SEARCH

SEARCH BY CITATION

Keywords:

  • brain tumors;
  • magnetic resonance imaging scan;
  • image analysis;
  • neurologic complications

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND

The primary objective of this study was to test the hypothesis that, among survivors of pediatric brain tumors, the association between reduced volumes of normal-appearing white matter (NAWM) and intellectual/academic achievement deficits can be explained by patient problems with memory and attention.

METHODS

Quantitative tissue volumes from magnetic resonance imaging scans and neurocognitive assessments were obtained for 40 long-term survivors of pediatric brain tumors. They were treated with radiotherapy (RT) with or without chemotherapy 2.6–15.3 years earlier (median, 5.7 years) at an age of 1.7–14.8 years (median, 6.5 years). Neurocognitive assessments included standardized tests of intellect (intelligence quotient [IQ]), attention, memory, and academic achievement.

RESULTS

Analyses revealed significant impairments in patients' neurocognitive test performance on all measures. After statistically controlling for age at RT and time from RT, significant associations were found between NAWM volumes and both attentional abilities and IQ, and between attentional abilities and IQ. Subsequent analyses supported the hypothesis that attentional abilities, but not memory, could explain a significant amount of the relationship between NAWM and IQ. The final developmental model predicting academic achievement based on NAWM, attentional abilities, and IQ explained approximately 60% of the variance in reading and spelling and almost 80% of the variance in math.

CONCLUSIONS

The authors demonstrated that the primary consequence of reduced NAWM among pediatric patients treated for brain tumors was decreased attentional abilities, leading to declining IQ and academic achievement. Cancer 2003;10:2512–9. © 2003 American Cancer Society.

DOI 10.1002/cncr.11355

Approximately 20% of all newly diagnosed childhood cancers are brain tumors (BT).1 Regardless of tumor histology or location, survivors of pediatric BT have some intellectual deficits.2–4 Due to the inherent risk of local disease recurrence or central nervous system (CNS) dissemination, these patients may receive aggressive CNS therapy, including maximal surgical resection followed by local irradiation with or without whole brain irradiation and/or chemotherapy. Consequently, long-term survivors are at increased risk for cognitive delays or deficits, which often impair future academic performance, employment, and quality of life.2, 3, 5–17 These long-term adverse effects of treatment, specifically deficits in intelligence quotient (IQ) and academic achievement, are due to a diminished ability to acquire new information rather than the loss of previously learned information.18 This inability to acquire new information may be secondary to one or more cognitive processing impairments, including deficits in attention, short-term memory, speed of processing, visual-motor coordination, and sequencing abilities. These impairments depend on the integrity of widely distributed neural networks believed to be supported by interhemispheric and intrahemispheric white matter tracts.19–22 Radiotherapy (RT) and some chemotherapeutic agents are well-established causes of structural alteration of cerebral white matter.23 Other potential sources of white matter damage include increased intracranial pressure and treatment with steroids.8, 23 This suggests that white matter damage may represent a useful index of the cumulative impact of multiple sources of CNS insult, regardless of tumor histology or tumor location.

Preliminary quantitative studies of white matter neurotoxicity and associations with cognitive deficits in children have focused primarily on patients surviving BT. These studies demonstrated a significant reduction in the volume of normal-appearing white matter (NAWM), but not a significant reduction in gray matter or cerebrospinal fluid (CSF) volume, among patients treated for medulloblastoma when compared with age-matched controls treated with surgery alone for low-grade tumors of the posterior fossa.24 The patients treated for medulloblastoma also demonstrated significantly lower IQ scores.25 However, due to its cross-sectional design, this study could not discern whether the lower white matter volumes were due to loss of NAWM or to decreased maturation of white matter. A subsequent longitudinal study of patients undergoing treatment for medulloblastoma followed patients throughout their treatment regimens, which included surgery, RT, and chemotherapy. The results of this longitudinal study revealed a significant loss of NAWM volumes, in contrast to normally expected maturation.26 Finally, a more recent cross-sectional study of 42 patients treated for medulloblastoma found that approximately 70% of the association between age at RT and IQ was explained by NAWM volumes.27

The current study further defines the relationships between NAWM and neurocognitive functioning among survivors of pediatric BT. Extensive measures of neurocognitive functioning (IQ, attention, memory, and academic achievement) were combined with simultaneous magnetic resonance image (MRI) scan assessment of brain tissue volumetric measures.28 This study explicitly tests a developmental model to further explain neurocognitive deficits in pediatric BT survivors through assessment of NAWM volumes (Fig. 1). In this model, decreased NAWM volumes in children surviving treatment for BT are hypothesized to be associated with deficits in memory and/or attention. These deficits are reflected in lower IQ scores and below-normal academic achievement.

thumbnail image

Figure 1. Proposed developmental model of the relationship between normal-appearing white matter (NAWM), attention, memory, intelligence, and academic achievement. Multiple pathways are possible to incorporate attention and memory between NAWM and intelligence.

Download figure to PowerPoint

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patient Population

Study participants were recruited from an institutional review board–approved clinical trial of patients diagnosed with either a BT or acute lymphoblastic leukemia at a single pediatric cancer center.29 To be eligible for participation, patients were required to be age 6–18 years, to be receiving no therapy, and to have documented disease control for 2 or more years. In addition, English was required to be the primary language of all patients. Of the 108 consenting participants, 57 met the study criteria. They were diagnosed with pediatric BT and were treated with surgical resection and RT, with or without chemotherapy. Of these 57 patients, 17 were excluded because of inadequate MRI scans for the quantitative volumetric measures.

Forty participants (22 male and 18 female) were included in the final analysis. Primary tumor histology consisted of medulloblastoma (n = 18), astrocytoma (n = 8), ependymoma (n = 6), primitive neuroectodermal tumors (n = 4), germinoma (n = 2), oligodendroglioma (n = 1), and craniopharyngioma (n = 1). Tumors were located in the posterior fossa (n = 26), third ventricle (n = 6), and cerebral hemisphere (n = 8) according to a previously reported classification system for tumor locations.2, 16 In response to concerns regarding the heterogeneity of tumor location, we compared patients with posterior fossa tumors with patients with tumors in other locations using Student t tests of each volume and neurocognitive measure. There were no statistical differences in the white matter volumes, attentional and memory measures, or academic achievement measures between the two groups of patients. Eighteen patients received chemotherapy consisting of one of more of the following agents: cisplatin, carboplatin, cyclophosphamide, vincristine, and MOPP. When combined with RT, these chemotherapy regimens did not have asignificant additional impact on white matter volumes.24 Sixteen patients received RT to the primary site alone (14 conventional, 1 hyperfractionated, 1 brachytherapy) and 24 patients received local and whole brain irradiation (Table 1).

Table 1. Type and Timing of Therapy Experienced by the Patients
CharacteristicsNo. of patientsMedian (range)
Age at examination (yrs)4012.8 (7.1–18.8)
Chemotherapy18 — (—)
Age at irradiation (yrs)40 6.5 (1.7–14.8)
Time since irradiation (yrs)40 5.7 (2.6–15.3)
Whole–brain irradiation (Gy)2435.2 (23.4–44.0)
Local irradiation only (Gy)16 
 Conventional1453.1 (37.8–59.4)
 Hyperfractionated170.2 (M)
 Brachytherapy165.5 (M)

Neurocognitive Assessment

Abbreviated Wechsler Intelligence Scales

All participants were given the Information (fund of factual knowledge), Similarities (verbal reasoning), and Block Design (nonverbal reasoning) subtests from either the Wechsler Intelligence Scale for Children-III30 or the Wechsler Adult Intelligence Scale–Revised.31 The screening process was utilized to provide age-corrected estimates of full-scale IQ (FSIQ) with a mean score of 100 and (standard deviation [SD] = 15). This abbreviated version has been found to correlate highly with FSIQ32 and has been used successfully in several previous studies relating brain volume to intellectual development.25, 27

Conners' Continuous Performance Test (CPT)

This computer-administered test measures selective and sustained attention, reaction time, and impulsivity.33 The CPT provides age-corrected standard scores on multiple indices of attentional abilities and impulsivity for healthy children and adolescents, as well as those diagnosed with attention deficit hyperactivity disorder. The CPT takes approximately 15 minutes to complete and is computer scored. Patients are instructed to press the space bar on the computer keyboard when they see any letter other than “X” and to withhold responding when the letter “X” appears. Pressing the bar after any letter other than “X” is termed a “hit,” not pressing the bar after any letter other than “X” is an error of omission, and pressing the bar after the “X” is an error of commission. Eleven age and gender-corrected indices of attention are derived from the patient's performance. We selected the overall index, a weighted algorithm of the 11 component scores, as a representation of attentional abilities. Unlike the other tests given, higher scores on the overall index indicate worse performance.

California Verbal Learning Test (CVLT)

The CVLT assesses short-term and long-term memory and verbal learning in children ages 5–16 years.34 A standardized word list is presented to the patient on five consecutive trials, and the patient is asked to recall the words after each trial. The test requires approximately 20 minutes to administer and includes a 20-minute delay for analysis of long-term free and cued recall and recognition memory. The computerized scoring program provides standardized (mean = 50; SE = 10) and z-scores corrected for age. Numerous indices are available. We selected the score from List A Total Recall to represent an index of verbal memory abilities.

Abbreviated Wechsler Individual Achievement Test (WIAT)

The WIAT is a standardized test of academic achievement, individually administered with acceptable reliability and validity.35 The screening WIAT, which included the Basic Reading, Spelling, and Mathematics Reasoning subtests only, takes approximately 30 minutes to administer. Standardized scores are based on age-adjusted normative samples (normative mean, 100; SD = 15).

MRI Scan

MRI evaluations were performed on a 1.5 T Magnetom (Siemens Medical Systems, Iselin, NJ) whole body imager using the standard circular polarized volume head coil. T1, T2, and PD images were acquired on all patients as transverse 5-mm thick slices with a 1-mm gap interleaved to avoid crosstalk between slice excitations. T1 images were acquired using a gradient-echo FLASH-2D imaging sequence (TR/TE = 266/6 ms, 90-degree flip angle, 192 phase encodes, 3 acquisitions). T2 and PD images were acquired simultaneously using a dual spin-echo sequence (TR/TE1/TE2 = 3500/19/93 ms, 2 echoes, 192 phase encodes, 1 acquisition). The imaging protocol for the current study was chosen because it is a routine element in the evaluation of patients at our institution. Standard positioning beams on the magnet and head immobilization devices built into the head coil by the manufacturer were sufficient to ensure adequate head positioning and immobilization in these studies.

Quantitative MRI Scan Volumetrics

Image registration, a process of alignment so that the individual points in an image correspond to the same anatomic tissue in a related image, was performed within each examination. A single transverse section at the level of the basal ganglia, including both the genu and splenium of the corpus callosum and generally showing the putamen and the lateral ventricle, was selected as the index slice for this investigation. This allowed quantification of both interhemispheric and intrahemispheric white matter tracts. This representative index slice was chosen to sample cortical gray matter, white matter, central gray matter structures, and ventricular CSF (Fig. 2), and it has been shown to be highly predictive of full cerebrum volumes in other patient populations.36

thumbnail image

Figure 2. The images (from left to right) are T1-weighted, T2-weighted, PD-weighted, and segmented output, at level of index slice, for a 10.6-year-old female patient treated with chemotherapy and craniospinal and local irradiation for posterior fossa medulloblastoma at 4.3 years of age. Each color in the segmented output image represents a classified tissue type: gray matter (yellow), white matter (green), and cerebrospinal fluid [CSF] (blue).

Download figure to PowerPoint

Volumes of brain parenchyma on MRI scan images were assessed quantitatively using a fully automated hybrid neural network segmentation and classification method that is able to separate NAWM from abnormal white matter, gray matter, and CSF.28, 37 The resulting classified regions were mapped to a color scheme similar to that used for positron emission tomography. Gray matter was colored yellow, whereas white matter was colored green. The CSF was colored a light blue, whereas blood vessels and membranes were colored dark violet. For contrast purposes, the background was colored black. A histogram for each color was then completed to determine the number of pixels present, which then was multiplied by pixel volume to determine the sampled volume of each tissue type. Robust reliability and validity have been established for these methods.24, 28

Statistical Analyses

Our primary objective was to test the hypothesis that the association between reduced NAWM volumes and intellectual deficits can be explained by deficits in memory and attention, ultimately resulting in declines in academic achievement. The hypothesized model for these relationships is illustrated in Figure 1. Analyses were performed to test five possible pathways: 1) attentional and memory abilities explain the relationship between NAWM volume and FSIQ scores; 2) memory abilities alone explain the relationship between NAWM volume and FSIQ scores; 3) attentional abilities alone explain the relationship between NAWM volume and FSIQ scores; 4) memory abilities explain the relationship between attentional abilities and FSIQ scores; and 5) FSIQ scores explain the relationship between attentional abilities and academic performance. The conditions necessary to test for statistical inference were evaluated by computing partial correlation coefficients, controlling for age at RT and time since RT in all analyses because of their well-documented influence on the relationship between RT and IQ. The final model was tested using multiple regression analysis. Because the hypotheses predicted directional relationships between variables, one-tailed probability tests of significance (alpha = 0.05) were used.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Performance on Neurocognitive Measures Relative to Normative Standards

All mean FSIQ scores for patients in the current study were significantly lower than expected based on age-corrected test norms (mean, 82.5 vs. 100; P < 0.001). The mean CPT Overall Index, which was used to assess attentional abilities, also was significantly worse (mean, 8.46 vs. 0.00; P < 0.001) when compared with age and gender-corrected normative values. The CVLT List A Total Recall score, which measures overall verbal memory performance, also was significantly lower than age-corrected test norms (mean, 39.4 vs. 50.0; P < 0.001). Finally, academic achievement scores for Reading (mean, 83.8 vs. 100), Spelling (mean, 81.1 vs. 100), and Math (mean, 81.7 vs. 100) all were significantly lower (P < 0.001) than age-adjusted test norms.

NAWM and FSIQ

The first hypothesis, that attention and memory abilities explain the relationship between decreased NAWM volume and lower FSIQ, was tested by computing five partial correlations (rp). The NAWM volume had a significant and positive association with FSIQ (rp = 0.32; P = 0.026) and with attentional ability (rp = 0.49; P = 0.001). Attentional ability also had a significant and positive association with memory ability (rp = 0.38; P = 0.016), as did memory ability with FSIQ (rp = 0.46, P = 0.005). The final analysis revealed that after controlling for attentional and memory abilities, the association between NAWM volume and FSIQ was diminished and nonsignificant (rp = 0.07; P > 0.05). This result supports the hypothesis that attention and/or memory abilities explain the association between NAWM volume and FSIQ.

The second hypothesis, that memory ability alone explains the relationship between decreased NAWM volume and lower FSIQ, was tested by computing additional partial correlations. It had been shown in the previous analysis that NAWM volume and memory ability each had a significant and positive association with FSIQ. However, memory ability was not associated significantly and directly with NAWM volume (P > 0.05), preventing an evaluation of the impact of memory abilities on the relationship between NAWM volume and FSIQ.

The third hypothesis, that attentional abilities alone explain the relationship between decreased NAWM volumes and lower FSIQ, was tested by computing additional partial correlations. It was shown in the previous analyses that NAWM volume had a significant and positive association with FSIQ and with attentional abilities. Attentional abilities also had a significant and positive association with FSIQ (rp = 0.65; P < 0.001). The final analysis revealed that after controlling for attentional abilities, the association between NAWM volume and FSIQ was diminished and became nonsignificant (rp = 0.01; P > 0.05). This result supports the hypothesis that attentional abilities can explain a significant portion of the association between NAWM volumes and FSIQ.

The fourth hypothesis, that memory abilities explain the relationship between deficits in attentional abilities and lower FSIQ, was tested by computing additional partial correlations. It was shown in the previous analyses that attentional abilities had a significant and positive association with FSIQ and with memory abilities. In addition, memory abilities had a significant and positive association with FSIQ. The final analysis revealed that after controlling for memory abilities, the association between attentional abilities and FSIQ was not diminished significantly and remained statistically significant (rp = 0.55; P = 0.001). This result suggests that memory abilities cannot account for the significance of the relationship between attentional abilities and FSIQ.

Attention and Academic Achievement

The hypothesis that FSIQ explains the relationship between deficits in attentional abilities and academic achievement was tested by computing partial correlations. It was shown in previous analyses that attentional abilities had a significant and positive association with FSIQ. Attentional abilities also had a significant and positive association with the academic achievement scores for reading (rp = 0.48, P < 0.001), spelling (rp = 0.55, P < 0.001), and math (rp = 0.68; P < 0.001). As expected, FSIQ had a significant and positive association with academic achievement (reading, rp = 0.68; spelling, rp = 0.67; math, rp = 0.85; all P < 0.001). The final analysis revealed that after controlling for FSIQ, the associations between attentional abilities and reading (rp = 0.06, P > 0.05) and spelling (rp = 0.20, P > 0.05) were diminished and not statistically significant. This result supports the hypothesis that FSIQ is responsible for a significant amount of the association between attentional abilities and reading and spelling. However, the association between attentional abilities and math was not diminished significantly and remained statistically significant (rp = 0.30; P = 0.034). This result suggests that FSIQ cannot account for the significance of the relationship between attentional abilities and math.

Developmental Model Relating NAWM and Academic Achievement

The final developmental model (Fig. 3) relates posttherapy decreases in NAWM volumes to subsequent deficits in attentional abilities, which then result in decreased FSIQ and ultimately poor academic achievement. The predictive validity of this model for academic achievement was evaluated by multiple regression analysis. Variables were entered in the following order: age at RT, time from RT, NAWM volume, attentional ability, and FSIQ. Table 2 provides the statistically significant results of this model for each of the three academic achievement measures.

thumbnail image

Figure 3. Developmental model relating normal-appearing white matter (NAWM) to academic achievement through attention and intelligence. FSIQ: full-scale intelligence quotience.

Download figure to PowerPoint

Table 2. Results of Multiple Regression Analyses with ANOVA for Developmental Model Relating NAWM to Academic Achievement through Attention and Intelligence
ModelR2SE of estimateSignificance
  1. ANOVA: analysis of variance; NAWM: normal-appearing white matter; SE: standard error.

Reading0.5912.78P < 0.001
Spelling0.5913.28P < 0.001
Math0.798.17P < 0.001

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Although the brain may appear normal with conventional clinical MRI scans in many long-term survivors of childhood cancer, recent studies have shown that subtle decreases in NAWM volume can be quantified and that this phenomenon is associated with deficits in neurocognitive function.25–27 It is not known whether the underlying cause of white matter injury is direct (i.e., treatment-induced glial cell death) or indirect (i.e., destruction of oligodendrocyte precursors and/or microvascular damage).38 However, the current study shows that intellectual decline and subsequent decreased academic achievement are associated with fundamental neurocognitive impairments, primarily in the area of attention. Correlational analyses of neurocognitive variables and brain parenchyma volumes suggest that there is a link between the changes that occur in the brain and the impairments that arise in cognitive functioning after treatment-related injury to the CNS. Specifically, changes in NAWM volume are associated with impaired performance on attentional measures for long-term survivors.

Our results were consistent with the hypothesis that declining IQ scores among pediatric BT survivors3 are related to their inability to keep up with peers in the acquisition of new learning,18 which is facilitated by cognitive processes that are affected by treatment-related injury to white matter.25–27 These findings resulted in a developmental model based on a diverse group of survivors of pediatric BT. The model illustrated a pathway by which treatment-induced damage of white matter results in deficient academic achievement and a potentially poor quality of life;7, 11 it explains approximately 60% of the variance in reading and spelling and almost 80% of the variance in math (Table 2). It also provides a framework for explaining neurocognitive functioning in childhood BT survivors and is being validated in a larger ongoing prospective clinical trial.

There is little question that axon myelination confers a considerable functional advantage in terms of efficiency and speed of transmission.23, 39 Recent evidence suggests that glial cells may communicate actively with neurons and participate in the initiation of impulses.23 Myelination of axons continues into early adulthood, resulting in a net increase in the percent of brain parenchyma that is represented by white matter.39 It is noteworthy that the myelination process is completed last in the frontal and prefrontal lobes. These areas are known to subserve ‘executive functions’ involving planning and organizing of behavior as well as allocation of attentional resources.23 Previous studies by our group and others have demonstrated that patients who are younger at treatment have poorer neurocognitive outcomes.2, 5, 6, 9, 27 It is our contention that loss of normal white matter or failure to develop normal white matter can at least partially explain this relationship.25–27 The current results identify neurocognitive problems involving attention as a result of decreased white matter and reinforce the importance of intact attentional abilities for maintaining normal intellectual and academic development.

Interpretation of these results must include consideration of the inherent limitations of the study design. First, all quantitative imaging analyses were performed in a single slice that was chosen to be representative of diffuse changes associated with therapy. Although this section, in the absence of pathology, has been shown to be highly predictive of full cerebrum volumes,36 it may not capture injuries that are not evenly distributed in the brain. In addition, all sections were 5 mm thick, a measurement that is used routinely for clinical evaluations at our institution. The thickness of the sections caused partial voluming of some thinner white matter tracks but was inherent across all patients in the study. Second, although all patients in this study received RT to the primary site, which varied in histology and location, treatment sometimes was augmented with additional risk-adapted cranial irradiation (60% of patients) and/or chemotherapy (45% of patients). Risk adapted refers to the risk of recurrence, with high-risk patients receiving higher doses of craniospinal irradiation. However, Hoppe-Hirsch et al.3 illustrated that even ependymoma patients treated with local posterior fossa RT exhibited deficits in IQ. Only 56% of patients had IQ scores above 90 at 10 years post therapy, compared with approximately 75% expected in a normal population. Although the current study is unable to discern the impact of the addition of craniospinal irradiation, an ongoing study has been designed to assess regional white matter response to specific doses of irradiation in children treated for medulloblastoma of the posterior fossa. For these studies, the quantitative volumetric imaging is fused with radiation dosimetry maps to produce a calculated dose for every voxel of white matter in specific regions.

This preliminary study proposed a conceptual model that should be tested more definitively in a larger sample. The summary memory score used in this study (List A Total Recall) represents only a restricted set of memory abilities, including immediate encoding and recall and learning efficiency for words. This measure does not include other memory attributes measured by the CVLT (e.g., cued and delayed) or by other memory tests (e.g., working memory). These other memory abilities, which were not tested in our model, may influence IQ and academic achievement. In addition, although age at RT was included in the analyses as a covariate, the resulting conceptual model may differ for younger versus older children. However, the ratio of patients to variables needed for either of these potential analyses is not sufficient in the current sample. A larger study would enable further development, expansion, and validation of the proposed conceptual model using more sophisticated processing procedures such as path analysis or structural equation modeling to evaluate the relations.

The developmental model identifies a novel specific point for intervention (attentional abilities) to avoid or minimize the impact of therapy on the quality of life for pediatric BT survivors. A randomized, double-blind, placebo-controlled trial of methylphenidate (MPH) as a potential intervention for learning-impaired survivors has resulted in a significant improvement on measures of attention abilities in a laboratory setting.29 However, the efficacy of long-term MPH therapy is not yet known. Cognitive remediation focused on attentional abilities also is being investigated in pediatric cancer survivors. This multiinstitutional randomized clinical trial uses intensive one-on-one tutoring sessions. The content of these sessions is based on literature from pediatric and adult closed-head injury rehabilitation, special education, and developmental psychology.40 Finally, one should not underestimate the importance of informing parents, teachers, and primary care providers of the potential for attentional deficits and the impact on school performance so that the patient's decreased performance is not attributed to lack of motivation or other adjustment problems, which could delay intervention.41

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    Steen RG, Mirro J Jr. Childhood cancer. Cambridge, MA: Perseus, 2000.
  • 2
    Mulhern RK, Hancock J, Fairclough DL, Kun LE. Neuropsychological status of children treated for brain tumors: a critical review and integrative analysis. Med Pediatr Oncol. 1992; 20: 181191.
  • 3
    Hoppe-Hirsch E, Brunet L, Laroussinie F, et al. Intellectual outcome in children with malignant tumors of the posterior fossa: influence of the field of irradiation and quality of surgery. Childs Nervous System. 1995; 11: 340345.
  • 4
    Jannoun L, Bloom HJG. Long-term psychological effects in children treated for intracranial tumors. Int J Radiat Oncol Biol Phys. 1990; 18: 747753.
  • 5
    Dennis M, Spiegler BJ, Hetherington CR, Greenberg ML. Neuropsychological sequelae of the treatment of children with medulloblastoma. J Neurooncol. 1996; 29: 91101.
  • 6
    Lannering B, Marky I, Lundberg A, Olsson E. Long-term sequelae after pediatric brain tumors: their effect on disability and quality of life. Med Pediatr Oncol. 1990; 18: 304310.
  • 7
    Armstrong FD, Blumberg MJ, Toledano SR. Neurobehavioral issues in childhood cancer. School Psychol Rev. 1999; 28: 194203.
  • 8
    Ris MD, Noll RB. Long-term neurobehavioral outcome in pediatric brain tumor patients: review and methodological critique. J Clin Exp Neuropsychol. 1994; 16: 2142.
  • 9
    Mulhern RK, Kepner JL, Thomas PR, Armstrong FD, Friedman HS, Kun LE. Neuropsychologic functioning of survivors of childhood medulloblastoma randomized to receive conventional or reduced-dose craniospinal irradiation: a pediatric oncology group study. J Clin Oncol. 1998; 16: 17231728.
  • 10
    Packer RJ, Meadows AT, Rorke LB, Goldwein JL, D'Angio G. Long-term sequelae of cancer treatment on the central nervous system in childhood. Med Pediatr Oncol. 1987; 15: 241253.
  • 11
    Johnson DL, McCabe MA, Nicholson HS, et al. Quality of long-term survival in young children with medulloblastoma. J Neurosurg. 1994; 80: 10041010.
  • 12
    Mostow EN, Byrne J, Connelly RR, Mulvihill JJ. Quality of life in long-term survivors of CNS tumors of childhood and adolescence. J Clin Oncol. 1991; 9: 592599.
  • 13
    Roman DD, Sperduto PW. Neuropsychological effects of cranial radiation: current knowledge and future directions. Int J Radiat Oncol Biol Phys. 1995; 31: 983998.
  • 14
    Walter A, 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: 37203728.
  • 15
    Duffner PK, Cohen ME, Thomas PR. Late effects of treatment on the intelligence of children with posterior fossa tumors. Cancer. 1983; 51: 233237.
  • 16
    Ellenberg L, McComb JG, Siegel SE. Factors affecting intellectual outcome in pediatric brain tumor patients. Neurosurgery. 1987; 21: 638644.
  • 17
    Chapman CA, Waber DP, Berstein JH. Neurobehavioral and neurological outcome in long-term survivors of posterior fossa brain tumors: role of age and perioperative factors. J Child Neurol. 1995; 10: 209212.
  • 18
    Palmer SL, Goloubeva O, Reddick WE, et al. Patterns of intellectual development among survivors of pediatric medulloblastoma: a longitudinal analysis. J Clin Oncol. 2001; 19: 23022308.
  • 19
    Dennis M, Spiegler BJ, Hoffman HJ, Hendrick EB, Humphreys RP, Becker LE. Brain tumors in children and adolescents. I. Effects on working, associative and serial-order memory of IQ, age at tumor onset and age of tumor. Neuropsychologia. 1991; 29: 813827.
  • 20
    Rodgers J, Horrocks J, Britton PG, Kernathan J. Attentional ability among survivors of leukaemia. Arch Dis Child. 1999; 80: 318323.
  • 21
    Lockwood KA, Bell TS, Colegrove RW Jr. Long-term effects of cranial radiation therapy on attention functioning in survivors of childhood leukemia. J Pediatr Psychol. 1999; 24: 5566.
  • 22
    Schatz H, Kramer JH, Ablin A, Matthay KK. Processing speed, working memory and IQ: a developmental model of cognitive deficits following cranial radiation therapy. Neuropsychology. 2000; 14: 189200.
  • 23
    Filley CM, Kleinschmidt-DeMasters BK. Toxic leukoencephalopathy. N Engl J Med. 2001; 345: 425432.
  • 24
    Reddick WE, Mulhern RK, Elkin TD, Glass JO, Merchant TE, Langston JW. A hybrid neural network analysis of subtle brain volume differences in children surviving brain tumors. Magn Reson Imaging. 1998; 16: 413421.
  • 25
    Mulhern RK, Reddick WE, Palmer SL, et al. Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol. 1999; 46: 834841.
  • 26
    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: 787793.
  • 27
    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: 472479.
  • 28
    Reddick WE, Glass JO, Cook EN, Elkin TD, Deaton R. Automated segmentation and classification of multispectral magnetic resonance images of brain using artificial neural networks. IEEE Trans Med Imaging. 1997; 16: 911918.
  • 29
    Thompson S, Leigh L, Christensen R, et al. Immediate neurocognitive effects of methylphenidate on learning-impaired survivors of childhood cancer. J Clin Oncol. 2001; 19: 18021808.
  • 30
    The Psychological Corporation. The Wechsler Intelligence Test for Children–III (3rd edition). New York: Harcourt, Brace, Jovanovich, 1997.
  • 31
    The Psychological Corporation. Wechsler Adult Intelligence Scale–revised (revised edition). New York: Harcourt, Brace, Jovanovich, 1989.
  • 32
    Sattler JM. Assessment of children, revised ed. La Mesa, CA: Jerome M. Sattler, 1992.
  • 33
    Conners CK. The Conner's Continuous Performance Test. Toronto: Multi-Health Systems, 1995.
  • 34
    Delis D, Kramer JH, Kaplan E, Ober B. California Verbal Learning Test. New York: Harcourt, Brace, Jovanovich, 1994.
  • 35
    The Psychological Corporation. The Wechsler Individual Achievement Test. New York: Harcourt, Brace, Jovanovich, 1992.
  • 36
    Glass JO, Reddick WE, Ji Q, Glas LS. Segmentation and classification of normal appearing brain - how much is enough? SPIE Medical Imaging Symposium, Conference on Image Processing 4684, February 23–28, 2002; San Diego, CA. Bellingham, WA: SPIE - The International Society for Optical Engineering, 2002: 216224.
  • 37
    Reddick WE, Glass JO, Langston JW, Helton KJ. Quantitative MRI assessment of leukoencephalopathy. Magn Reson Med. 2002; 47: 912920.
  • 38
    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: 265275.
  • 39
    Barkovitch AJ. Concepts of myelin and myelination in neuroradiology. Am J Neuroradiol. 2000; 21: 10991109.
  • 40
    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 Neuropsych Soc. 2002; 8: 115124.
  • 41
    Mulhern RK, Palmer SL. Neurocognitive late effects in pediatric cancer. Curr Problems Cancer. In press.