3. Benefits of dose decreases in planning of craniospinal radiotherapy shown with total-brain volumes of cerebral white matter, or dose-volume histograms (DVH), comparison of 36·0 Gy (yellow), 23·4 Gy (red), and 18 Gy (blue). specific regional volumes of white matter in the prefrontal, or frontal, cerebrospinal fluid, and an equal volume of grey matter.28 As lobe and cingulate gyrus.32 Although reliable, associations expected, survivors of medulloblastoma had significantly between decreased white matter and decline in cognitive lower IQ scores, which had a positive and statistically function could represent the influence of a third, as yet significant association with volumes of cerebral white unidentified variable that has an adverse effect on both the matter.15 However, because of the cross-sectional design of development of white matter and cognition. the studies, it has not been possible to discern whether the The next logical step in this line of investigation would smaller volume of cerebral white matter is due to loss of be to integrate maps on digital radiation dosimetry with the tissue, failure to develop white matter at an appropriate rate, data on tissue volume to assess the relation between the or both. volume of white matter and dose response. A crucial dose Subsequently, a longitudinal study showed a significant threshold could be established and regional analyses could decrease in the volume of cerebral white matter in patients be combined with more sensitive and specific neurowho underwent treatment for medulloblastoma; this cognitive testing. These studies are necessary to establish the decrease was more rapid in patients who received 36 Gy precise relation between therapy, the location of the craniospinal irradiation than in patients who received developmental area causing cognitive deficit, the location of 23·4 Gy craniospinal irradiation.29 These findings that the decrease in the volume of cerebral white matter, and the associate cerebral white matter with irradiation dose were effect on neurocognitive performance. confirmed in another longitudinal study on volumes of Global or regional volumes of neuronal structures might corpus callosum, which found the greatest deviation from be relatively insensitive to more subtle changes that occur in normal development occurred in the most posterior the microstructure and organisation of fibre tracts in white subregions of the brain—the area that also received the matter. Other updated measures of the integrity of white highest total dose of irradiation.30 matter should also be incorporated, such as diffusion tensor Previous studies have established a relation between the imaging of apparent diffusion coefficient and fractional volume of cerebral white matter and both radiation dose and anisotropy. The apparent diffusion coefficient measures the IQ. A further study established that the volume of cerebral average diffusion distance in a region, which reflects the white matter could explain about 70% of the association volume ratio of extracellular to intracellular white matter between IQ score and age at time of irradiation.13 A cross- and is also a sensitive measure of acute ischaemia. Fractional sectional study, by Reddick and colleagues,31 of patients anisotropy measures the directional organisation of a region treated for medulloblastoma showed significantly impaired and thus reflects myelin integrity. A pilot study of children performance on all neurocognitive measures of intellect, who survived medulloblastoma found significant decreases
Oncology Vol 5 July 2004
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
403
Review
Neurocognition and brain tumours
fractionation; chemotherapy was not included in many cases. To address mainly the neuropsychological sequelae commonly seen in long-term survivors, patients with average-risk medulloblastoma were randomly assigned a lower dose of craniospinal radiotherapy (up to 23·4 Gy or 36·0 Gy).36 In a separate study, results with a lower dose of craniospinal irradiation combined with multiagent chemotherapy compared favourably with those achieved by standard-dose craniospinal irraditation. Combination treatment has since been used as standard treatment for average-risk patients, on the premise that a lower dose of craniospinal irradiation would have fewer long-term Innovations in radiotherapy effects.37 Despite decreases in the dose of craniospinal The side-effects attributed to radiotherapy have been a main radiotherapy, cognitive impairment remains significant even concern in the design of clinical trials for medulloblastoma with a decrease in IQ of 4·3 points per year reported from a during the past 20 years. Concern has been greatest for national trial of low-dose craniospinal radiotherapy and patients with average-risk medulloblastoma, for whom long- chemotherapy.14 Although a decreased dose of craniospinal radiotherapy term control of disease is likely and the side-effects of therapy might be long-lasting. Radiotherapy for can help to reduce the dose given to the entire brain and medulloblastoma is technically demanding, requiring both temporal lobes, it does little to decrease the volume of tissue craniospinal radiotherapy and boost treatment of the that receives the highest doses because the boost dose is anatomic posterior fossa. If the prescribed dose of radiation prescribed to the entire posterior fossa (figure 3). For patients can be delivered to these well-defined volumes, there is an at average risk, lowering of the boost volume from the entire posterior fossa to the tumour bed with a limited margin can increased likelihood of obtaining disease control.35 Only 10 years ago, the standard of care for patients with significantly decrease the volume of healthy tissue that average-risk medulloblastoma was postoperative cranio- receives the highest doses (figure 4). Results obtained spinal radiotherapy up to a dose of 36 Gy and boost with smaller target volumes and sophisticated, image-guided treatment up to 54 Gy by use of 1·8 Gy per day conventional radiation techniques indicate excellent disease control; the intended improvement in functional measures—including neurocognitive status—awaits further confirmation.38 23·4 23·4 A larger study on decreased irradiation of target volume that used focal, image-guided delivery of 54·0 54·0 radiation has been reported in preliminary form. Among 84 children, the cumulative rate of failure of the posterior fossa at 3 years from 100 diagnosis was 6·3%.39 The treatment 90 regimen for patients enrolled on the trial (SJMB-96) included postoperative 80 Total Brain DVH craniospinal radiotherapy up to 70 23·4 Gy, conformal posterior fossa boost up to 36·0 Gy, and focal 60 treatment of the tumour bed with a 50 2-cm margin. High-dose chemotherapy with peripheral stem-cell 40 support was administered after 30 radiotherapy. Lowering of the target volume to more closely encompass the 20 tumour bed can further spare the 10 temporal lobes from the effects of irradiation—a relation believed to be 0 important in preservation of 0 10 20 30 40 50 60 70 neurocognitive function. The current Dose (Gy) trial at St Jude Children’s Research Hospital, TN, USA (SJMB-03) is using Figure 4. Benefits of dose decreases in planning of radiotherapy to posterior fossa shown with totalfocal treatment of the primary site of brain dose-volume histograms (DVH), comparison of conventional boost (blue) to posterior fossa with conformal boost (yellow) to the primary site after 23·4 Gy craniospinal irradiation. the tumour after craniospinal Volume (%)
in fractional anisotropy of white matter, which correlated significantly with deterioration of performance at school.33 Functional MRI, a method of mapping brain activation, has been used extensively to investigate and characterise the basic neural networks that support normal cognitive function and disease-associated changes in the networks. MRI might also provide important insights into the relation between morphological abnormalities induced by disease and treatment, as well as the behavioural deficits that negatively affect the quality of life of survivors of medulloblastoma.34
404
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Review
Neurocognition and brain tumours
radiotherapy with a limited margin of 1 cm. About 25% of the posterior fossa will be spared from receiving the prescribed boost dose of 55·8 Gy. In the Children’s Oncology Group ACNS0331—an ambitious four-group randomised trial for children aged 3–8 years—participants with medulloblastoma are assigned 18·0 Gy or 23·4 Gy craniospinal radiotherapy, followed by a second randomisation to either conventional boost treatment or treatment of less than the entire posterior fossa. For patients older than 8 years, the craniospinal dose will remain at 23·4 Gy. The risks associated with a lowering of the dose of craniospinal radiotherapy and limitation of the boost treatment to the postoperative tumour bed with a limited margin must be balanced against the observed effects of radiotherapy on cognition, neurological function, growth, and development. Dose-volume data can now be acquired for functional regions of the brain and correlated with outcome for the patient.40 Establishment of dose-volume relations can be used to predict side-effects of treatment and could thus help tailor treatment. The usefulness of these models will be increased if factors other than radiation dose and age are tested for inclusion: tumour-related insults, effects of surgery, hydrocephalus, combination chemotherapy, and other factors should be investigated so that accurate models of the effects of treatment can be developed. Only then will we be able to investigate whether the planned deceases in dose for treatment in clinical trials are likely to be beneficial. Investigators in Europe and the UK are testing different strategies to decrease treatment-related toxic effects by use of hyperfractionated radiotherapy for craniospinal radiotherapy and the boost phase.35 The underlying idea is that smaller doses administered twice a day will decrease the occurrence of long-term effects. Hyperfractionation seems to be equally effective when given once a day, provided the total dose is increased slightly. Comparison of disease control and functional outcomes between the strategies of the American, European, and UK cooperative groups might some day provide the basis for a worldwide trial for patients with medulloblastoma.
Cognitive remediation Few published studies have assessed cognitive behavioural interventions for survivors of brain tumours in childhood. One case study investigated whether a compensatory memory notebook would improve neurocognitive function in an adolescent with severe memory impairment.41 A calendar section was used to record forthcoming activities and events, and to keep track of time. The notebook had a section on things to do for forthcoming assignments; an orientation section contained the names of teachers and the counsellor, classroom numbers, and other important personal information. Finally, a transport section included a bus schedule and maps of the school. The patient was systematically trained to use the memory book. On testing before and after the intervention, the patient showed signs of significantly impaired memory, although academic achievement increased slightly and there was an improvement in class attendance and timely completion of assignments.
Oncology Vol 5 July 2004
The most systematic effort to apply ideas of cognitive remediation to children with neurocognitive deficits as a result of cancer and subsequent treatment is the programme developed by Butler, Copeland, and colleagues.42 Their tripartite model uses techniques and methods from three disciplines: brain-injury rehabilitation, special education (or educational psychology), and clinical psychology. The programme consists of 20 sessions, each of 2 h, with an individual therapist. One component, called attentionprocess training, addresses deficits in control of sustained, selective, divided, and executive attention.43 If a patient does not obtain a score of 50% during a task, the activity is substituted for a more basic task. Once the child reaches 80% accuracy, the next stage of difficulty for a given test is applied. Each patient also receives instruction in metacognitive strategies, which are grouped within the three areas of task preparedness, on-task performance, and posttask strategies. Other methods to study cognitive behaviour specifically assess the ability to withstand distraction. Cognitive-behavioural psychotherapy is also used, which includes reframing of cognitive struggles in a positive light, psychotherapeutic support, acknowledgment of weaknesses, and blocks to successful improvement; other components are monitoring of internal dialogue, stress management, support for the patient to become his or her own best friend, and reinforcement of realistic, positive, and optimistic selfstatements. Preliminary data on efficacy have suggested an improvement in laboratory measures of abilities in attention, but no improvement in academic achievement.42 The results of a continuing phase III clinical trial based at seven institutions in the USA should soon be ready for publication.
Pharmacotherapy Research done over the past 50 years in children diagnosed with attention deficit hyperactivity disorder (ADHD) who are otherwise healthy has shown the effectiveness of stimulant medications—most commonly methylphenidate—to improve cognitive performance.44 Methylphenidate is a mixed dopaminergic–noradrenergic agonist that is thought to improve the function of the attention network in the frontostriatal region of the brain.45 Use of methylphenidate is found to have the most consistent and significant benefits in measures of vigilance and sustained attention, but improvements in reaction time, paired-associate learning, and perceptual efficiency are also commonly seen.44 Positive effects of methylphenidate on higher-order cognitive functions, such as problem-solving or language processing, have not been observed. Studies that assess the effects of methylphenidate on actual academic achievement—as opposed to behavioural improvements in the classroom—are equivocal among children with ADHD. Improvement in academic learning attributed to the action of methylphenidate is mainly the result of the drug’s effect on attention and ability to concentrate. In children with ADHD, methylphenidate might also improve formation of social relationships with peers, as assessed by behavioural observations and peer ratings.44 There have been very few studies on the
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
405
Review effectiveness of methylphenidate for attention problems in children who experience various forms of acquired brain damage. There is evidence that methylphenidate can be as effective against attention problems in children with epilepsy as against ADHD, without a significant lowering of the seizure threshold.45 Two preliminary studies were among the first to investigate the use of methylphenidate in children with learning problems that were presumed to be secondary to cancer treatment. In the first of these studies, 12 children who survived either brain tumours or acute lymphoblastic leukaemia were treated with methylphenidate for 6 months to 6 years (median duration of treatment was 23 months). The response was rated as good in eight children, fair in two, and poor in two.46 In the second study, six children who had received craniospinal radiotherapy 3–12 years earlier for treatment of brain tumours were given methylphenidate.47 With a consistent dose of 0·3 mg/kg methylphenidate, there was no significant immediate, or delayed, benefit to patients. An open-label study on adults treated for malignant brain tumours used neurocognitive tests and objective inventories to quantify the response to methylphenidate.48 The study found significant improvements in cognitive function (including psychomotor speed, memory, and executive functions) as well as mood and activities of daily living, commonly despite the presence of progressive disease. Thompson and co-workers49 investigated the effects of methylphenidate in a randomised, double-blind trial on survivors of acute lymphoblastic leukaemia or malignant brain tumours who had impaired learning. The researchers screened for a particular cognitive phenotype that was thought to elicit an optimum response to methylphenidate treatment—deficits in academic achievement and a concurrent problem with vigilance. 32 children were randomly assigned placebo or methylphenidate (0·6 mg/kg, up to maximum dose of 20 mg). They were retested on a completion of a computerised continuous-performance test of vigilance 90 min after the treatment. The group assigned methylphenidate showed significantly greater improvement than the placebo group. The results of that 1-day study encouraged implementation of a multisite trial on the effects of longterm exposure to methylphenidate, which has similar criteria for eligibility and exclusion to the study by Thompson and co-workers.49 After an initial laboratory challenge with methylphenidate, children participate in a randomised, double-blind, 3-week home crossover trial. Participants take two capsules each day for 5 days per week: 1 week of placebo, 1 week of low-dose methylphenidate (0·3 mg/kg, up to a maximum of 10 mg), and 1 week of moderate-dose methylphenidate (0·6 mg/kg, up to a maximum of 20 mg). At the end of each week, parents and teachers complete the Conners’ rating scales of attention and behaviour, and also complete scales to assess side-effects. At the end of the third week, the masking code is broken and the ratings are compared. If a patient has shown objective improvement with methylphenidate, the medication is continued for the next 12 months, then attention function and school achievement are reassessed. At present,
406
Neurocognition and brain tumours
preliminary analyses are in progress. Of the first 325 patients screened for participation in the trial, 40% qualified for methylphenidate treatment and 70% of those agreed to participate in the trial. Most of the participants have shown at least a minimum response to treatment with methylphenidate, and the parents of most of those who have shown a response agreed to continue treatment with methylphenidate. Whether this drug ultimately facilitates academic achievement in survivors of childhood cancer who have problems with attention is not yet known.
Environmental interventions In addition to cognitive remediation and pharmacological approaches to overcome neurocognitive deficits, the importance of ecological or environmentally based interventions for children with brain injuries—especially with regard to school—should not be underestimated.50 Many such children might need extended time limits for completion of school examinations, use of true-or-false questions and multiple-choice formats in tests rather than essay-based examinations, and encouragement to record classroom lectures for later review. Other ecological interventions include use of written handouts to decrease demands for copying of material from the blackboard, and substitution of computers for handwritten assignments. Whether the main approach to treatment is cognitive remediation, pharmacotherapy, ecological intervention, or a combination, explanation of the affected areas of neurocognitive deficit to patients, providers of care, and others in the community (eg, primary health-care providers and especially teachers) is important. Leigh and colleagues50 have recommended that all paediatric oncology centres should have a structured programme of school re-entry, a component of which should be education of the patient’s teacher about childhood cancer and the specific signs, symptoms, and special needs associated with the patient’s treatment and the likely outcome. Some special arrangements are obvious, such as ensuring that a patient with hemiparesis has extra time to move between classrooms. However, some neurological and neurocognitive deficits can be quite subtle, including mild loss of vision or hearing, or slow visual-motor production. Furthermore, the neurocognitive status of the patient might not remain stable after completion of treatment. The onset of some deficits can be delayed, and other deficits might not be evident until the ability is normally expected.51 One of the greatest dangers in the lack of communication of new problems to the patient’s teachers is that struggles the child may have in the classroom could be wrongly attributed to a lack of motivation, attitude problems, daydreaming, or emotional maladjustment. Simple arrangements in the classroom, such as fewer items on multiple-choice tests, preferential seating in the classroom, and lower expectations for amount of homework are sometimes needed. Since older children generally have many teachers in a school year and that most, if not all, of teachers change with each school year, communication of older patients’ special needs can be more difficult. Many children undergoing treatment for cancer, as well as those who have completed treatment, will be classified as “other
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Review
Neurocognition and brain tumours
Search strategy and selection criteria Data for this review were identified by searches of PubMed with the terms “children”, “ependymoma”, “medulloblastoma”, “PNET”, and “posterior fossa”. Only papers published in English since 1990 were selected. Additional papers were identified from references of relevant articles and for historical value.
4 5 6 7
health impaired” by their local school systems to access additional resources for their special needs. One area of progress in the development of interventions for cognitive deficits in children has been the increase in appreciation of the influence of the family environment and resources available after recovery from traumatic brain injury. Chaotic and dysfunctional family environments have a substantial adverse effect on neurocognitive recovery from traumatic brain injury in school-age children. This relation seems to be valid, even when there is control for severity of brain injury and other medical factors.52 The implications of these findings for the potential success of interventions to remediate neurocognitive deficits in children who survive brain tumours has not yet been investigated.
8
9
10
11 12 13
Conclusions Neurosurgeons, paediatric oncologists, and radiation oncologists have been successful in improving cure rates for most types of childhood brain tumours, including those of the posterior fossa. Nevertheless, the risks of neurocognitive impairment remain substantial, especially among individuals who were treated aggressively and at a young age. Thus, there is an imperative need for effective, but less neurotoxic, tumour therapy. Until such therapy is developed, cognitive, behavioural, pharmacological, and environmental interventions will be needed to address the neurocognitive effects that cannot be avoided. However, these interventions are only beginning to be investigated in randomised clinical trials, and information on the efficacy of different interventions will not be available for some time. We strongly encourage the further development of innovative strategies to preserve cognitive function while cure rates for children with malignant brain tumours continue to be maintained and improved. Conflict of interest
14
15 16
17 18 19
20 21
None declared. Acknowledgments
Our research is supported by the American Lebanese Syrian Associated Charities (ALSAC) and Cancer Center Support (CORE) grant P30CA21765, R01CA78957, U01CA81445, and R01CA90246 from the US National Cancer Institute.
22 23
References
1 Bloom HJG, Wallace ENK, Henk JM. The treatment and prognosis of medulloblastoma in children. AJR Am J Roentgenol 1969; 105: 43–62. 2 Mostow EN, Byrne J, Connelly RR, Mulvihill JJ. Quality of life in long-term survivors of CNS tumours of childhood and adolescence. J Clin Oncol 1991; 9: 592–99. 3 Strother DR, Pollack IF, Fisher PG, et al. Tumours of the central nervous system. In: Pizzo PA, Poplack DG, eds. Principles and
Oncology Vol 5 July 2004
24 25 26
practice of paediatric oncology, 4th edn. Lippincott, Williams and Wilkins: Philadelphia, 2002. Gurney JG, Severson RK, Davis S, Robinson LL. Incidence of cancer in children in the United States. Cancer 1995; 75: 2186. Young J, Miller R. Incidence of malignant tumours in US children. J Pediatr 1975; 86: 254. Schoenburg B, Christine B, Wishnant J. The descriptive epidemiology of primary intracranial neoplasm of childhood: a population study. Mayo Clin Proc 1976; 51: 51. Ellison D, Clifford SC, Gajjar A, Gilbertson RJ. What’s new in neurooncology? Recent advances in medulloblastoma. Eur J Paediatr Neurol 1991; 7: 53–66. Strother D, Ashley D, Kellie SJ, et al. Feasibility of four consecutive high-dose chemotherapy cycles with stem-cell rescue for patients with newly diagnosed medulloblastoma or supratentorial PNET after craniospinal rt: results of a collaborative study. J Clin Oncol 2001; 19: 2696–704. Merchant TE, Zhu Y, Thompson SJ, et al. Preliminary results from a phase II trial of conformal radiation therapy for paediatric patients with localized low-grade astrocytoma and ependymoma. Int J Radiat Oncol Biol Phys 2002; 52: 325–32. Duffner PK, Horowitz ME, Krischer JP, et al Postoperative chemotherapy and delayed radiation therapy and children less than three years of age with malignant brain tumours. N Engl J Med 1993; 328: 1725–31. Palmer SL, Gajjar A, Reddick WE, et al. Predicting intellectual outcome among children treated with 35–40 Gy craniospinal irradiation for medulloblastoma. Neuropsychology 2003; 17: 548–55. Palmer SL, Goloubeva O, Reddick WE, et al. Patterns of intellectual development among survivors of paediatric medulloblastoma: a longitudinal analysis. J Clin Oncol 2001; 19: 2302–08. Mulhern RK, Palmer SL, Reddick WE, et al. Risks of young age for selected neurocognitive deficits in medulloblastoma are associated with white matter loss. J Clin Oncol 2001; 19: 472–79. Ris MD, Packer R, Goldwein J, et al. Intellectual outcome after reduced-dose radiation therapy plus adjuvant chemotherapy for medulloblastoma: a Children’s Cancer Group study. J Clin Oncol 2001; 19: 3470–76. Mulhern RK, Reddick WE, Palmer SL, et al. Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol 1999; 46: 834–41. Mulhern RK, Kepner JL, Thomas PR, et al. Neuropsychologic functioning of survivors of childhood medulloblastoma randomized to receive conventional or reduced-dose craniospinal irradiation: a Paediatric Oncology Group study. J Clin Oncol 1998; 16: 1723–28. Dennis M, Spiegler BJ, Hetherington CR, Greenberg ML. Neuropsychological sequelae of the treatment of children with medulloblastoma. J Neurooncol 1996; 29: 91–101. Grill J, Renaux VK, Bulteau C, et al. Long-term intellectual outcome in children with posterior fossa tumours according to radiation doses and volumes. Int J Radiat Biol Phys 1999; 45: 137–45. Hoppe-Hirsch E, Brunet L, Laroussinie F, et al. Intellectual outcome in children with malignant tumours of the posterior fossa: influence of the field of irradiation and quality of surgery. Childs Nerv Syst 1995; 11: 340–46. Copeland DR, DeMoor C, Moore BD, Ater JL. Neurocognitive development of children after a cerebellar tumour in infancy: a longitudinal study. J Clin Oncol 1999; 17: 3476–86. Walter AW, Mulhern RK, Gajjar A, et al. Survival and neurodevelopmental outcome of young children with medulloblastoma at St Jude Children’s Research Hospital. J Clin Oncol 1999; 17: 3720–28. Kiltie AE, Lashford LS, Gattamaneni HR. Survival and late effects in medulloblastoma patients treated with craniospinal irradiation under three years old. Med Pediatr Oncol 1997; 28: 348–54. Chapman CA, Waber DP, Gernstein JH, et al. Neurobehavioral and neurologic outcome in long-term survivors of posterior fossa brain tumours: role of age and perioperative factors. J Child Neurol 1995; 10: 209–12. Fry AS, Hale S. Processing speed, working memory, and fluid intelligence: evidence for a developmental cascade. Psychol Sci 1996; 4: 237–41. Hopewell JW, van der Kogel AJ. Pathophysiological mechanisms leading to the development of late radiation-induced damage to the central nervous system. Front Radiat Ther Oncol 1999; 33: 265–75. Tofilon PJ, Fike J. The radioresponse of the central nervous system: a dynamic process. Radiat Res 2000; 153: 357–70.
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
407
Review 27 Monje ML, Mizumatsu S, Fike JR, Palmer TD. Irradiation induces neural precursor-cell dysfunction. Nat Med 2002; 8: 955–62. 28 Reddick WE, Mulhern RK, Elkin TD, et al. A hybrid neural network analysis of subtle brain volume differences in children surviving brain tumours. Magn Reson Imaging 1998; 16: 413–21. 29 Reddick WE, Russell JM, Glass JO, et al. Subtle white matter volume differences in children treated for medulloblastoma with conventional or reduced-dose cranial-spinal irradiation. Magn Reson Imaging 2000; 18: 787–93. 30 Palmer SL, Reddick WE, Glass JO, et al. Decline in corpus callosum volume among paediatric patients with medulloblastoma: longitudinal MR imaging study. Am J Neuro Radiol 2002; 23: 1088–94. 31 Reddick WE, White H, Glass JO, et al. Developmental model relating white matter volume with neurocognitive deficits in paediatric brain tumour survivors. Cancer 2003; 97: 2512–19. 32 Mulhern RK, White HA, Glass JO, et al. Attentional functioning and white matter integrity among survivors of malignant brain tumours of childhood. J Int Neuropsychol Soc 2004; 10: 180–89. 33 Khong PL, Kwong DLW, Chan GCF, et al. Diffusion-tensor imaging for the detection and quantification of treatment-induced white matter injury in children with medulloblastoma: a pilot study. Am J Neuro Radiol 2003; 24: 734–40. 34 Ogg R, Zou P, White H, et al. Attention deficits in survivors of childhood cancer: an fMRI study. J Int Neuropsychol Soc 2002; 8: 494. 35 Carrie C, Muracciole X, Gomez F, et al. Conformal radiotherapy, reduced boost volume, hyperfractionnated radiotherapy and on-line quality control in standard risk medulloblastoma without chemotherapy: results of the French M-SFOP 98 protocol. Int J Radiat Oncol Biol Phys 2003; 57: S195. 36 Thomas PR, Deutsch M, Kepner JL, et al. Low-stage medulloblastoma: final analysis of trial comparing standard-dose with reduced-dose neuraxis irradiation. J Clin Oncol 2000; 18: 3004–11. 37 Packer RJ, Goldwein J, Nicholson HS, et al. Treatment of children with medulloblastomas with reduced-dose craniospinal radiation therapy and adjuvant chemotherapy: a Children’s Cancer Group Study. J Clin Oncol 1999; 17: 2127–36. 38 Wolden SL, Dunkel IJ, Souweidane MM, et al. Patterns of failure using a conformal radiation therapy tumour bed boost for medulloblastoma. J Clin Oncol 2003; 21: 3079–83. 39 Merchant TE, Kun LE, Krasin MJ, et al. A multi-institution prospective trial of reduced-dose craniospinal irradiation (23·4 Gy) followed by conformal posterior fossa (36 Gy) and primary site
408
Neurocognition and brain tumours
40 41 42
43 44
45 46 47 48 49
50
51 52
irradiation (55·8 Gy) and dose-intensive chemotherapy for average-risk medulloblastoma. Int J Radiat Oncol Biol Phys 2003; 57: S194–95. Merchant TE, Goloubeva O, Pritchard DL, et al. Radiation dosevolume effects on growth hormone secretion. Int J Radiat Oncol Biol Phys 2002; 52: 1264–70. Kerns KA, Thomson J. Case study: implementation of a compensatory memory system in a school age child with severe memory impairment. Pediatr Rehabil 1998; 2: 77–87. Butler RW, Copeland DR. Attentional processes and their remediation in children treated for cancer: a literature review and the development of a therapeutic approach. J Int Soc Neuropsychol 2002; 8: 115–24. Sohlberg MM, Mateer CA. Cognitive rehabilitation: an integrative neuropsychological approach. Guilford Press: New York; 2001. Brown RT, Dingle, A, Dreelin B. Neuropsychological effects of stimulant medication on children’s learning and behavior. In: Reynolds CR, Fletcher-Janzen E, eds. Handbook of clinical child pharmacology. Wiley: New York; 1997. Weber P, Lutschg J. Methylphenidate treatment. Pediatr Neurol 2002; 26: 261–66. DeLong R, Friedman H, Friedman N, et al. Methylphenidate in neuropsychological sequelae of rt and chemotherapy of childhood brain tumours and leukemia. J Child Neurol 1992; 7: 462–63. Torres C, Korones D, Palumbo D, et al. Effect of methylphenidate in the post-radiation attention and memory deficits in children. Ann Neurol 1996; 40: 331–32 (abstr). Meyers CA, Weitzner MA, Valentine AD, et al. Methylphenidate therapy improves cognition, mood, and function of brain tumour patients. J Clin Oncol 1998; 16: 2522–27. Thompson SJ, Leigh L, Christensen R, et al. Immediate neurocognitive effects of methylphenidate on learningimpaired survivors of childhood cancer. J Clin Oncol 2001; 19: 1802–08. Leigh L, Miles MA. Educational issues for children with cancer. In: Pizzo PA, Poplack DG, eds. Principles and practice of paediatric oncology, 4th edn. Lippincott, Williams and Wilkins: Philadelphia; 2002. Armstrong FD, Blumberg MJ, Toledano SR. Neurobehavioral issues in childhood cancer. School Psychol Rev 1997; 3: 617–30. Max JE, Roberts MA, Koele SL, et al. Cognitive outcome in children and adolescents following severe traumatic brain injury: influence of psychosocial, psychiatric and injury-related variables. J Int Neuropsychol Soc 1999; 5: 58–68.
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
As survival among children treated for cancer continues to improve, more attention is being focussed on the late effects of cancer treatment. In children treated for brain tumours, chronic neurocognitive effects are especially challenging. Deficits in cognitive development have been described most thoroughly among children treated for posterior-fossa tumours, specifically medulloblastomas and ependymomas, which account for about 30% of all newly diagnosed cases of brain tumours in children. Most children who have survived brain tumours have required surgical resection and focal or craniospinal radiotherapy (irradiation of the entire subarachnoid volume of the brain and spine), with or without systemic chemotherapy. Historically, intelligence quotient (IQ) scores have provided a benchmark against which to measure changes in cognitive development after treatment. Observed declines in IQ are most likely a result of failure to learn at a rate that is appropriate for the age of the child, rather than from a loss of previously acquired knowledge. The rate of IQ decline is associated with a several risk factors, including younger age at time of treatment, longer time since treatment, female sex, as well as clinical variables such as hydrocephalus, use of radiotherapy and radiotherapy dose, and the volume of the brain that received treatment. Loss of cerebral white matter and failure to develop white matter at a rate appropriate to the developmental stage of the child could partly account for changes in IQ score. Technical advances in radiotherapy hold promise for lowering the frequency of neurocognitive sequelae. Further efforts to limit neurocognitive sequelae have included design of clinical trials to test the effectiveness of cognitive, behavioural, and pharmacological interventions. Lancet Oncol 2004; 5: 399–408
In 1969, HJG Bloom was one of the first physicians to recognise the threat that brain tumours and their treatment present to the quality of life of children through a study of 82 patients with medulloblastoma (figure 1) treated at the Royal Marsden Hospital, London, UK.1 These early observations showed the high risk of dementia associated with use of cranial or craniospinal radiotherapy for children younger than 2 years. Since then, studies that have assessed various components of quality of life (eg, sensory, motor, social, emotional, and cognitive components) have become increasingly sophisticated, and many compare patients with healthy children in the general population or with healthy controls matched for age and sex. These studies have
Oncology Vol 5 July 2004
Figure 1. Sagittal T1-weighted MRI showing planned cranial radiation dosimetry for contemporary treatment of medulloblastoma. Note the postsurgical tumour void anterior to the cerebellum and posterior to the brainstem. Dosimetry: magenta, 55·8 Gy; light blue, 36·0 Gy; and green, 36·0 Gy. Boundaries are progressively lower doses of radiotherapy. Conformal planning of dosimetry could help reduce neurocognitive sequelae by limiting irradiation of healthy structures in the brain.
confirmed that children treated for brain tumours are at increased risk of a poor quality of life. For example, a large cross-sectional survey2 that included 342 survivors of brain tumours in childhood and 479 sibling controls showed that former patients were 10·8 times less likely to be employed and 28·8 times less likely to be able to drive a car than their siblings. In terms of mental function, survivors were 16·2 times more likely to be incompetent; a greater likelihood of incompetence was found among survivors who were younger at time of treatment, who received cranial or craniospinal radiotherapy, or who had supratentorial tumours. RKM is Chief of the Division of Behavioral Medicine; TEM is Chief of the Division of Radiation Oncology; AG is Director of Neurooncology in the Department of Hematology/Oncology; WER is Director of Signal Processing, Division of Diagnostic Imaging; and LEK is Chair, Department of Radiological Sciences. All authors are at St Jude Children’s Research Hospital, Memphis, TN, USA. Correspondence: Dr Raymond Mulhern, Division of Behavioral Medicine, St Jude Children’s Research Hospital, 332 North Lauderdale, Memphis, TN 38105-2794, USA. Tel: +1 901 495 3580. Fax: +1 901 495 3121. Email: [email protected]
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
399
Review
Neurocognition and brain tumours
Here, we review the neurocognitive sequelae of posterior-fossa brain tumours of childhood and their treatment. Discussion is restricted to late effects on neurocognition, or cognitive problems that arise 2 or more
years after completion of treatment, as opposed to acute or subacute complications of treatment. Emphasis on late effects encompasses a range of neurocognitive problems that are therefore expected to be chronic, if not progressive, in
Table 1. Studies of intellectual development in survivors of medulloblastoma and ependymoma Sample
n
Study design
Treatment
Results
Ref
Medulloblastoma
50
Longitudinal
Radiotherapy (n=50) CSI: 35–40 Gy PF: 51–59 Gy
Mean loss of 2·2 IQ points per year after radiotherapy Greater IQ deficits with younger age at radiotherapy Greater IQ deficits among female patients
11
Medulloblastoma
44
Longitudinal
Radiotherapy (n=44) CSI: 23·4–48·0 Gy PF: 49–55 Gy Chemotherapy (n=23)
Mean loss of 2·6 IQ points per year after radiotherapy Rate of learning 50–60% of normal Greater IQ deficits with younger age at radiotherapy
12
Medulloblastoma
42
Cross-sectional
Radiotherapy (n=42) CSI: 49–54 Gy PF: 23·4–36·0 Gy Chemotherapy (n=29)
Greater IQ deficits with younger age at radiotherapy Greater IQ deficits with increasing time from radiotherapy
13
Medulloblastoma
43
Longitudinal
Radiotherapy (n=43) CSI: 23·4 Gy PF: 32·4 Gy
Mean IQ loss of 4·3 points per year after radiotherapy Greater IQ deficits with younger age at radiotherapy
14
Medulloblastoma
18
Cross-sectional
Radiotherapy (n=18) CSI: 23·4–36·0 Gy PF: 49–54 Gy Chemotherapy (n=9)
Mean IQ lower in patients with medulloblastoma than age-matched patients with low-grade astrocytoma
15
Low-grade astrocytoma Medulloblastoma
18 22
Cross-sectional
Medulloblastoma
25
Cross-sectional
Medulloblastoma
19
Cross-sectional
Ependymoma
12
Medulloblastoma
59
Ependymoma
37
Very young (<4 years) PF
27
Longitudinal
Very young (<4 years) medulloblastoma
19
Longitudinal
Very young (<4 years) medulloblastoma
37
Cross-sectional
Very young (<4 years) medulloblastoma
10
Longitudinal
Surgery only Radiotherapy (n=22) PF: 54 Gy (n=22) CSI: 36·0 Gy (n=13) CSI: 23·4 Gy (n=9) Radiotherapy (n=25) (dose not specified) Chemotherapy (n=7) Radiotherapy (n=19) CSI: 25–35 Gy (n=19) PF: 55 Gy (n=19) Chemotherapy (n=17)
Greater IQ deficits with younger age at radiotherapy
17
Greater IQ deficits with increased time from radiotherapy IQ score lower in patients with medulloblastoma than in patients with ependymoma Patients with medulloblastoma given 25 Gy (CSI) had better IQ scores than patients given 35 Gy (CSI)
Radiotherapy (n=12) Chemotherapy (n=4) PF: 25 Gy (n=11) Cross-sectional
15 16
Greater IQ deficits with younger age at radiotherapy Greater IQ deficits in patients given 36 Gy (CSI)
18
18
Radiotherapy (CSI, PR; n=59)
10% of patients had IQ of greater than 90
19
Radiotherapy (PF, n=28)
60% IQ >90 in ependymoma
19
Radiotherapy (n=7) CSI: 30–40 Gy PF: 40–50 Gy Chemotherapy (n=20) Radiotherapy (n=19) CSI: 35·2 Gy PF: 53·4 Gy Chemotherapy (n=19) Radiotherapy (n=37) CSI, PF doses varied Chemotherapy (n=15) Radiotherapy (n=10) CSI: 18 Gy PF: 54 Gy Chemotherapy (n=10)
Greater decline in IQ with radiotherapy
20
Mean IQ loss of 3·9 points per year after radiotherapy
21
8 of 16 long-term survivors in special schools
22
No change in baseline IQ score at 3 years after radiotherapy
23
CSI, craniospinal irradiation; PF, posterior fossa.
400
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Review
Neurocognition and brain tumours
long-term survivors. Other types of late effects (such as endocrinopathies) are reviewed elsewhere.3 First, we briefly review the epidemiology and contemporary approaches for the treatment (including conformal radiotherapy) of two histologically distinct tumours of the posterior fossa—ependymoma and medulloblastoma. The core and secondary neurocognitive symptoms associated with treatment of these tumours are presented along with known or suspected factors that place subsets of children at increased risk of cognitive deficits, such as area of radiotherapy dose and the dose given to some brain structures. Figure 1 shows the planning of radiotherapy for a postoperative patient with medulloblastoma. We also discuss changes that occur in the developing brain that can be detected by neuroimaging and that could explain the process of neurocognitive decline. Finally, we describe three methods of treatment that try to avoid or limit the augmentation of potentially unavoidable neurocognitive deficits: threedimensional conformal radiotherapy, cognitive or behavioural remediation, and pharmacotherapy.
Epidemiology and contemporary treatment There are two recognised peaks in the incidence of brain tumours in children and adolescents (<18 years). The first peak, with a value of 2·2–2·5 cases per 100 000 children per year, occurs the first decade of life; there is a slight male predominance at this age. This first incidence peak accounts for most cases of CNS embryonal tumours and ependymomas located in the posterior fossa. A second, much larger peak occurs in late adolescence and early adulthood, and characteristically accounts for most cases of glial tumours that occur mainly in the supratentorial compartment.4–6 A multidisciplinary team that involves specialists in neurosurgery, diagnostic imaging, radiation oncology, neuro-oncology, rehabilitation medicine, and behavioural medicine is the cornerstone for delivery of contemporary therapy at most academic neuro-oncology centres. Therapy for children diagnosed with medulloblastoma has been refined over the past 30 years. Early studies emphasised strategies to increase the rate of cure, whereas present therapeutic protocols in North America emphasise a riskbased approach to lower neurotoxicity. In patients staged with clinically standard-risk disease (defined as little or no gross evidence of tumour as shown by postoperative MRI and no evidence of metastatic disease), emphasis on a decrease in the dose of radiation to the neuraxis—including full-cranial radiation—has been a central strategy to lower the frequency of long-term neurocognitive deficits observed in earlier studies. In patients at high risk (defined as patients with >1·5 cm2 residual disease after surgery, or presence of metastatic disease), high-dose radiotherapy to the neuraxis is still the standard of care.7 Several strategies have been used to improve outcome for high-risk patients, but early results from a study that used high-dose adjuvant chemotherapy with stem-cell support seem to be most promising.8 At present, although new therapeutic approaches are being developed for high-risk patients, many could experience the more serious consequences of use of high-dose radiotherapy to the whole brain.
Oncology Vol 5 July 2004
Complete surgical resection followed by radiotherapy is the standard of care for intracranial ependymoma. Advanced neurosurgical techniques have facilitated the procedure of gross total resection of this tumour, with acceptable morbidity. New techniques of conformal radiation have enabled precise delivery of radiotherapy with less damage to the surrounding brain tissue. Early results from a study that used three-dimensional conformal radiotherapy are encouraging, especially because preliminary neurocognitive data show no adverse effect of use of this technique.9 Longterm neurocognitive follow-up is needed to validate the use of conformal radiotherapy as the standard of care for patients with intracranial ependymoma. Treatment approaches for infants diagnosed with medulloblastoma and ependymoma have previously been based on the delaying of radiotherapy to the infant brain to preserve neurocognitive function. Several studies have used an extended chemotherapy regimen after surgical resection to maintain disease control.10 For infants with localised disease, the emphasis on contemporary protocols has now shifted to use of early focal radiotherapy in an attempt to prevent local tumour recurrence. Early results from studies that have used this treatment strategy seem promising. For infants with disseminated disease, outcome remains poor because use of radiotherapy to the neuraxis is avoided.
Late effects of tumour and treatment The most common method used to document neurocognitive effects in children treated for brain tumours has been IQ tests with a normative mean score of 100 and standard deviations of 15 or 16. IQ tests are standardised on large numbers of the general population, so scores can be corrected for the age of the person who is to complete the test. The IQ score of a healthy person is thus not expected to change substantially over time. Table 1 summarises findings from 12 studies of IQ outcomes in children treated for medulloblastoma or ependymoma.11–22 Despite substantial variability in the methods used in the studies, several conclusions can be made. First, longitudinal studies consistently show significant declines in IQ over time in patients treated given craniospinal radiotherapy for medulloblastoma. Second, cross-sectional comparative studies have shown less severe deficits in IQ in those for ependymoma by use of postoperative irradiation confined to the posterior fossa alone, or with patients treated for lowgrade astrocytoma by use of surgery alone, than in those treated for medulloblastoma with craniospinal radiotherapy. Third, increasing age at time of treatment increases the risk of a decline in IQ. Studies that focused on very young patients (typically patients younger than 3 or 4 years old at time of diagnosis) showed that changes in IQ can have devastating effects; one series21 on salvage irradiation showed a median IQ score of 62 at 5 years after treatment. Although a decline in IQ has mainly been attributed to the effects of radiotherapy, other factors—such as hydrocephalus and posterior-fossa syndrome—are also known to have affect cognitive development.23 Furthermore, very young children have lower cognitive scores before radiotherapy treatment than their older counterparts,
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
401
Review
Neurocognition and brain tumours
Table 2. Main cognitive deficits among survivors of medulloblastoma Sample
n
Treatment
Main deficits
Findings
Ref
Medulloblastoma
42
Radiotherapy (n=42) CSI: 49–54 Gy PF: 23·4–36·0 Gy Chemotherapy (n=29)
Verbal memory Visual attention
Greater deficits with increasing time from radiotherapy Greater deficits with younger age at radiotherapy
13
Medulloblastoma
22
Radiotherapy (n=22) PF: 54 Gy (n=22) CSI: 36·0 Gy (n=13) CSI: 23·4 Gy (n=9)
Attention
Greater deficits in younger patients (<8 years at diagnosis) Greater deficits in patients given 36 Gy (CSI)
16
Radiotherapy (n=27) CSI: 30–40 Gy PF: 40–50 Gy Chemotherapy (n=20)
Attention
Overall decline greater in patients given radiotherapy than patients given no radiotherapy Overall decline Overall decline greater in patients given radiotherapy than patients given no radiotherapy
20
Very young (<4 years) 27 PF
Verbal memory Spatial memory
CSI, craniospinal irradiation; PF, posterior fossa.
implying that very young children are more vulnerable to the effects of the tumour, surgery, and perioperative factors.21 Another important point is that, because of assumptions of the normal distribution of IQ in the general population, 50% of patients would be predicted to have an IQ score of less than 100. A decline in IQ has frequently been associated with a commensurate decline in areas of basic academic achievement,14–16,20 but these neurocognitive outcomes are not functionally specific enough to help elucidate the neural networks that may be involved, or to help the process of rehabilitation. Increasingly, knowledge-based measures of IQ and academic achievement are viewed as the distant indicators of core cognitive functions such as attention, speed of processing, and working memory that provide the foundations for the ability to learn efficiently and retain information.13,16,20 This view is reinforced by evidence that normal, age-related improvements in processing speed and working memory account for nearly half of the age-related improvements that occur in intelligence.24 Table 2 summarises studies that have emphasised assessment of these core deficits in patients who survived medulloblastoma.13,16,20 In general, these studies confirm that deficits in the core areas of attention and memory are common,
Figure 2. External view of segmented cerebral cortex (left). Interior view of healthy white matter (green) segmented from the cerebral hemispheres with lateral ventricles shown in the segmented white matter to improve orientation (right).
402
and—similarly to IQ—that the more severe core deficits are associated with radiation treatment at a younger age, increased time from radiotherapy, and higher doses of radiotherapy. However, whether these core deficits truly cause declines in IQ and academic achievement has not yet been shown in children who survive brain tumours.
Neurodevelopmental consequences The pathophysiology of late CNS damage induced by radiation is not fully understood, especially with regard to the vulnerability of white matter to injury. Some hypotheses attribute primary mechanisms of CNS damage to the death of neuronal cells, oligodendrocytes, or endothelial cells and the subsequent microvascular damage. Ultimately, this damage seems to be a primary or secondary effect associated with administration of treatment to the CNS.25 Secondary processes, such as damage to the myelin membrane as a result of oxidative stress after radiotherapy, have also been proposed as putative mechanisms of CNS damage.26 Furthermore, study of the adverse effects of irradiation on the microenvironment of neural precursor cells has helped to integrate the findings from these competing hypotheses.27 At present, research on imaging techniques has progressed beyond diagnosis and surveillance to begin an assessment of neuropathological and neurodevelopmental correlates. One mechanism thought to account partly for the observed rate of IQ decline in survivors is loss of cerebral white matter or failure to develop white matter at a rate appropriate for the developmental stage of the survivor. This mechanism has begun to receive more attention because of newly developed methods for quantification of volumes of brain tissue (figure 2).28 Studies that have quantified toxic effects on white matter and investigated the association between neurotoxicity and cognitive deficits in children have mainly focused on survivors of medulloblastoma. For example, in a comparison of patients treated for medulloblastoma with age-matched controls who had received surgery alone for treatment of lowgrade tumours of the posterior fossa, survivors of medulloblastoma had a significantly smaller volume of cerebral white matter, a substantially greater volume of
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Neurocognition and brain tumours
Review
Volume (%)
attention, memory, and academic achievement. The researchers produced a developmental model in which academic achievement could be predicted by the volume of cerebral 54·0 54·0 54·0 white matter, ability to show attention, 36·0 18·0 23·4 and IQ score. These factors explained about 60% of the variance observed in tasks that measured reading and 100 spelling, and almost 80% of the 90 variance in mathematical tasks. The main cognitive change associated with 80 lower volume of cerebral white matter 70 was a decreased ability to pay attention, which was also associated 60 with a decrease in IQ score and 50 academic achievement. The relation between attention and cerebral white 40 matter in the same group of patients 30 was investigated in more detail by use of a self-administered computer test of 20 Total Brain DVH visual attention. After controlling 10 statistically for the effects of age at diagnosis and time elapsed from 0 treatment, the study showed a 0 10 20 30 40 50 60 70 significant association between functDose (Gy) ions that require attention and Figure 3. Benefits of dose decreases in planning of craniospinal radiotherapy shown with total-brain volumes of cerebral white matter, or dose-volume histograms (DVH), comparison of 36·0 Gy (yellow), 23·4 Gy (red), and 18 Gy (blue). specific regional volumes of white matter in the prefrontal, or frontal, cerebrospinal fluid, and an equal volume of grey matter.28 As lobe and cingulate gyrus.32 Although reliable, associations expected, survivors of medulloblastoma had significantly between decreased white matter and decline in cognitive lower IQ scores, which had a positive and statistically function could represent the influence of a third, as yet significant association with volumes of cerebral white unidentified variable that has an adverse effect on both the matter.15 However, because of the cross-sectional design of development of white matter and cognition. the studies, it has not been possible to discern whether the The next logical step in this line of investigation would smaller volume of cerebral white matter is due to loss of be to integrate maps on digital radiation dosimetry with the tissue, failure to develop white matter at an appropriate rate, data on tissue volume to assess the relation between the or both. volume of white matter and dose response. A crucial dose Subsequently, a longitudinal study showed a significant threshold could be established and regional analyses could decrease in the volume of cerebral white matter in patients be combined with more sensitive and specific neurowho underwent treatment for medulloblastoma; this cognitive testing. These studies are necessary to establish the decrease was more rapid in patients who received 36 Gy precise relation between therapy, the location of the craniospinal irradiation than in patients who received developmental area causing cognitive deficit, the location of 23·4 Gy craniospinal irradiation.29 These findings that the decrease in the volume of cerebral white matter, and the associate cerebral white matter with irradiation dose were effect on neurocognitive performance. confirmed in another longitudinal study on volumes of Global or regional volumes of neuronal structures might corpus callosum, which found the greatest deviation from be relatively insensitive to more subtle changes that occur in normal development occurred in the most posterior the microstructure and organisation of fibre tracts in white subregions of the brain—the area that also received the matter. Other updated measures of the integrity of white highest total dose of irradiation.30 matter should also be incorporated, such as diffusion tensor Previous studies have established a relation between the imaging of apparent diffusion coefficient and fractional volume of cerebral white matter and both radiation dose and anisotropy. The apparent diffusion coefficient measures the IQ. A further study established that the volume of cerebral average diffusion distance in a region, which reflects the white matter could explain about 70% of the association volume ratio of extracellular to intracellular white matter between IQ score and age at time of irradiation.13 A cross- and is also a sensitive measure of acute ischaemia. Fractional sectional study, by Reddick and colleagues,31 of patients anisotropy measures the directional organisation of a region treated for medulloblastoma showed significantly impaired and thus reflects myelin integrity. A pilot study of children performance on all neurocognitive measures of intellect, who survived medulloblastoma found significant decreases
Oncology Vol 5 July 2004
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
403
Review
Neurocognition and brain tumours
fractionation; chemotherapy was not included in many cases. To address mainly the neuropsychological sequelae commonly seen in long-term survivors, patients with average-risk medulloblastoma were randomly assigned a lower dose of craniospinal radiotherapy (up to 23·4 Gy or 36·0 Gy).36 In a separate study, results with a lower dose of craniospinal irradiation combined with multiagent chemotherapy compared favourably with those achieved by standard-dose craniospinal irraditation. Combination treatment has since been used as standard treatment for average-risk patients, on the premise that a lower dose of craniospinal irradiation would have fewer long-term Innovations in radiotherapy effects.37 Despite decreases in the dose of craniospinal The side-effects attributed to radiotherapy have been a main radiotherapy, cognitive impairment remains significant even concern in the design of clinical trials for medulloblastoma with a decrease in IQ of 4·3 points per year reported from a during the past 20 years. Concern has been greatest for national trial of low-dose craniospinal radiotherapy and patients with average-risk medulloblastoma, for whom long- chemotherapy.14 Although a decreased dose of craniospinal radiotherapy term control of disease is likely and the side-effects of therapy might be long-lasting. Radiotherapy for can help to reduce the dose given to the entire brain and medulloblastoma is technically demanding, requiring both temporal lobes, it does little to decrease the volume of tissue craniospinal radiotherapy and boost treatment of the that receives the highest doses because the boost dose is anatomic posterior fossa. If the prescribed dose of radiation prescribed to the entire posterior fossa (figure 3). For patients can be delivered to these well-defined volumes, there is an at average risk, lowering of the boost volume from the entire posterior fossa to the tumour bed with a limited margin can increased likelihood of obtaining disease control.35 Only 10 years ago, the standard of care for patients with significantly decrease the volume of healthy tissue that average-risk medulloblastoma was postoperative cranio- receives the highest doses (figure 4). Results obtained spinal radiotherapy up to a dose of 36 Gy and boost with smaller target volumes and sophisticated, image-guided treatment up to 54 Gy by use of 1·8 Gy per day conventional radiation techniques indicate excellent disease control; the intended improvement in functional measures—including neurocognitive status—awaits further confirmation.38 23·4 23·4 A larger study on decreased irradiation of target volume that used focal, image-guided delivery of 54·0 54·0 radiation has been reported in preliminary form. Among 84 children, the cumulative rate of failure of the posterior fossa at 3 years from 100 diagnosis was 6·3%.39 The treatment 90 regimen for patients enrolled on the trial (SJMB-96) included postoperative 80 Total Brain DVH craniospinal radiotherapy up to 70 23·4 Gy, conformal posterior fossa boost up to 36·0 Gy, and focal 60 treatment of the tumour bed with a 50 2-cm margin. High-dose chemotherapy with peripheral stem-cell 40 support was administered after 30 radiotherapy. Lowering of the target volume to more closely encompass the 20 tumour bed can further spare the 10 temporal lobes from the effects of irradiation—a relation believed to be 0 important in preservation of 0 10 20 30 40 50 60 70 neurocognitive function. The current Dose (Gy) trial at St Jude Children’s Research Hospital, TN, USA (SJMB-03) is using Figure 4. Benefits of dose decreases in planning of radiotherapy to posterior fossa shown with totalfocal treatment of the primary site of brain dose-volume histograms (DVH), comparison of conventional boost (blue) to posterior fossa with conformal boost (yellow) to the primary site after 23·4 Gy craniospinal irradiation. the tumour after craniospinal Volume (%)
in fractional anisotropy of white matter, which correlated significantly with deterioration of performance at school.33 Functional MRI, a method of mapping brain activation, has been used extensively to investigate and characterise the basic neural networks that support normal cognitive function and disease-associated changes in the networks. MRI might also provide important insights into the relation between morphological abnormalities induced by disease and treatment, as well as the behavioural deficits that negatively affect the quality of life of survivors of medulloblastoma.34
404
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Review
Neurocognition and brain tumours
radiotherapy with a limited margin of 1 cm. About 25% of the posterior fossa will be spared from receiving the prescribed boost dose of 55·8 Gy. In the Children’s Oncology Group ACNS0331—an ambitious four-group randomised trial for children aged 3–8 years—participants with medulloblastoma are assigned 18·0 Gy or 23·4 Gy craniospinal radiotherapy, followed by a second randomisation to either conventional boost treatment or treatment of less than the entire posterior fossa. For patients older than 8 years, the craniospinal dose will remain at 23·4 Gy. The risks associated with a lowering of the dose of craniospinal radiotherapy and limitation of the boost treatment to the postoperative tumour bed with a limited margin must be balanced against the observed effects of radiotherapy on cognition, neurological function, growth, and development. Dose-volume data can now be acquired for functional regions of the brain and correlated with outcome for the patient.40 Establishment of dose-volume relations can be used to predict side-effects of treatment and could thus help tailor treatment. The usefulness of these models will be increased if factors other than radiation dose and age are tested for inclusion: tumour-related insults, effects of surgery, hydrocephalus, combination chemotherapy, and other factors should be investigated so that accurate models of the effects of treatment can be developed. Only then will we be able to investigate whether the planned deceases in dose for treatment in clinical trials are likely to be beneficial. Investigators in Europe and the UK are testing different strategies to decrease treatment-related toxic effects by use of hyperfractionated radiotherapy for craniospinal radiotherapy and the boost phase.35 The underlying idea is that smaller doses administered twice a day will decrease the occurrence of long-term effects. Hyperfractionation seems to be equally effective when given once a day, provided the total dose is increased slightly. Comparison of disease control and functional outcomes between the strategies of the American, European, and UK cooperative groups might some day provide the basis for a worldwide trial for patients with medulloblastoma.
Cognitive remediation Few published studies have assessed cognitive behavioural interventions for survivors of brain tumours in childhood. One case study investigated whether a compensatory memory notebook would improve neurocognitive function in an adolescent with severe memory impairment.41 A calendar section was used to record forthcoming activities and events, and to keep track of time. The notebook had a section on things to do for forthcoming assignments; an orientation section contained the names of teachers and the counsellor, classroom numbers, and other important personal information. Finally, a transport section included a bus schedule and maps of the school. The patient was systematically trained to use the memory book. On testing before and after the intervention, the patient showed signs of significantly impaired memory, although academic achievement increased slightly and there was an improvement in class attendance and timely completion of assignments.
Oncology Vol 5 July 2004
The most systematic effort to apply ideas of cognitive remediation to children with neurocognitive deficits as a result of cancer and subsequent treatment is the programme developed by Butler, Copeland, and colleagues.42 Their tripartite model uses techniques and methods from three disciplines: brain-injury rehabilitation, special education (or educational psychology), and clinical psychology. The programme consists of 20 sessions, each of 2 h, with an individual therapist. One component, called attentionprocess training, addresses deficits in control of sustained, selective, divided, and executive attention.43 If a patient does not obtain a score of 50% during a task, the activity is substituted for a more basic task. Once the child reaches 80% accuracy, the next stage of difficulty for a given test is applied. Each patient also receives instruction in metacognitive strategies, which are grouped within the three areas of task preparedness, on-task performance, and posttask strategies. Other methods to study cognitive behaviour specifically assess the ability to withstand distraction. Cognitive-behavioural psychotherapy is also used, which includes reframing of cognitive struggles in a positive light, psychotherapeutic support, acknowledgment of weaknesses, and blocks to successful improvement; other components are monitoring of internal dialogue, stress management, support for the patient to become his or her own best friend, and reinforcement of realistic, positive, and optimistic selfstatements. Preliminary data on efficacy have suggested an improvement in laboratory measures of abilities in attention, but no improvement in academic achievement.42 The results of a continuing phase III clinical trial based at seven institutions in the USA should soon be ready for publication.
Pharmacotherapy Research done over the past 50 years in children diagnosed with attention deficit hyperactivity disorder (ADHD) who are otherwise healthy has shown the effectiveness of stimulant medications—most commonly methylphenidate—to improve cognitive performance.44 Methylphenidate is a mixed dopaminergic–noradrenergic agonist that is thought to improve the function of the attention network in the frontostriatal region of the brain.45 Use of methylphenidate is found to have the most consistent and significant benefits in measures of vigilance and sustained attention, but improvements in reaction time, paired-associate learning, and perceptual efficiency are also commonly seen.44 Positive effects of methylphenidate on higher-order cognitive functions, such as problem-solving or language processing, have not been observed. Studies that assess the effects of methylphenidate on actual academic achievement—as opposed to behavioural improvements in the classroom—are equivocal among children with ADHD. Improvement in academic learning attributed to the action of methylphenidate is mainly the result of the drug’s effect on attention and ability to concentrate. In children with ADHD, methylphenidate might also improve formation of social relationships with peers, as assessed by behavioural observations and peer ratings.44 There have been very few studies on the
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
405
Review effectiveness of methylphenidate for attention problems in children who experience various forms of acquired brain damage. There is evidence that methylphenidate can be as effective against attention problems in children with epilepsy as against ADHD, without a significant lowering of the seizure threshold.45 Two preliminary studies were among the first to investigate the use of methylphenidate in children with learning problems that were presumed to be secondary to cancer treatment. In the first of these studies, 12 children who survived either brain tumours or acute lymphoblastic leukaemia were treated with methylphenidate for 6 months to 6 years (median duration of treatment was 23 months). The response was rated as good in eight children, fair in two, and poor in two.46 In the second study, six children who had received craniospinal radiotherapy 3–12 years earlier for treatment of brain tumours were given methylphenidate.47 With a consistent dose of 0·3 mg/kg methylphenidate, there was no significant immediate, or delayed, benefit to patients. An open-label study on adults treated for malignant brain tumours used neurocognitive tests and objective inventories to quantify the response to methylphenidate.48 The study found significant improvements in cognitive function (including psychomotor speed, memory, and executive functions) as well as mood and activities of daily living, commonly despite the presence of progressive disease. Thompson and co-workers49 investigated the effects of methylphenidate in a randomised, double-blind trial on survivors of acute lymphoblastic leukaemia or malignant brain tumours who had impaired learning. The researchers screened for a particular cognitive phenotype that was thought to elicit an optimum response to methylphenidate treatment—deficits in academic achievement and a concurrent problem with vigilance. 32 children were randomly assigned placebo or methylphenidate (0·6 mg/kg, up to maximum dose of 20 mg). They were retested on a completion of a computerised continuous-performance test of vigilance 90 min after the treatment. The group assigned methylphenidate showed significantly greater improvement than the placebo group. The results of that 1-day study encouraged implementation of a multisite trial on the effects of longterm exposure to methylphenidate, which has similar criteria for eligibility and exclusion to the study by Thompson and co-workers.49 After an initial laboratory challenge with methylphenidate, children participate in a randomised, double-blind, 3-week home crossover trial. Participants take two capsules each day for 5 days per week: 1 week of placebo, 1 week of low-dose methylphenidate (0·3 mg/kg, up to a maximum of 10 mg), and 1 week of moderate-dose methylphenidate (0·6 mg/kg, up to a maximum of 20 mg). At the end of each week, parents and teachers complete the Conners’ rating scales of attention and behaviour, and also complete scales to assess side-effects. At the end of the third week, the masking code is broken and the ratings are compared. If a patient has shown objective improvement with methylphenidate, the medication is continued for the next 12 months, then attention function and school achievement are reassessed. At present,
406
Neurocognition and brain tumours
preliminary analyses are in progress. Of the first 325 patients screened for participation in the trial, 40% qualified for methylphenidate treatment and 70% of those agreed to participate in the trial. Most of the participants have shown at least a minimum response to treatment with methylphenidate, and the parents of most of those who have shown a response agreed to continue treatment with methylphenidate. Whether this drug ultimately facilitates academic achievement in survivors of childhood cancer who have problems with attention is not yet known.
Environmental interventions In addition to cognitive remediation and pharmacological approaches to overcome neurocognitive deficits, the importance of ecological or environmentally based interventions for children with brain injuries—especially with regard to school—should not be underestimated.50 Many such children might need extended time limits for completion of school examinations, use of true-or-false questions and multiple-choice formats in tests rather than essay-based examinations, and encouragement to record classroom lectures for later review. Other ecological interventions include use of written handouts to decrease demands for copying of material from the blackboard, and substitution of computers for handwritten assignments. Whether the main approach to treatment is cognitive remediation, pharmacotherapy, ecological intervention, or a combination, explanation of the affected areas of neurocognitive deficit to patients, providers of care, and others in the community (eg, primary health-care providers and especially teachers) is important. Leigh and colleagues50 have recommended that all paediatric oncology centres should have a structured programme of school re-entry, a component of which should be education of the patient’s teacher about childhood cancer and the specific signs, symptoms, and special needs associated with the patient’s treatment and the likely outcome. Some special arrangements are obvious, such as ensuring that a patient with hemiparesis has extra time to move between classrooms. However, some neurological and neurocognitive deficits can be quite subtle, including mild loss of vision or hearing, or slow visual-motor production. Furthermore, the neurocognitive status of the patient might not remain stable after completion of treatment. The onset of some deficits can be delayed, and other deficits might not be evident until the ability is normally expected.51 One of the greatest dangers in the lack of communication of new problems to the patient’s teachers is that struggles the child may have in the classroom could be wrongly attributed to a lack of motivation, attitude problems, daydreaming, or emotional maladjustment. Simple arrangements in the classroom, such as fewer items on multiple-choice tests, preferential seating in the classroom, and lower expectations for amount of homework are sometimes needed. Since older children generally have many teachers in a school year and that most, if not all, of teachers change with each school year, communication of older patients’ special needs can be more difficult. Many children undergoing treatment for cancer, as well as those who have completed treatment, will be classified as “other
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com
Review
Neurocognition and brain tumours
Search strategy and selection criteria Data for this review were identified by searches of PubMed with the terms “children”, “ependymoma”, “medulloblastoma”, “PNET”, and “posterior fossa”. Only papers published in English since 1990 were selected. Additional papers were identified from references of relevant articles and for historical value.
4 5 6 7
health impaired” by their local school systems to access additional resources for their special needs. One area of progress in the development of interventions for cognitive deficits in children has been the increase in appreciation of the influence of the family environment and resources available after recovery from traumatic brain injury. Chaotic and dysfunctional family environments have a substantial adverse effect on neurocognitive recovery from traumatic brain injury in school-age children. This relation seems to be valid, even when there is control for severity of brain injury and other medical factors.52 The implications of these findings for the potential success of interventions to remediate neurocognitive deficits in children who survive brain tumours has not yet been investigated.
8
9
10
11 12 13
Conclusions Neurosurgeons, paediatric oncologists, and radiation oncologists have been successful in improving cure rates for most types of childhood brain tumours, including those of the posterior fossa. Nevertheless, the risks of neurocognitive impairment remain substantial, especially among individuals who were treated aggressively and at a young age. Thus, there is an imperative need for effective, but less neurotoxic, tumour therapy. Until such therapy is developed, cognitive, behavioural, pharmacological, and environmental interventions will be needed to address the neurocognitive effects that cannot be avoided. However, these interventions are only beginning to be investigated in randomised clinical trials, and information on the efficacy of different interventions will not be available for some time. We strongly encourage the further development of innovative strategies to preserve cognitive function while cure rates for children with malignant brain tumours continue to be maintained and improved. Conflict of interest
14
15 16
17 18 19
20 21
None declared. Acknowledgments
Our research is supported by the American Lebanese Syrian Associated Charities (ALSAC) and Cancer Center Support (CORE) grant P30CA21765, R01CA78957, U01CA81445, and R01CA90246 from the US National Cancer Institute.
22 23
References
1 Bloom HJG, Wallace ENK, Henk JM. The treatment and prognosis of medulloblastoma in children. AJR Am J Roentgenol 1969; 105: 43–62. 2 Mostow EN, Byrne J, Connelly RR, Mulvihill JJ. Quality of life in long-term survivors of CNS tumours of childhood and adolescence. J Clin Oncol 1991; 9: 592–99. 3 Strother DR, Pollack IF, Fisher PG, et al. Tumours of the central nervous system. In: Pizzo PA, Poplack DG, eds. Principles and
Oncology Vol 5 July 2004
24 25 26
practice of paediatric oncology, 4th edn. Lippincott, Williams and Wilkins: Philadelphia, 2002. Gurney JG, Severson RK, Davis S, Robinson LL. Incidence of cancer in children in the United States. Cancer 1995; 75: 2186. Young J, Miller R. Incidence of malignant tumours in US children. J Pediatr 1975; 86: 254. Schoenburg B, Christine B, Wishnant J. The descriptive epidemiology of primary intracranial neoplasm of childhood: a population study. Mayo Clin Proc 1976; 51: 51. Ellison D, Clifford SC, Gajjar A, Gilbertson RJ. What’s new in neurooncology? Recent advances in medulloblastoma. Eur J Paediatr Neurol 1991; 7: 53–66. Strother D, Ashley D, Kellie SJ, et al. Feasibility of four consecutive high-dose chemotherapy cycles with stem-cell rescue for patients with newly diagnosed medulloblastoma or supratentorial PNET after craniospinal rt: results of a collaborative study. J Clin Oncol 2001; 19: 2696–704. Merchant TE, Zhu Y, Thompson SJ, et al. Preliminary results from a phase II trial of conformal radiation therapy for paediatric patients with localized low-grade astrocytoma and ependymoma. Int J Radiat Oncol Biol Phys 2002; 52: 325–32. Duffner PK, Horowitz ME, Krischer JP, et al Postoperative chemotherapy and delayed radiation therapy and children less than three years of age with malignant brain tumours. N Engl J Med 1993; 328: 1725–31. Palmer SL, Gajjar A, Reddick WE, et al. Predicting intellectual outcome among children treated with 35–40 Gy craniospinal irradiation for medulloblastoma. Neuropsychology 2003; 17: 548–55. Palmer SL, Goloubeva O, Reddick WE, et al. Patterns of intellectual development among survivors of paediatric medulloblastoma: a longitudinal analysis. J Clin Oncol 2001; 19: 2302–08. Mulhern RK, Palmer SL, Reddick WE, et al. Risks of young age for selected neurocognitive deficits in medulloblastoma are associated with white matter loss. J Clin Oncol 2001; 19: 472–79. Ris MD, Packer R, Goldwein J, et al. Intellectual outcome after reduced-dose radiation therapy plus adjuvant chemotherapy for medulloblastoma: a Children’s Cancer Group study. J Clin Oncol 2001; 19: 3470–76. Mulhern RK, Reddick WE, Palmer SL, et al. Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol 1999; 46: 834–41. Mulhern RK, Kepner JL, Thomas PR, et al. Neuropsychologic functioning of survivors of childhood medulloblastoma randomized to receive conventional or reduced-dose craniospinal irradiation: a Paediatric Oncology Group study. J Clin Oncol 1998; 16: 1723–28. Dennis M, Spiegler BJ, Hetherington CR, Greenberg ML. Neuropsychological sequelae of the treatment of children with medulloblastoma. J Neurooncol 1996; 29: 91–101. Grill J, Renaux VK, Bulteau C, et al. Long-term intellectual outcome in children with posterior fossa tumours according to radiation doses and volumes. Int J Radiat Biol Phys 1999; 45: 137–45. Hoppe-Hirsch E, Brunet L, Laroussinie F, et al. Intellectual outcome in children with malignant tumours of the posterior fossa: influence of the field of irradiation and quality of surgery. Childs Nerv Syst 1995; 11: 340–46. Copeland DR, DeMoor C, Moore BD, Ater JL. Neurocognitive development of children after a cerebellar tumour in infancy: a longitudinal study. J Clin Oncol 1999; 17: 3476–86. Walter AW, Mulhern RK, Gajjar A, et al. Survival and neurodevelopmental outcome of young children with medulloblastoma at St Jude Children’s Research Hospital. J Clin Oncol 1999; 17: 3720–28. Kiltie AE, Lashford LS, Gattamaneni HR. Survival and late effects in medulloblastoma patients treated with craniospinal irradiation under three years old. Med Pediatr Oncol 1997; 28: 348–54. Chapman CA, Waber DP, Gernstein JH, et al. Neurobehavioral and neurologic outcome in long-term survivors of posterior fossa brain tumours: role of age and perioperative factors. J Child Neurol 1995; 10: 209–12. Fry AS, Hale S. Processing speed, working memory, and fluid intelligence: evidence for a developmental cascade. Psychol Sci 1996; 4: 237–41. Hopewell JW, van der Kogel AJ. Pathophysiological mechanisms leading to the development of late radiation-induced damage to the central nervous system. Front Radiat Ther Oncol 1999; 33: 265–75. Tofilon PJ, Fike J. The radioresponse of the central nervous system: a dynamic process. Radiat Res 2000; 153: 357–70.
http://oncology.thelancet.com
For personal use. Only reproduce with permission from The Lancet.
407
Review 27 Monje ML, Mizumatsu S, Fike JR, Palmer TD. Irradiation induces neural precursor-cell dysfunction. Nat Med 2002; 8: 955–62. 28 Reddick WE, Mulhern RK, Elkin TD, et al. A hybrid neural network analysis of subtle brain volume differences in children surviving brain tumours. Magn Reson Imaging 1998; 16: 413–21. 29 Reddick WE, Russell JM, Glass JO, et al. Subtle white matter volume differences in children treated for medulloblastoma with conventional or reduced-dose cranial-spinal irradiation. Magn Reson Imaging 2000; 18: 787–93. 30 Palmer SL, Reddick WE, Glass JO, et al. Decline in corpus callosum volume among paediatric patients with medulloblastoma: longitudinal MR imaging study. Am J Neuro Radiol 2002; 23: 1088–94. 31 Reddick WE, White H, Glass JO, et al. Developmental model relating white matter volume with neurocognitive deficits in paediatric brain tumour survivors. Cancer 2003; 97: 2512–19. 32 Mulhern RK, White HA, Glass JO, et al. Attentional functioning and white matter integrity among survivors of malignant brain tumours of childhood. J Int Neuropsychol Soc 2004; 10: 180–89. 33 Khong PL, Kwong DLW, Chan GCF, et al. Diffusion-tensor imaging for the detection and quantification of treatment-induced white matter injury in children with medulloblastoma: a pilot study. Am J Neuro Radiol 2003; 24: 734–40. 34 Ogg R, Zou P, White H, et al. Attention deficits in survivors of childhood cancer: an fMRI study. J Int Neuropsychol Soc 2002; 8: 494. 35 Carrie C, Muracciole X, Gomez F, et al. Conformal radiotherapy, reduced boost volume, hyperfractionnated radiotherapy and on-line quality control in standard risk medulloblastoma without chemotherapy: results of the French M-SFOP 98 protocol. Int J Radiat Oncol Biol Phys 2003; 57: S195. 36 Thomas PR, Deutsch M, Kepner JL, et al. Low-stage medulloblastoma: final analysis of trial comparing standard-dose with reduced-dose neuraxis irradiation. J Clin Oncol 2000; 18: 3004–11. 37 Packer RJ, Goldwein J, Nicholson HS, et al. Treatment of children with medulloblastomas with reduced-dose craniospinal radiation therapy and adjuvant chemotherapy: a Children’s Cancer Group Study. J Clin Oncol 1999; 17: 2127–36. 38 Wolden SL, Dunkel IJ, Souweidane MM, et al. Patterns of failure using a conformal radiation therapy tumour bed boost for medulloblastoma. J Clin Oncol 2003; 21: 3079–83. 39 Merchant TE, Kun LE, Krasin MJ, et al. A multi-institution prospective trial of reduced-dose craniospinal irradiation (23·4 Gy) followed by conformal posterior fossa (36 Gy) and primary site
408
Neurocognition and brain tumours
40 41 42
43 44
45 46 47 48 49
50
51 52
irradiation (55·8 Gy) and dose-intensive chemotherapy for average-risk medulloblastoma. Int J Radiat Oncol Biol Phys 2003; 57: S194–95. Merchant TE, Goloubeva O, Pritchard DL, et al. Radiation dosevolume effects on growth hormone secretion. Int J Radiat Oncol Biol Phys 2002; 52: 1264–70. Kerns KA, Thomson J. Case study: implementation of a compensatory memory system in a school age child with severe memory impairment. Pediatr Rehabil 1998; 2: 77–87. Butler RW, Copeland DR. Attentional processes and their remediation in children treated for cancer: a literature review and the development of a therapeutic approach. J Int Soc Neuropsychol 2002; 8: 115–24. Sohlberg MM, Mateer CA. Cognitive rehabilitation: an integrative neuropsychological approach. Guilford Press: New York; 2001. Brown RT, Dingle, A, Dreelin B. Neuropsychological effects of stimulant medication on children’s learning and behavior. In: Reynolds CR, Fletcher-Janzen E, eds. Handbook of clinical child pharmacology. Wiley: New York; 1997. Weber P, Lutschg J. Methylphenidate treatment. Pediatr Neurol 2002; 26: 261–66. DeLong R, Friedman H, Friedman N, et al. Methylphenidate in neuropsychological sequelae of rt and chemotherapy of childhood brain tumours and leukemia. J Child Neurol 1992; 7: 462–63. Torres C, Korones D, Palumbo D, et al. Effect of methylphenidate in the post-radiation attention and memory deficits in children. Ann Neurol 1996; 40: 331–32 (abstr). Meyers CA, Weitzner MA, Valentine AD, et al. Methylphenidate therapy improves cognition, mood, and function of brain tumour patients. J Clin Oncol 1998; 16: 2522–27. Thompson SJ, Leigh L, Christensen R, et al. Immediate neurocognitive effects of methylphenidate on learningimpaired survivors of childhood cancer. J Clin Oncol 2001; 19: 1802–08. Leigh L, Miles MA. Educational issues for children with cancer. In: Pizzo PA, Poplack DG, eds. Principles and practice of paediatric oncology, 4th edn. Lippincott, Williams and Wilkins: Philadelphia; 2002. Armstrong FD, Blumberg MJ, Toledano SR. Neurobehavioral issues in childhood cancer. School Psychol Rev 1997; 3: 617–30. Max JE, Roberts MA, Koele SL, et al. Cognitive outcome in children and adolescents following severe traumatic brain injury: influence of psychosocial, psychiatric and injury-related variables. J Int Neuropsychol Soc 1999; 5: 58–68.
Oncology Vol 5 July 2004
For personal use. Only reproduce with permission from The Lancet.
http://oncology.thelancet.com