paediatrics Brussels 17

Redmond et al.: Radiation to neural progenitor niches and neurocognitive outcomes

cancer patients, which may have limited their ability to present for all scheduled appointments. Additionally, some of the youngest patients were not tested until the later time points. Our analyses presume that the existing data are reflective of the entire group, but future studies with potentially more complete datasets will be impor- tant to provide confirmation. In addition, while our data suggest that limiting radi- ation dose to the temporal lobe and hippocampus is im- portant in reducing the neurocognitive sequelae of radiation to the brain, we do not have sufficient data to determine a safe radiation tolerance of these struc- tures. In order to effectively spare these areas using tech- niques such as intensity-modulated RT, it will be critical to better define the radiation tolerance and dose response of these structures. Similarly, further evaluation for a dose-volume effect in these structures will be important. Finally, we report results on a limited number of neu- rocognitive tests, and our analysis includes radiation dose to only a limited number of brain structures. Analysis of a broader spectrum of neurocognitive func- tions and CNS anatomy in future studies will be impor- tant. It is possible that the associations we report are confounded by radiation dose to adjacent structures that were not included in this analysis. To conclude, we report a significant relationship between radiation dose to the hippocampus and tempo- ral lobes and performance on select neurocognitive tests but do not find a relationship between RT dose to the SVZ and neurocognitive function. To our knowledge, this study is one of the first prospective studies to date to examine the relationship between radiation dose to NPC-containing niches and neurocognitive function. Our results have important implications for ongoing clinical trials such as RTOG 0933, which is a phase II trial of hippocampal avoidance during whole brain RT for brain metastases and has a primary objective of per- formance on a test of verbal learning and memory.

in performance on the test of motor speed between patients with infratentorial versus supratentorial tumors at any time point. There was a significant relationship between radia- tion dose to the left temporal lobe and performance on the visual perception test. However, the effect was very small and may not be clinically significant. There was no difference in performance on this test between left- and right-handed individuals at any time point. The absence of a significant association between radiation dose to the hippocampi / temporal lobes and performance on tests of verbal learning and memory, visuospatial working memory, and vocabulary may be due to low statistical power in detecting a difference, since the motor function and spatial perception tests are dependent on a complex interplay among multiple neural pathways and may be more sensitive than the other tests examined in this study. Second, given the relatively small number of patients enrolled in this study, we were not able to perform a detailed analysis of all potential variables that might contribute to cognitive outcomes. For example, con- founding variables such as tumor location, recurrence patterns, and surgical interventions may have contribut- ed to the changes in neurocognitive function that we report. Our data suggest potential recovery in function over time from tumor and surgery-related intervention. The relationships among other disease- or treatment- related factors beyond radiation dose to the structures evaluated remain unclear. Future studies enrolling a larger and potentially more homogeneous patient popu- lation would be helpful in further evaluating this interac- tion. Similarly, longer-term follow-up will be important in ascertaining whether the changes reported in this study were impacted by acute effects associated with the disease, surgery, and adjuvant therapy rather than strictly cognitive late effects of radiation. In addition, although the eligibility criteria for study entry were comparable for both the patient and control groups, there were no children age 0–4 years in the control group, whereas this age group comprised 21% of the patient group. Because we used standard- ized, age-adjusted scores for our analysis of the neuro- cognitive tests, we would not expect this difference to have an impact upon our results. Similarly, there were a larger percentage of African Americans in the control group than in the patient group, but we are not aware of evidence to suggest that this would impact upon test performance or interpretation of results. Third, our series contains missing data, in that not all patients were evaluated on all tests at all time points, and missing patients varied among time points. This limita- tion may be in part related to the poor prognosis and severity of illness in our patients as well as the complex social situations of families of pediatric

at Universitaet Leipzig, Institut fuer Informatik/URZ, Bibliothek on August 25, 2014 http://neuro-oncology.oxfordjournals.org/ Downloaded from

Conflict of interest statement . None declared.

Funding

This work was supported by NIH / NINDS grant no. R01 NS04285; Intellectual and Developmental Disabi- lities Research Center-NIH / NICHD grant no. P30H- D024061-16; the Intellectual and Developmental Disabilities Research Center, grant no. HD-24061; Jo- hns Hopkins University School of Medicine Institute f- or Clinical and Translational Research; and the NIH / National Center for Research Resources Clinical and Translational Sciences Awards Program, grant no. UL1-RR025005.

NEURO-ONCOLOGY † M A R C H 2 0 1 3

367