paediatrics Brussels 17

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

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

Fig. 5. Performance on Purdue Pegboard 2-Hand test ( Z -scores) at 6 mo following completion of RT relative to (A) mean left temporal lobe radiation dose, P ¼ .033, and (B) mean right temporal lobe radiation dose, P ¼ .015. Standardized scores were used in this analysis to account for the impact of age on test performance.

the left or right hippocampus (dose, P ≥ .2), cerebrum ( P ¼ .22), or SVZ ( P ≥ .21) and performance on the vocabulary test. Visuospatial working memory (Bead Memory).— There was no relationship between mean radiation dose to the left or right hippocampus ( P ≥ .66), temporal lobes ( P ≥ 0.87), SVZ ( P ≥ 0.18), or cerebrum ( P ≥ .25) and performance on the spatial working memory test. We present a prospective study examining the relation- ship between radiation dose to NPC-containing niches, temporal lobes, and cerebrum and neurocognitive func- tion following RT. We demonstrate a significant associ- ation between increasing mean RT dose to the hippocampus and temporal lobes and decline in select neurocognitive skills following cranial irradiation, but no association between mean dose to SVZ or cerebrum and test performance. This is one of the first human studies to corroborate animal data suggesting a relation- ship between radiation-induced damage to the hippo- campus and neurocognitive dysfunction. 8 – 14 These prospective data are consistent with prior ret- rospective studies on long-term cancer survivors that have demonstrated a significant association between ra- diation dose to the temporal lobes and neurocognitive dysfunction. 25 , 39 A prior prospective study found that patients receiving . 43.2 Gy to 13% of the volume of the left temporal lobe were significantly more likely to demonstrate a . 10% decline in performance in full- scale IQ. 24 An analysis of patients with nasopharyngeal carcinoma reported significantly lower cognitive func- tioning scores in patients with a mean dose to the tempo- ral lobes of . 36 Gy. 23 Changes in the development of neuropsychological skills (especially motor speed, declarative memory, and visuoperceptual skills) following cranial irradiation may result from radiation-induced structural damage to the brain. For example, Nagel et al 40 demonstrated Discussion

that the volume of both the right and left hippocampi de- creased during the first 2–3 years following craniospinal radiation for medulloblastoma. Diffusion tensor MRI has been used to detect increased diffusion of water mol- ecules in the hippocampus of patients receiving RT. 41 Similarly, changes in white matter integrity 42 and volume 43 occur following cranial irradiation, and reduced volumes of the cerebellar vermis predict reduced performance on neuropsychological testing. 44 Future studies will be critical in gaining additional insight into the mechanism of radiation-induced struc- tural changes in the brain and its relationship to neuro- cognitive dysfunction. Our study did not demonstrate a relationship between RT dose to the SVZ and neurocognitive func- tion, a finding that is consistent with a prior study. 31 Improved neurocognitive performance has also been re- ported in patients with central nervous system germ cell tumors treated with whole ventricle irradiation com- pared with craniospinal RT, 45 suggesting that the most critical areas for neurocognitive dysfunction are likely to reside outside of the SVZ. There are several limitations to our study. First, al- though hippocampal and temporal lobe functions are classically associated with declarative memory and learning, the most significant relationship between radi- ation dose to these areas and reduced performance was on the test of motor speed, which may reflect the sensi- tivity of tests of motor function to neural dysfunction outside the frontally mediated motor systems. These changes to motor function may reflect earlier effects on more widespread subcortical white matter pathways in- volved in the development of motor speed. In addition, the test of motor function is the only timed measure in- cluded in this battery, and the results may therefore reflect a difference in processing speed. Performance on the test of motor speed may also have been impacted by radiation dose to other structures, such as the cerebel- lum, which received a mean dose of 29.5 Gy (range 2–50.4 Gy). Although the relatively large percentage of patients with infratentorial tumors could have con- founded our results, there was no significant difference

NEURO-ONCOLOGY † M A R C H 2 0 1 3

366