Intellectual and Functional Outcome of Children 3 Years Old or Younger Who Have CNS Malignancies
http://www.100md.com
▲還散笫雖悝◎
the Departments of Hematology-Oncology, Biostatistics and Epidemiology, Behavioral Medicine, and Radiological Sciences, Neurology, St Jude Children's Research Hospital, Memphis, TN
ABSTRACT
PURPOSE: To evaluate the impact of tumor location, clinical parameters, and therapy on neurocognitive, neuroendocrine, and functional outcomes in children 3 years old with intracranial CNS malignancies who survived at least 2 years after diagnosis.
PATIENTS AND METHODS: Records were retrospectively reviewed for 194 children diagnosed from 1985 to 1999 at St Jude Children's Research Hospital (Memphis, TN).
RESULTS: The median age at diagnosis was 1.8 years (range, 0.1 to 3.5 years). Median follow-up was 7.64 years (2.0 to 19.4 years). Tumors were infratentorial (102), diencephalic (53), and hemispheric (39); 47% required ventriculoperitoneal shunts, 36% developed seizure disorders, and 20% developed severe ototoxicity. Therapy included no radiation therapy (RT) in 57 (30%), local RT in 87 (45%), and craniospinal irradiation (CSI) in 49 (25%). Overall survival at 10 years was 78 ㊣ 4%. In a longitudinal analysis of 126 patients with at least one neurocognitive evaluation (NE), the mean rate of intelligence quotient (IQ) change for patients who received CSI (每1.34 points per year) and local RT (每0.51 points per year) was significantly different from the no RT group (0.91 points per year; P = .005 and P = .036, respectively). Patients with hemispheric tumors had a significantly greater IQ decline (每1.52 points per year) than those with midline tumors (0.59 points per year; P = .038). Among those with NE 5 years after diagnosis, 71.4% of CSI recipients compared with 23% of local RT recipients had IQ less than 70 (P = .021). Patients undergoing CSI were more likely to develop endocrinopathies (P < .0001) and to require special education (P = .0007).
CONCLUSION: In young children with CNS tumors, CSI and hemispheric location are associated with significant declines in IQ scores.
INTRODUCTION
Approximately 20% of pediatric CNS tumors occur in children younger than 3 years old. Although treatment for resectable low-grade gliomas includes surgery followed by observation, most young children with CNS tumors require more aggressive treatment, including chemotherapy and/or radiotherapy (RT). Studies have demonstrated the adverse neurocognitive sequelae associated with RT in older children with brain tumors, reporting significant declines in intelligence quotient (IQ) over time among patients treated with cranial irradiation.1-13 Limitations of these reports have included small patient numbers and considerable variability in follow-up and treatment. Unfortunately, there are no comprehensive reports of neurocognitive, functional, and neuroendocrine outcome in children younger than 3 years treated for CNS malignancies. Such data would prove invaluable in defining future treatment strategies, including the role of early RT, in this poor-prognosis patient population.
We examine the impact of tumor location, clinical parameters, and therapy on neurocognitive, neuroendocrine, and functional outcome in children 3 years old with CNS malignancies in the largest reported cohort diagnosed and treated at a single institution.
PATIENTS AND METHODS
Patients
Between January 1985 and December 1999, 224 patients with intracranial CNS tumors who were diagnosed at 3 years old and survived for at least 2 years after diagnosis were evaluated and treated at St Jude Children's Research Hospital. Thirty patients with ependymoma who were treated with up-front focal conformal radiation therapy on an institutional protocol, representing a significant change in treatment strategy, have been reported separately.14 The 194 remaining patients are the subject of this report. Patients were enrolled on protocols approved by the institutional review board. Informed consent was obtained from parents or guardians at protocol enrollment. Seventy-six had a gross total resection at diagnosis (no surgical or imaging evidence of residual disease), nine had nearly total resection (< 10% residual tumor by surgical report and postoperative imaging), 53 had subtotal resection (< 50% residual tumor by surgical report and postoperative imaging), 37 had biopsy, and 19 had no tumor surgery (17 with optic pathway gliomas and two with brainstem gliomas).
A chart review documented clinical features, including location of initial tumor (infratentorial, midline [diencephalic], or hemispheric), type of therapy (including type of radiation therapy [RT]: local, craniospinal [CSI], or none), presence or absence of ventriculoperitoneal (VP) shunt, grade 3 or 4 hearing loss, seizure disorder, endocrinopathy, neurocognitive outcome, and other measures of functional outcome (ie, type of schooling, employment, vision, dexterity, ambulation, speech) as well as overall and progression-free survival.
Therapy
The intent of neurosurgery was to perform a gross total resection in all patients with tumors amenable to surgery. In general, the role of chemotherapy was to delay the institution of radiotherapy for at least a year, if possible. Depending on tumor type, radiation therapy was used as planned consolidative therapy (n = 52) or at the time of progression (n = 84). For those who received irradiation, the RT exposure was categorized as local (including local conformal), craniospinal, or none.
Disease Surveillance and Follow-Up
Patients were monitored regularly after completion of therapy. Neuroimaging and clinical assessments were performed every 3 to 4 months for the first 2 to 3 years, every 4 to 6 months until 5 years after diagnosis, and annually thereafter. Functional outcome, including education type, marital status, living arrangements, speech, ambulation, and dexterity, were assessed at the time of regular follow-up. The data reported for functional outcome reflect information obtained at the last follow-up. Patients were referred for endocrinologic assessment for routine surveillance and/or clinical suspicion of endocrinologic deficits and were followed on a semiannual or annual basis.
During platinum-containing therapy, audiology assessments were routinely conducted using auditory brainstem responses or audiograms before and after RT, and annually thereafter. The ototoxicity reported is the worst grade recorded for each patient.
Opthalmologic testing was conducted in all patients with hypothalamic optic pathway tumors and coincided with their neuroimaging assessments. Others were followed by an ophthalmologist if patients or clinicians suspected specific visual problems. Data from the last examination are reported.
Neuropsychological Evaluation
The institutional standard is to conduct psychological testing at least annually in all patients. During the study interval, patients 30 months of age or younger were evaluated with the Bayley Scales of Infant Development,15,16 which results in a Mental Development Index score. Those older than 30 months were most often evaluated with the age-appropriate Wechsler Scales of Intelligence,17-21 administered either in the standard format or in an abbreviated format at selected intervals, resulting in an estimation of full-scale IQ.22 These tests, regardless of version, result in age-corrected scores with a mean of 100 and a standard deviation of 15.
To estimate premorbid developmental trajectories, we computed an approximation of IQ based on a demographic algorithm (parental education, race) validated on the normative sample of the Wechsler Intelligence Scale for Children-III as well as children experiencing traumatic brain injury,23 where ethnicity is coded as two variables (white = 1 or non-white = 0, and black = 1 or non-black = 0) and parental education for both parents is also coded (0 to 8 years = 1, 9 to 11 years = 2, 12 years or GED = 3, 13 to 15 years = 4, 16 + years = 5) in the formula: premorbid IQ = 5.44 (mean parental education) + 2.80 (white/non-white) 每 9.01 (black/non-black) + 81.68.
Statistical Considerations
The frequencies of endocrinopathies, clinical factors, and functional outcome were descriptively provided, and exact 2 tests were used to assess associations with RT type.
Progression-free survival (PFS) was measured from the date of diagnosis to the date of first progression (or death, whichever was first) or last contact. Overall survival (OS) was measured from the date of diagnosis to death or last contact. Distributions of PFS and OS were estimated using Kaplan-Meier methods. The current cohort included a selected group of patients who lived 2 years after their diagnosis, making direct comparisons of survival estimates with other studies difficult.
Mean rate of change in IQ scores was assessed using a random coefficients model in which the IQ score is a linear function of the time since diagnosis, and both the intercept and the slope are considered random effects. The model fit was obtained using the MIXED procedure in SAS (SAS/STAT User＊s Guide, Version 6, 1990; SAS Institute, Cary, NC). The longitudinal effects of RT, tumor location, shunt, seizure disorder, neurofibromatosis type 1 (NF1), and ototoxicity on the IQ scores were investigated independently. Factors significant at the P < .05 level were investigated in a multivariate model. Because status of RT, shunt, seizure disorder, and ototoxicity variables vary over time, they were included in the models as time-dependent variables.24,25
Patients who had only one neurocognitive assessment were kept in the analysis to improve the baseline estimate and reduce variation. Removing these patients did not significantly affect the slope estimates.
RESULTS
Characteristics of the Study Population
The distribution of clinical characteristics of the 194 patients diagnosed with an intracranial CNS tumor at 3 years old and surviving at least 2 years are summarized in Table 1. Histologic confirmation was available in 175 patients. Locations included infratentorial (102), diencephalic (53), and hemispheric (39). Ninety-one patients (47%) required VP shunts at a median of 0.03 years from diagnosis, 70 (36%) had a seizure disorder diagnosed at a median of 0.49 years, and 22 (11.3%) had NF1.
Treatment and Outcome
The OS for the whole group at 10 years was 78 ㊣ 4%, with a PFS of 34 ㊣ 5% (median follow-up, 7.64 years; range, 2.0 to 19.4 years; Fig 1A). OS in patients with high-grade gliomas, ependymomas, and medulloblastoma/primitive neuroectodermal tumor are summarized in Figure 1B.
Chemotherapy
Patients were enrolled on a number of different protocols. Most patients with malignant embryonal tumors of infancy received a combination of cyclophosphamide, vincristine, etoposide, and platinum-containing regimens as their front-line therapy. Those with midline low-grade tumors generally received platinum-containing regimen with or without vincristine. Overall, 145 patients received chemotherapy at some point during therapy. Thirty-seven patients underwent neither chemotherapy nor RT at any time.
Radiation Therapy
Overall, 49 patients received CSI (22 as part of their primary therapy and 27 as part of their retrieval strategy after progression) and 87 received only local RT (30 as part of their primary therapy and 57 as part of their retrieval strategy). Types of local RT included three-dimensional conformal/intensity-modulated RT (n = 42), parallel opposed fields RT (n = 41), Gamma Knife (n = 2), and intracystic 32P (n = 2). Fifty-eight patients did not receive RT at any time.
The median age at RT for the entire cohort was 3.1 years (range, 0.7 to 11.2 years). The median age at initial diagnosis was 1.79 years. Among those who received CSI (median age, 3.1 years; range, 1.1 to 14.6 years), 34 patients received cranial doses greater than 30 Gy (range, 30 to 48.4 Gy), 11 patients received 23.4 Gy, and four patients received 18 Gy. CSI was followed by local RT to the primary site with a median dose of 51.9 Gy.
Among the 42 patients who received at most local conformal RT (median age, 3.3 years; range, 1.0 to 10.8 years), the median dose of RT was 54.0 Gy (range, 45.0 to 59.4 Gy); of the 41 patients who received at most parallel opposed field local RT (median age, 3.1 years; range, 0.7 to 9.7 years), the median dose was 50.4 Gy (range, 40 to 70.2 Gy).
Follow-Up Information
Eighty percent of patients had follow-up within 3 years of the data extraction date, with a median of 8 months since the prior contact date. Data reported regarding education, ambulation, dexterity, marital status, and living arrangements were obtained from the last follow-up. The median follow-up was 7.64 years (range, 2.0 to 19.4 years).
Table 2 summarizes the number of patients with neurocognitive and audiology assessments at selected time points from diagnosis, demonstrating that the majority were obtained in the first 5 years after diagnosis.
Functional Outcome
Functional outcome, including endocrinopathies, ototoxicity, visual acuity, education type, marital status, living arrangements, hearing, vision, speech, ambulation, and dexterity are summarized in Table 3. Neurocognitive outcomes are summarized in Tables 4 and 5. Overall, 52% developed at least one endocrinopathy, 21% grade 3 or 4 ototoxicity (69% had infratentorial tumors, 13% had diencephalic tumors), and 18% were legally blind (visual acuity 20/200 in at least one eye).
Twenty-four percent attended regular school and 36% were in special education (Table 3). The former group was significantly less likely to have received CSI (8.5%) than the latter (40%; P = .0007).
Only 13 patients were 18 years old at last follow-up (range, 18.2 to 20.7 years). For the nine with available information, one was living independently, one was living in a group home, and seven were living at home with their parents. None was married, and only one had part time employment. Abnormalities in speech, ambulation, and dexterity were noted in 51%, 40%, and 34%, respectively, of the entire cohort, at last follow-up.
Endocrinopathies
Patients who received RT had a significantly greater likelihood of developing endocrinopathy than those who did not (P < .0001); 90% of patients who received CSI had an endocrinopathy compared with 68% and 26% of those who received local RT and no RT, respectively. There was a significantly higher incidence of growth hormone deficiency (P < .0001), adrenocorticotropic hormone deficiency (P = .0042), and hypothyroidism (P < .0001) in patients who received CSI compared with those who received local RT or no RT.
Effects of RT on Neurocognitive Outcome
Cross-sectional IQ analyses. Among 67 patients who had at least one IQ exam 5 years after diagnosis, we estimated the frequency of severe mental retardation (ie, IQ < 70) based on the type of RT received (Table 4). Among patients who received CSI, 71.4% had severe mental retardation, compared with 23% of those who received at most local RT and 20% of patients who received no RT (P = .021).
Longitudinal IQ analysis. Of the 194 patients, 101 had sufficient data to estimate premorbid IQ. The obtained estimate was 100.86 (standard deviation, 7.0), suggesting that the mean IQ is projected to have been in the normal range for age. Of the 194 patients, 126 (65%) had at least one IQ assessment and are included in the longitudinal analysis, examining the effect of tumor location, RT, VP shunt, seizure, hearing loss, and NF1 on IQ over time. The mean IQ at the time of first assessment (median, 1.1 years) was 82 (range, 46 to 130). The median number of IQ assessments was four (range, one to seven). Among those with an IQ assessment within 6 months of diagnosis (n = 44), the mean initial IQ was 84 (range, 50 to 120). Table 5 summarizes the clinical characteristics of the patients included in the longitudinal IQ analysis.
The subgroup of 126 patients with at least one IQ assessment was compared with the cohort of 194 patients using the exact 2 test. Of all the demographic and clinical variables, only seizure disorder was distributed significantly differently, with 25% of those with IQ assessments having seizures compared with 44% of those without (P = .007).
Overall, there was a mean IQ decline of 每0.29 points per year in all patients. The IQ increase of 0.91 points per year in patients who received no RT was significantly different from the IQ decline in patients who underwent CSI (每1.34 points per year; P = .005; Fig 2A) or local RT (每0.51 points per year; P = .036), respectively. Although patients who received CSI (每1.34 points per year) seemed to have a worse neurocognitive outcome than those who received local RT (每0.51 points per year), the difference was not statistically significance (P = .26). Recipients of more than 33 Gy CSI experienced a greater IQ decline (每1.7 points per year) than those receiving 24 Gy CSI (每0.63 points per year). However, the numbers of patients in the two categories were too small to report statistical significances.
The IQ decline of patients who received CSI as planned consolidative therapy (每0.31 points per year) was not statistically different from those who received CSI at the time of progression (每1.26 points per year; P = .39).
Among patients with infratentorial tumors, those receiving CSI had a significant decline in IQ (每1.34 points per year) compared with those who received local RT (0.35 points per year; P = .032) and those who received no RT (1.28 points per year; P = .009). Although patients undergoing local RT or no RT tended to have a slight gain in IQ points per year, the rate of change did not differ significantly (P = .34) between these two groups. There was no statistically significant difference in the rate of IQ decline for those who received conformal RT compared with parallel opposed fields RT (P = .71). However, the latter results should be interpreted with caution because of the small sample size.
To assess the effect of age at RT on neurocognitive outcome, we reanalyzed the data to determine rates of IQ decline in patients who underwent CSI at less than 3 years or at 3 years. The IQ decline (每1.97 points per year) for the younger group was not significantly different from the decline of 0.72 IQ points per year for the latter group (P = .19). Because current infant protocols include patients younger than 3 years at presentation, we reanalyzed the data to include only patients younger than 3 years. All model estimates were statistically similar to those obtained for the whole cohort.
Effects of Other Patient Variables on Neurocognitive Outcome
Patients with hemispheric tumors had a significantly greater IQ decline (每1.52 points per year) than those with midline tumors (0.59 points per year; P = .038; Fig 2B). Although patients with VP shunts had lower initial IQ scores, shunting did not significantly affect the rate of IQ decline (P = .69; Fig 2C). Although patients with chronic seizures had a greater decline in IQ scores (每1.02 points per year) than those without seizures (0.004 points per year), the effect was not statistically significant (P = .123; Fig 2D). Severe ototoxicity did not significantly affect the IQ scores over time (P = .58; Fig 1F). Finally, patients with NF1 had an increase in IQ scores (1.7 points per year) compared with those without (每0.5 IQ points per year; P = .024). This may reflect the fact that only 7% of NF1 patients received CSI compared with 30% of those without NF1 (Fig 2E).
Effects of Tumor Location and RT on Neurocognitive Outcome
Only RT and tumor location had a significant effect (P < .05) on neurocognitive outcome independently. We further investigated the combination of RT and tumor location on the change in neurocognitive outcome. To increase the number of patients in each category, we combined CSI and local RT, resulting in no RT/midline (n = 17), no RT/infratentorial (n = 16), no RT/hemispheric (n = 14), RT/midline (n = 13), RT/infratentorial (n = 56), and RT/hemispheric (n = 10). In the multivariable analysis, patients with midline or infratentorial tumors who did not receive RT had an increase in IQ of 3.59 or 1.39 points per year, respectively, which was significantly different from the IQ declines for the other four subgroups (maximum P value is .03). In contrast, patients with hemispheric tumors who received RT suffered the most with an IQ decline of 2.81 points per year, followed by patients with midline tumors who had RT with a decline of 1.66 points per year.
DISCUSSION
This study demonstrates that, in young children with intracranial CNS tumors, treatment with CSI and hemispheric location are associated with significant declines in IQ scores. Patients undergoing CSI are more likely to develop endocrinopathies (P < .0001) and require special education (P = .0007). Among patients with infratentorial tumors, patients undergoing local RT (whether conformal or parallel opposed fields) had a slight gain in IQ that did not differ significantly from the gain in the no-RT group (P = .34).
Functional Outcome
The recent study by Packer et al26 of 1,607 patients in the Childhood Cancer Survivor Study is the only comprehensive review of the neurologic and neurosensory sequelae of children with brain tumors. Relative to sibling controls, survivors of brain tumors were at elevated risk for hearing and visual impairment. Coordination and motor problems were reported in 49% and 26%, respectively, seizure disorders in 25%, and hearing impairment in 12%. Although these data were self-reported and included patients up to 20 years old, the results are similar to those of the current cohort.
The relative effect of therapy on the development of endocrinopathies in children with CNS malignancies has been reported previously, with various reports confirming or questioning the role of RT in the development of endocrinopathies.10,27-31 In the present study, patients who received RT had a significantly greater likelihood of developing endocrinopathy than those who did not (P < .0001); 90% of those who received CSI had an endocrinopathy compared with 68% and 26% of those who received local RT and no RT, respectively.
The current report offers a comprehensive review of social attainment in this group of patients. The 36% of patients who required special education were significantly more likely to have received CSI than those in regular school. Only one of the 13 surviving patients older than 18 years was working independently, and none was married. Likewise, Kiltie et al11 reported that of 16 medulloblastoma patients alive 10 years after treatment, 15 required special education, only one was working independently, and none was married. Hoppe-Hirsch et al8 reported that, 10 years after therapy, 71% of patients with ependymoma (median age at presentation, 5.8 years) attended regular school compared with 20% of medulloblastoma patients (median age at presentation, 5.8 years). Zebrack et al32 recently reviewed the psychological outcomes in 1,101 long-term survivors of childhood CNS tumors younger than 18 years at diagnosis and reported that 74% were currently employed, 63% had graduated from high school, and 74% remained single; they concluded that diminished social functioning contributed to increased psychological distress.
Factors Affecting Neurocognitive Outcome
Previous studies have reported progressive neurocognitive decline after RT.4,6,9-12,14 A study of 19 infants with medulloblastoma from this institution treated with chemotherapy and CSI at progression demonstrated a decline of 3.9 IQ points per year,10 which is similar to the decline of the current larger cohort. The IQ decline for patients who received CSI or local RT only was significantly different from the IQ gain in the no RT group. The likelihood of having an IQ less than 70 at least 5 years after diagnosis was much higher in patients with CSI compared with those who received local or no RT. These trends are similar to those reported by Grille et al,7 in which an older group of patients with posterior fossa tumors receiving 0, 25, and 35 Gy had mean IQ scores of 84.5, 76.9, and 63.7, respectively. Similarly, Hoppe-Hirsch et al8 reported that in older children, 56% of patients with ependymoma (who received local RT only) maintained an IQ of 90 5 years after therapy compared with 22% of patients with medulloblastoma who received CSI.8 In our cohort, patients who received cranial doses 33 Gy also had a greater IQ decline than those receiving 24 Gy, but the numbers were too small to assess statistical significance. Consistent evidence of neurocognitive decline with CSI in children of different ages with brain tumors has also been reported by others.3,5,9,11,13
For infratentorial tumors, there was no significant IQ decline for those treated only with local irradiation, whether parallel opposed fields or conformal RT. The latter finding is in keeping with a report from this institution in which patients with ependymoma receiving conformal RT demonstrated no statistically significant change in IQ 24 months after completion of therapy.14
VP shunting did not have a significant impact on the rate of IQ decline, although patients with VP shunts tended to have lower initial IQs than those without. The relative effects of the tumor, surgery, and perioperative factors compared with treatment effects have always been difficult to delineate.9,33-35 In the current cohort, patients who had neurocognitive assessments done within 6 months of diagnosis demonstrated a lower baseline IQ than expected, confirming that the presence of tumor, surgery, and perioperative factors do have an effect on neurocognitive outcome regardless of subsequent treatment.
Although this study provides the most comprehensive overview of functional, social, and neurocognitive outcomes in a large cohort of infants with a long follow-up period, it has several limitations. Neurocognitive data were available in only 65% of the cohort. Our analysis demonstrated the relative similarity of the cohorts with and without neurocognitive follow-up. However, patients who were followed or had surgery only were less likely to undergo neurocognitive testing (54.9%) than those who received chemotherapy or RT (68.5%), probably because of the lower risk for neurocognitive sequelae. Similarly, there may have been a bias not to test patients who were more severely affected. Of those with neurocognitive data, a subset was included in an estimation of premorbid functioning using a prediction formula previously published.23 This constitutes an initial attempt to use such an inferential model. However, the model was developed for estimating full-scale IQ scores comparable with subsequent scores derived from the Wechsler Intelligence Scale for Children每Third Edition, recommended for children aged 6 through 16 years 11 months.18 In addition, the coding for ethnic classifications used to determine the model are restrictive (black/non-black, white/non-white) and may not be suitable for highly diverse populations. Extrapolating outside the data on which the model was developed may not be appropriate. Among patients with endocrinopathy, the endocrinopathies were coded as present or absent without regard to time of onset, thus making any conclusions regarding the relationship between endocrinopathy and type of therapy difficult to assess. Finally, the diversity of histology and tumor location in this cohort has implications for choice of therapy and, hence, long-term outcome in this particularly vulnerable group. It is imperative that future studies include prospective and comprehensive neurocognitive and functional assessments to accurately evaluate the impact of novel therapeutic approaches on long-term functional outcome.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Dr James Boyett for his input regarding the statistical analysis and Patsy Burnside for typing the manuscript.
NOTES
Supported by Center of Research Excellence support Grant No. CA21765 and by American Lebanese Syrian Associated Charities.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Palmer SL, Goloubeva O, Reddick WE, et al: Patterns of intellectual development among survivors of pediatric medulloblastoma: A longitudinal analysis. J Clin Oncol 19:2302-2308, 2001
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 19:472-479, 2001
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 19:3470-3476, 2001
Mulhern RK, Reddick WE, Palmer SL, et al: Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol 46:834-841, 1999
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 Pediatric Oncology Group study. J Clin Oncol 16:1723-1728, 1998
Dennis M, Spiegler BJ, Hetherington CR, et al: Neuropsychological sequelae of the treatment of children with medulloblastoma. J Neurooncol 29:91-101, 1996
Grill J, Renaux VK, Bulteau C, et al: Long-term intellectual outcome in children with posterior fossa tumors according to radiation doses and volumes. Int J Radiat Oncol Biol Phys 45:137-145, 1999
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 Nerv Syst 11:340-345, 1995
Copeland DR, deMoor C, Moore BD III, et al: Neurocognitive development of children after a cerebellar tumor in infancy: A longitudinal study. J Clin Oncol 17:3476-3486, 1999
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 17:3720-3728, 1999
Kiltie AE, Lashford LS, Gattamaneni HR: Survival and late effects in medulloblastoma patients treated with craniospinal irradiation under three years old. Med Pediatr Oncol 28:348-354, 1997
Chapman CA, Waber DP, Bernstein JH, et al: Neurobehavioral and neurologic outcome in long-term survivors of posterior fossa brain tumors: Role of age and perioperative factors. J Child Neurol 10:209-212, 1995
Spiegler BJ, Bouffet E, Greenberg ML, et al: Change in neurocognitive functioning after treatment with cranial radiation in childhood. J Clin Oncol 22:706-713, 2004
Merchant TE, Mulhern RK, Krasin MJ, et al: Preliminary results from a phase II trial of conformal radiation therapy and evaluation of radiation-related CNS effects for pediatric patients with localized ependymoma. J Clin Oncol 22:3156-3162, 2004
Bayley N: Bayley Scales of Infant Development (ed 2). New York, NY, The Psychological Corporation, 1993
Bayley N: Bayley Scales of Infant Development (ed 2). New York, NY, The Psychological Corporation, 1969
Wechsler D: Wechsler Adult Intelligence Scale〞Revised. New York, NY, The Psychological Corporation, 1989
Wechsler D: Wechsler Intelligence Scale for Children (ed 3). New York, NY, The Psychological Corporation, 1991
Wechsler D: Manual for the Weschler Intelligence Scale for Children〞Revised. San Antonio, TX, The Psychological Corporation, 1974
Wechsler D: Wechsler Preschool and Primary School Scales of Intelligence〞Revised. New York, NY, The Psychological Corporation, 1989
Wechsler D: Wechsler Preschool and Primary School Scales of Intelligence〞Revised. New York, NY, The Psychological Corporation, 1963
Sattler JM. Assessment of children. San Diego, CA, JM Sattler, 1992, pp 1170
Vanderploeg RD, Schinka JA, Baum KM, et al: WISC-III premorbid prediction strategies: Demographic and best performance approaches. Psychol Assess 10:277-284, 1998
Hardin J, Hilbe J: Generalized Estimating Equations. Boca Raton, FL, Chapman & Hall/CRC, 2003, pp 46-47
SAS Institute Inc: Longitudinal Data Analysis with Discrete and Continuous Responses; course notes. Cary, NC, SAS Institute Inc, 2004, pp 77-99
Packer RJ, Gurney JG, Punyko JA, et al: Long-term neurologic and neurosensory sequelae in adult survivors of a childhood brain tumor: Childhood cancer survivor study. J Clin Oncol 21:3255-3261, 2003
Constine LS, Woolf PD, Cann D, et al: Hypothalamic-pituitary dysfunction after radiation for brain tumors. N Engl J Med 328:87-94, 1993
Duffner PK, Cohen ME, Thomas PR, et al: The long-term effects of cranial irradiation on the central nervous system. Cancer 56:1841-1846, 1985
Kanev PM, Lefebvre JF, Mauseth RS, et al: Growth hormone deficiency following radiation therapy of primary brain tumors in children. J Neurosurg 74:743-748, 1991
Collet-Solberg PF, Sernyak H, Satin-Smith M, et al: Endocrine outcome in long-term survivors of low-grade hypothalamic/chiasmatic glioma. Clin Endocrinol (Oxf) 47:79-85, 1997
Merchant TE, Williams T, Smith JM, et al: Preirradiation endocrinopathies in pediatric brain tumor patients determined by dynamic tests of endocrine function. Int J Radiat Oncol Biol Phys 54:45-50, 2002
Zebrack BJ, Gurney JG, Oeffinger K, et al: Psychological outcomes in long-term survivors of childhood brain cancer: A report from the childhood cancer survivor study. J Clin Oncol 22:999-1006, 2004
Ater JL, Moore BD III, Francis DJ, et al: Correlation of medical and neurosurgical events with neuropsychological status in children at diagnosis of astrocytoma: Utilization of a neurological severity score. J Child Neurol 11:462-469, 1996
Chin HW, Maruyama Y: Age at treatment and long-term performance results in medulloblastoma. Cancer 53:1952-1958, 1984
Mulhern RK, Horowitz ME, Kovnar EH, et al: Neurodevelopmental status of infants and young children treated for brain tumors with preirradiation chemotherapy. J Clin Oncol 7:1660-1666, 1989(Maryam Fouladi, Elizabeth)
ABSTRACT
PURPOSE: To evaluate the impact of tumor location, clinical parameters, and therapy on neurocognitive, neuroendocrine, and functional outcomes in children 3 years old with intracranial CNS malignancies who survived at least 2 years after diagnosis.
PATIENTS AND METHODS: Records were retrospectively reviewed for 194 children diagnosed from 1985 to 1999 at St Jude Children's Research Hospital (Memphis, TN).
RESULTS: The median age at diagnosis was 1.8 years (range, 0.1 to 3.5 years). Median follow-up was 7.64 years (2.0 to 19.4 years). Tumors were infratentorial (102), diencephalic (53), and hemispheric (39); 47% required ventriculoperitoneal shunts, 36% developed seizure disorders, and 20% developed severe ototoxicity. Therapy included no radiation therapy (RT) in 57 (30%), local RT in 87 (45%), and craniospinal irradiation (CSI) in 49 (25%). Overall survival at 10 years was 78 ㊣ 4%. In a longitudinal analysis of 126 patients with at least one neurocognitive evaluation (NE), the mean rate of intelligence quotient (IQ) change for patients who received CSI (每1.34 points per year) and local RT (每0.51 points per year) was significantly different from the no RT group (0.91 points per year; P = .005 and P = .036, respectively). Patients with hemispheric tumors had a significantly greater IQ decline (每1.52 points per year) than those with midline tumors (0.59 points per year; P = .038). Among those with NE 5 years after diagnosis, 71.4% of CSI recipients compared with 23% of local RT recipients had IQ less than 70 (P = .021). Patients undergoing CSI were more likely to develop endocrinopathies (P < .0001) and to require special education (P = .0007).
CONCLUSION: In young children with CNS tumors, CSI and hemispheric location are associated with significant declines in IQ scores.
INTRODUCTION
Approximately 20% of pediatric CNS tumors occur in children younger than 3 years old. Although treatment for resectable low-grade gliomas includes surgery followed by observation, most young children with CNS tumors require more aggressive treatment, including chemotherapy and/or radiotherapy (RT). Studies have demonstrated the adverse neurocognitive sequelae associated with RT in older children with brain tumors, reporting significant declines in intelligence quotient (IQ) over time among patients treated with cranial irradiation.1-13 Limitations of these reports have included small patient numbers and considerable variability in follow-up and treatment. Unfortunately, there are no comprehensive reports of neurocognitive, functional, and neuroendocrine outcome in children younger than 3 years treated for CNS malignancies. Such data would prove invaluable in defining future treatment strategies, including the role of early RT, in this poor-prognosis patient population.
We examine the impact of tumor location, clinical parameters, and therapy on neurocognitive, neuroendocrine, and functional outcome in children 3 years old with CNS malignancies in the largest reported cohort diagnosed and treated at a single institution.
PATIENTS AND METHODS
Patients
Between January 1985 and December 1999, 224 patients with intracranial CNS tumors who were diagnosed at 3 years old and survived for at least 2 years after diagnosis were evaluated and treated at St Jude Children's Research Hospital. Thirty patients with ependymoma who were treated with up-front focal conformal radiation therapy on an institutional protocol, representing a significant change in treatment strategy, have been reported separately.14 The 194 remaining patients are the subject of this report. Patients were enrolled on protocols approved by the institutional review board. Informed consent was obtained from parents or guardians at protocol enrollment. Seventy-six had a gross total resection at diagnosis (no surgical or imaging evidence of residual disease), nine had nearly total resection (< 10% residual tumor by surgical report and postoperative imaging), 53 had subtotal resection (< 50% residual tumor by surgical report and postoperative imaging), 37 had biopsy, and 19 had no tumor surgery (17 with optic pathway gliomas and two with brainstem gliomas).
A chart review documented clinical features, including location of initial tumor (infratentorial, midline [diencephalic], or hemispheric), type of therapy (including type of radiation therapy [RT]: local, craniospinal [CSI], or none), presence or absence of ventriculoperitoneal (VP) shunt, grade 3 or 4 hearing loss, seizure disorder, endocrinopathy, neurocognitive outcome, and other measures of functional outcome (ie, type of schooling, employment, vision, dexterity, ambulation, speech) as well as overall and progression-free survival.
Therapy
The intent of neurosurgery was to perform a gross total resection in all patients with tumors amenable to surgery. In general, the role of chemotherapy was to delay the institution of radiotherapy for at least a year, if possible. Depending on tumor type, radiation therapy was used as planned consolidative therapy (n = 52) or at the time of progression (n = 84). For those who received irradiation, the RT exposure was categorized as local (including local conformal), craniospinal, or none.
Disease Surveillance and Follow-Up
Patients were monitored regularly after completion of therapy. Neuroimaging and clinical assessments were performed every 3 to 4 months for the first 2 to 3 years, every 4 to 6 months until 5 years after diagnosis, and annually thereafter. Functional outcome, including education type, marital status, living arrangements, speech, ambulation, and dexterity, were assessed at the time of regular follow-up. The data reported for functional outcome reflect information obtained at the last follow-up. Patients were referred for endocrinologic assessment for routine surveillance and/or clinical suspicion of endocrinologic deficits and were followed on a semiannual or annual basis.
During platinum-containing therapy, audiology assessments were routinely conducted using auditory brainstem responses or audiograms before and after RT, and annually thereafter. The ototoxicity reported is the worst grade recorded for each patient.
Opthalmologic testing was conducted in all patients with hypothalamic optic pathway tumors and coincided with their neuroimaging assessments. Others were followed by an ophthalmologist if patients or clinicians suspected specific visual problems. Data from the last examination are reported.
Neuropsychological Evaluation
The institutional standard is to conduct psychological testing at least annually in all patients. During the study interval, patients 30 months of age or younger were evaluated with the Bayley Scales of Infant Development,15,16 which results in a Mental Development Index score. Those older than 30 months were most often evaluated with the age-appropriate Wechsler Scales of Intelligence,17-21 administered either in the standard format or in an abbreviated format at selected intervals, resulting in an estimation of full-scale IQ.22 These tests, regardless of version, result in age-corrected scores with a mean of 100 and a standard deviation of 15.
To estimate premorbid developmental trajectories, we computed an approximation of IQ based on a demographic algorithm (parental education, race) validated on the normative sample of the Wechsler Intelligence Scale for Children-III as well as children experiencing traumatic brain injury,23 where ethnicity is coded as two variables (white = 1 or non-white = 0, and black = 1 or non-black = 0) and parental education for both parents is also coded (0 to 8 years = 1, 9 to 11 years = 2, 12 years or GED = 3, 13 to 15 years = 4, 16 + years = 5) in the formula: premorbid IQ = 5.44 (mean parental education) + 2.80 (white/non-white) 每 9.01 (black/non-black) + 81.68.
Statistical Considerations
The frequencies of endocrinopathies, clinical factors, and functional outcome were descriptively provided, and exact 2 tests were used to assess associations with RT type.
Progression-free survival (PFS) was measured from the date of diagnosis to the date of first progression (or death, whichever was first) or last contact. Overall survival (OS) was measured from the date of diagnosis to death or last contact. Distributions of PFS and OS were estimated using Kaplan-Meier methods. The current cohort included a selected group of patients who lived 2 years after their diagnosis, making direct comparisons of survival estimates with other studies difficult.
Mean rate of change in IQ scores was assessed using a random coefficients model in which the IQ score is a linear function of the time since diagnosis, and both the intercept and the slope are considered random effects. The model fit was obtained using the MIXED procedure in SAS (SAS/STAT User＊s Guide, Version 6, 1990; SAS Institute, Cary, NC). The longitudinal effects of RT, tumor location, shunt, seizure disorder, neurofibromatosis type 1 (NF1), and ototoxicity on the IQ scores were investigated independently. Factors significant at the P < .05 level were investigated in a multivariate model. Because status of RT, shunt, seizure disorder, and ototoxicity variables vary over time, they were included in the models as time-dependent variables.24,25
Patients who had only one neurocognitive assessment were kept in the analysis to improve the baseline estimate and reduce variation. Removing these patients did not significantly affect the slope estimates.
RESULTS
Characteristics of the Study Population
The distribution of clinical characteristics of the 194 patients diagnosed with an intracranial CNS tumor at 3 years old and surviving at least 2 years are summarized in Table 1. Histologic confirmation was available in 175 patients. Locations included infratentorial (102), diencephalic (53), and hemispheric (39). Ninety-one patients (47%) required VP shunts at a median of 0.03 years from diagnosis, 70 (36%) had a seizure disorder diagnosed at a median of 0.49 years, and 22 (11.3%) had NF1.
Treatment and Outcome
The OS for the whole group at 10 years was 78 ㊣ 4%, with a PFS of 34 ㊣ 5% (median follow-up, 7.64 years; range, 2.0 to 19.4 years; Fig 1A). OS in patients with high-grade gliomas, ependymomas, and medulloblastoma/primitive neuroectodermal tumor are summarized in Figure 1B.
Chemotherapy
Patients were enrolled on a number of different protocols. Most patients with malignant embryonal tumors of infancy received a combination of cyclophosphamide, vincristine, etoposide, and platinum-containing regimens as their front-line therapy. Those with midline low-grade tumors generally received platinum-containing regimen with or without vincristine. Overall, 145 patients received chemotherapy at some point during therapy. Thirty-seven patients underwent neither chemotherapy nor RT at any time.
Radiation Therapy
Overall, 49 patients received CSI (22 as part of their primary therapy and 27 as part of their retrieval strategy after progression) and 87 received only local RT (30 as part of their primary therapy and 57 as part of their retrieval strategy). Types of local RT included three-dimensional conformal/intensity-modulated RT (n = 42), parallel opposed fields RT (n = 41), Gamma Knife (n = 2), and intracystic 32P (n = 2). Fifty-eight patients did not receive RT at any time.
The median age at RT for the entire cohort was 3.1 years (range, 0.7 to 11.2 years). The median age at initial diagnosis was 1.79 years. Among those who received CSI (median age, 3.1 years; range, 1.1 to 14.6 years), 34 patients received cranial doses greater than 30 Gy (range, 30 to 48.4 Gy), 11 patients received 23.4 Gy, and four patients received 18 Gy. CSI was followed by local RT to the primary site with a median dose of 51.9 Gy.
Among the 42 patients who received at most local conformal RT (median age, 3.3 years; range, 1.0 to 10.8 years), the median dose of RT was 54.0 Gy (range, 45.0 to 59.4 Gy); of the 41 patients who received at most parallel opposed field local RT (median age, 3.1 years; range, 0.7 to 9.7 years), the median dose was 50.4 Gy (range, 40 to 70.2 Gy).
Follow-Up Information
Eighty percent of patients had follow-up within 3 years of the data extraction date, with a median of 8 months since the prior contact date. Data reported regarding education, ambulation, dexterity, marital status, and living arrangements were obtained from the last follow-up. The median follow-up was 7.64 years (range, 2.0 to 19.4 years).
Table 2 summarizes the number of patients with neurocognitive and audiology assessments at selected time points from diagnosis, demonstrating that the majority were obtained in the first 5 years after diagnosis.
Functional Outcome
Functional outcome, including endocrinopathies, ototoxicity, visual acuity, education type, marital status, living arrangements, hearing, vision, speech, ambulation, and dexterity are summarized in Table 3. Neurocognitive outcomes are summarized in Tables 4 and 5. Overall, 52% developed at least one endocrinopathy, 21% grade 3 or 4 ototoxicity (69% had infratentorial tumors, 13% had diencephalic tumors), and 18% were legally blind (visual acuity 20/200 in at least one eye).
Twenty-four percent attended regular school and 36% were in special education (Table 3). The former group was significantly less likely to have received CSI (8.5%) than the latter (40%; P = .0007).
Only 13 patients were 18 years old at last follow-up (range, 18.2 to 20.7 years). For the nine with available information, one was living independently, one was living in a group home, and seven were living at home with their parents. None was married, and only one had part time employment. Abnormalities in speech, ambulation, and dexterity were noted in 51%, 40%, and 34%, respectively, of the entire cohort, at last follow-up.
Endocrinopathies
Patients who received RT had a significantly greater likelihood of developing endocrinopathy than those who did not (P < .0001); 90% of patients who received CSI had an endocrinopathy compared with 68% and 26% of those who received local RT and no RT, respectively. There was a significantly higher incidence of growth hormone deficiency (P < .0001), adrenocorticotropic hormone deficiency (P = .0042), and hypothyroidism (P < .0001) in patients who received CSI compared with those who received local RT or no RT.
Effects of RT on Neurocognitive Outcome
Cross-sectional IQ analyses. Among 67 patients who had at least one IQ exam 5 years after diagnosis, we estimated the frequency of severe mental retardation (ie, IQ < 70) based on the type of RT received (Table 4). Among patients who received CSI, 71.4% had severe mental retardation, compared with 23% of those who received at most local RT and 20% of patients who received no RT (P = .021).
Longitudinal IQ analysis. Of the 194 patients, 101 had sufficient data to estimate premorbid IQ. The obtained estimate was 100.86 (standard deviation, 7.0), suggesting that the mean IQ is projected to have been in the normal range for age. Of the 194 patients, 126 (65%) had at least one IQ assessment and are included in the longitudinal analysis, examining the effect of tumor location, RT, VP shunt, seizure, hearing loss, and NF1 on IQ over time. The mean IQ at the time of first assessment (median, 1.1 years) was 82 (range, 46 to 130). The median number of IQ assessments was four (range, one to seven). Among those with an IQ assessment within 6 months of diagnosis (n = 44), the mean initial IQ was 84 (range, 50 to 120). Table 5 summarizes the clinical characteristics of the patients included in the longitudinal IQ analysis.
The subgroup of 126 patients with at least one IQ assessment was compared with the cohort of 194 patients using the exact 2 test. Of all the demographic and clinical variables, only seizure disorder was distributed significantly differently, with 25% of those with IQ assessments having seizures compared with 44% of those without (P = .007).
Overall, there was a mean IQ decline of 每0.29 points per year in all patients. The IQ increase of 0.91 points per year in patients who received no RT was significantly different from the IQ decline in patients who underwent CSI (每1.34 points per year; P = .005; Fig 2A) or local RT (每0.51 points per year; P = .036), respectively. Although patients who received CSI (每1.34 points per year) seemed to have a worse neurocognitive outcome than those who received local RT (每0.51 points per year), the difference was not statistically significance (P = .26). Recipients of more than 33 Gy CSI experienced a greater IQ decline (每1.7 points per year) than those receiving 24 Gy CSI (每0.63 points per year). However, the numbers of patients in the two categories were too small to report statistical significances.
The IQ decline of patients who received CSI as planned consolidative therapy (每0.31 points per year) was not statistically different from those who received CSI at the time of progression (每1.26 points per year; P = .39).
Among patients with infratentorial tumors, those receiving CSI had a significant decline in IQ (每1.34 points per year) compared with those who received local RT (0.35 points per year; P = .032) and those who received no RT (1.28 points per year; P = .009). Although patients undergoing local RT or no RT tended to have a slight gain in IQ points per year, the rate of change did not differ significantly (P = .34) between these two groups. There was no statistically significant difference in the rate of IQ decline for those who received conformal RT compared with parallel opposed fields RT (P = .71). However, the latter results should be interpreted with caution because of the small sample size.
To assess the effect of age at RT on neurocognitive outcome, we reanalyzed the data to determine rates of IQ decline in patients who underwent CSI at less than 3 years or at 3 years. The IQ decline (每1.97 points per year) for the younger group was not significantly different from the decline of 0.72 IQ points per year for the latter group (P = .19). Because current infant protocols include patients younger than 3 years at presentation, we reanalyzed the data to include only patients younger than 3 years. All model estimates were statistically similar to those obtained for the whole cohort.
Effects of Other Patient Variables on Neurocognitive Outcome
Patients with hemispheric tumors had a significantly greater IQ decline (每1.52 points per year) than those with midline tumors (0.59 points per year; P = .038; Fig 2B). Although patients with VP shunts had lower initial IQ scores, shunting did not significantly affect the rate of IQ decline (P = .69; Fig 2C). Although patients with chronic seizures had a greater decline in IQ scores (每1.02 points per year) than those without seizures (0.004 points per year), the effect was not statistically significant (P = .123; Fig 2D). Severe ototoxicity did not significantly affect the IQ scores over time (P = .58; Fig 1F). Finally, patients with NF1 had an increase in IQ scores (1.7 points per year) compared with those without (每0.5 IQ points per year; P = .024). This may reflect the fact that only 7% of NF1 patients received CSI compared with 30% of those without NF1 (Fig 2E).
Effects of Tumor Location and RT on Neurocognitive Outcome
Only RT and tumor location had a significant effect (P < .05) on neurocognitive outcome independently. We further investigated the combination of RT and tumor location on the change in neurocognitive outcome. To increase the number of patients in each category, we combined CSI and local RT, resulting in no RT/midline (n = 17), no RT/infratentorial (n = 16), no RT/hemispheric (n = 14), RT/midline (n = 13), RT/infratentorial (n = 56), and RT/hemispheric (n = 10). In the multivariable analysis, patients with midline or infratentorial tumors who did not receive RT had an increase in IQ of 3.59 or 1.39 points per year, respectively, which was significantly different from the IQ declines for the other four subgroups (maximum P value is .03). In contrast, patients with hemispheric tumors who received RT suffered the most with an IQ decline of 2.81 points per year, followed by patients with midline tumors who had RT with a decline of 1.66 points per year.
DISCUSSION
This study demonstrates that, in young children with intracranial CNS tumors, treatment with CSI and hemispheric location are associated with significant declines in IQ scores. Patients undergoing CSI are more likely to develop endocrinopathies (P < .0001) and require special education (P = .0007). Among patients with infratentorial tumors, patients undergoing local RT (whether conformal or parallel opposed fields) had a slight gain in IQ that did not differ significantly from the gain in the no-RT group (P = .34).
Functional Outcome
The recent study by Packer et al26 of 1,607 patients in the Childhood Cancer Survivor Study is the only comprehensive review of the neurologic and neurosensory sequelae of children with brain tumors. Relative to sibling controls, survivors of brain tumors were at elevated risk for hearing and visual impairment. Coordination and motor problems were reported in 49% and 26%, respectively, seizure disorders in 25%, and hearing impairment in 12%. Although these data were self-reported and included patients up to 20 years old, the results are similar to those of the current cohort.
The relative effect of therapy on the development of endocrinopathies in children with CNS malignancies has been reported previously, with various reports confirming or questioning the role of RT in the development of endocrinopathies.10,27-31 In the present study, patients who received RT had a significantly greater likelihood of developing endocrinopathy than those who did not (P < .0001); 90% of those who received CSI had an endocrinopathy compared with 68% and 26% of those who received local RT and no RT, respectively.
The current report offers a comprehensive review of social attainment in this group of patients. The 36% of patients who required special education were significantly more likely to have received CSI than those in regular school. Only one of the 13 surviving patients older than 18 years was working independently, and none was married. Likewise, Kiltie et al11 reported that of 16 medulloblastoma patients alive 10 years after treatment, 15 required special education, only one was working independently, and none was married. Hoppe-Hirsch et al8 reported that, 10 years after therapy, 71% of patients with ependymoma (median age at presentation, 5.8 years) attended regular school compared with 20% of medulloblastoma patients (median age at presentation, 5.8 years). Zebrack et al32 recently reviewed the psychological outcomes in 1,101 long-term survivors of childhood CNS tumors younger than 18 years at diagnosis and reported that 74% were currently employed, 63% had graduated from high school, and 74% remained single; they concluded that diminished social functioning contributed to increased psychological distress.
Factors Affecting Neurocognitive Outcome
Previous studies have reported progressive neurocognitive decline after RT.4,6,9-12,14 A study of 19 infants with medulloblastoma from this institution treated with chemotherapy and CSI at progression demonstrated a decline of 3.9 IQ points per year,10 which is similar to the decline of the current larger cohort. The IQ decline for patients who received CSI or local RT only was significantly different from the IQ gain in the no RT group. The likelihood of having an IQ less than 70 at least 5 years after diagnosis was much higher in patients with CSI compared with those who received local or no RT. These trends are similar to those reported by Grille et al,7 in which an older group of patients with posterior fossa tumors receiving 0, 25, and 35 Gy had mean IQ scores of 84.5, 76.9, and 63.7, respectively. Similarly, Hoppe-Hirsch et al8 reported that in older children, 56% of patients with ependymoma (who received local RT only) maintained an IQ of 90 5 years after therapy compared with 22% of patients with medulloblastoma who received CSI.8 In our cohort, patients who received cranial doses 33 Gy also had a greater IQ decline than those receiving 24 Gy, but the numbers were too small to assess statistical significance. Consistent evidence of neurocognitive decline with CSI in children of different ages with brain tumors has also been reported by others.3,5,9,11,13
For infratentorial tumors, there was no significant IQ decline for those treated only with local irradiation, whether parallel opposed fields or conformal RT. The latter finding is in keeping with a report from this institution in which patients with ependymoma receiving conformal RT demonstrated no statistically significant change in IQ 24 months after completion of therapy.14
VP shunting did not have a significant impact on the rate of IQ decline, although patients with VP shunts tended to have lower initial IQs than those without. The relative effects of the tumor, surgery, and perioperative factors compared with treatment effects have always been difficult to delineate.9,33-35 In the current cohort, patients who had neurocognitive assessments done within 6 months of diagnosis demonstrated a lower baseline IQ than expected, confirming that the presence of tumor, surgery, and perioperative factors do have an effect on neurocognitive outcome regardless of subsequent treatment.
Although this study provides the most comprehensive overview of functional, social, and neurocognitive outcomes in a large cohort of infants with a long follow-up period, it has several limitations. Neurocognitive data were available in only 65% of the cohort. Our analysis demonstrated the relative similarity of the cohorts with and without neurocognitive follow-up. However, patients who were followed or had surgery only were less likely to undergo neurocognitive testing (54.9%) than those who received chemotherapy or RT (68.5%), probably because of the lower risk for neurocognitive sequelae. Similarly, there may have been a bias not to test patients who were more severely affected. Of those with neurocognitive data, a subset was included in an estimation of premorbid functioning using a prediction formula previously published.23 This constitutes an initial attempt to use such an inferential model. However, the model was developed for estimating full-scale IQ scores comparable with subsequent scores derived from the Wechsler Intelligence Scale for Children每Third Edition, recommended for children aged 6 through 16 years 11 months.18 In addition, the coding for ethnic classifications used to determine the model are restrictive (black/non-black, white/non-white) and may not be suitable for highly diverse populations. Extrapolating outside the data on which the model was developed may not be appropriate. Among patients with endocrinopathy, the endocrinopathies were coded as present or absent without regard to time of onset, thus making any conclusions regarding the relationship between endocrinopathy and type of therapy difficult to assess. Finally, the diversity of histology and tumor location in this cohort has implications for choice of therapy and, hence, long-term outcome in this particularly vulnerable group. It is imperative that future studies include prospective and comprehensive neurocognitive and functional assessments to accurately evaluate the impact of novel therapeutic approaches on long-term functional outcome.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
Acknowledgment
We thank Dr James Boyett for his input regarding the statistical analysis and Patsy Burnside for typing the manuscript.
NOTES
Supported by Center of Research Excellence support Grant No. CA21765 and by American Lebanese Syrian Associated Charities.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Palmer SL, Goloubeva O, Reddick WE, et al: Patterns of intellectual development among survivors of pediatric medulloblastoma: A longitudinal analysis. J Clin Oncol 19:2302-2308, 2001
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 19:472-479, 2001
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 19:3470-3476, 2001
Mulhern RK, Reddick WE, Palmer SL, et al: Neurocognitive deficits in medulloblastoma survivors and white matter loss. Ann Neurol 46:834-841, 1999
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 Pediatric Oncology Group study. J Clin Oncol 16:1723-1728, 1998
Dennis M, Spiegler BJ, Hetherington CR, et al: Neuropsychological sequelae of the treatment of children with medulloblastoma. J Neurooncol 29:91-101, 1996
Grill J, Renaux VK, Bulteau C, et al: Long-term intellectual outcome in children with posterior fossa tumors according to radiation doses and volumes. Int J Radiat Oncol Biol Phys 45:137-145, 1999
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 Nerv Syst 11:340-345, 1995
Copeland DR, deMoor C, Moore BD III, et al: Neurocognitive development of children after a cerebellar tumor in infancy: A longitudinal study. J Clin Oncol 17:3476-3486, 1999
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 17:3720-3728, 1999
Kiltie AE, Lashford LS, Gattamaneni HR: Survival and late effects in medulloblastoma patients treated with craniospinal irradiation under three years old. Med Pediatr Oncol 28:348-354, 1997
Chapman CA, Waber DP, Bernstein JH, et al: Neurobehavioral and neurologic outcome in long-term survivors of posterior fossa brain tumors: Role of age and perioperative factors. J Child Neurol 10:209-212, 1995
Spiegler BJ, Bouffet E, Greenberg ML, et al: Change in neurocognitive functioning after treatment with cranial radiation in childhood. J Clin Oncol 22:706-713, 2004
Merchant TE, Mulhern RK, Krasin MJ, et al: Preliminary results from a phase II trial of conformal radiation therapy and evaluation of radiation-related CNS effects for pediatric patients with localized ependymoma. J Clin Oncol 22:3156-3162, 2004
Bayley N: Bayley Scales of Infant Development (ed 2). New York, NY, The Psychological Corporation, 1993
Bayley N: Bayley Scales of Infant Development (ed 2). New York, NY, The Psychological Corporation, 1969
Wechsler D: Wechsler Adult Intelligence Scale〞Revised. New York, NY, The Psychological Corporation, 1989
Wechsler D: Wechsler Intelligence Scale for Children (ed 3). New York, NY, The Psychological Corporation, 1991
Wechsler D: Manual for the Weschler Intelligence Scale for Children〞Revised. San Antonio, TX, The Psychological Corporation, 1974
Wechsler D: Wechsler Preschool and Primary School Scales of Intelligence〞Revised. New York, NY, The Psychological Corporation, 1989
Wechsler D: Wechsler Preschool and Primary School Scales of Intelligence〞Revised. New York, NY, The Psychological Corporation, 1963
Sattler JM. Assessment of children. San Diego, CA, JM Sattler, 1992, pp 1170
Vanderploeg RD, Schinka JA, Baum KM, et al: WISC-III premorbid prediction strategies: Demographic and best performance approaches. Psychol Assess 10:277-284, 1998
Hardin J, Hilbe J: Generalized Estimating Equations. Boca Raton, FL, Chapman & Hall/CRC, 2003, pp 46-47
SAS Institute Inc: Longitudinal Data Analysis with Discrete and Continuous Responses; course notes. Cary, NC, SAS Institute Inc, 2004, pp 77-99
Packer RJ, Gurney JG, Punyko JA, et al: Long-term neurologic and neurosensory sequelae in adult survivors of a childhood brain tumor: Childhood cancer survivor study. J Clin Oncol 21:3255-3261, 2003
Constine LS, Woolf PD, Cann D, et al: Hypothalamic-pituitary dysfunction after radiation for brain tumors. N Engl J Med 328:87-94, 1993
Duffner PK, Cohen ME, Thomas PR, et al: The long-term effects of cranial irradiation on the central nervous system. Cancer 56:1841-1846, 1985
Kanev PM, Lefebvre JF, Mauseth RS, et al: Growth hormone deficiency following radiation therapy of primary brain tumors in children. J Neurosurg 74:743-748, 1991
Collet-Solberg PF, Sernyak H, Satin-Smith M, et al: Endocrine outcome in long-term survivors of low-grade hypothalamic/chiasmatic glioma. Clin Endocrinol (Oxf) 47:79-85, 1997
Merchant TE, Williams T, Smith JM, et al: Preirradiation endocrinopathies in pediatric brain tumor patients determined by dynamic tests of endocrine function. Int J Radiat Oncol Biol Phys 54:45-50, 2002
Zebrack BJ, Gurney JG, Oeffinger K, et al: Psychological outcomes in long-term survivors of childhood brain cancer: A report from the childhood cancer survivor study. J Clin Oncol 22:999-1006, 2004
Ater JL, Moore BD III, Francis DJ, et al: Correlation of medical and neurosurgical events with neuropsychological status in children at diagnosis of astrocytoma: Utilization of a neurological severity score. J Child Neurol 11:462-469, 1996
Chin HW, Maruyama Y: Age at treatment and long-term performance results in medulloblastoma. Cancer 53:1952-1958, 1984
Mulhern RK, Horowitz ME, Kovnar EH, et al: Neurodevelopmental status of infants and young children treated for brain tumors with preirradiation chemotherapy. J Clin Oncol 7:1660-1666, 1989(Maryam Fouladi, Elizabeth)