Neurocognitive outcomes after proton radiation for pediatric brain tumors

Reporter:  Arpi Thukral, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 6, 2010

Presenter: M. B. Pulsifer, Massachusetts General Hospital, Boston, MA; Children's Hospital Medical Center, Boston, MA; Harvard Medical School, Boston, MA


  • Pediatric brain tumors account for 20% of all pediatric cancers, and treatment generally involves surgery, chemotherapy, and/or radiation therapy.
  • Radiation therapy, which is delivered with photons at most centers in the U.S., has been demonstrated to be associated with long-term neurocognitive sequelae, including decrements in IQ and difficulties with attention, processing speed, and other executive skills.
  • Progressive IQ decline has been seen following radiation treatment, and the greatest decline has been in patients <7 years old.
  • It has been shown that radiation with protons allows for greater conformality and sparing of surrounding normal tissues. This is due to its characteristic Bragg Peak, which allows maximal deposition of dose in tumor, without an exit dose.
  • For this reason, one can expect that neurocognitive deficits resulting from radiation to the brain would be lower with protons as compared to photons.
  • The current study was performed to examining long-term neurocognitive outcomes of 56 patients with various brain tumors treated with proton radiation at MGH.
    • The questions to be answered in this study included:
      1. What is the neurocognitive profile of pediatric brain tumor patients before and after radiation therapy?
      2. Are outcomes in this population better with proton irradiation than would be expected after photon irradiation?


  • This is a longitudinal study that examined the neurocognitive and academic function in184 patients with brain tumors who received proton radiation using a multitude of standardized assessment tools with age-based scores.
  • Eligibility criteria:
    • Age < 21 years
    • Tumor location in brain with histologic diagnosis centrally reviewed at MGH
    • Treatment with proton XRT only; no photon XRT
  • All patients had surgery of the primary tumor followed by proton radiotherapy with or without concurrent chemotherapy and adjuvant platinum-based chemotherapy.
    • If patients required craniospinal irradiation (CSI), doses ranged from 18-36 Gray-Equivalents (GyE), with median dose 23.4 GyE. The tumor bed or posterior fossa was boosted to 54 GyE.
  • Neurocognitive areas that were examined included, but were not limited to:
    • IQ, language, visual spatial/motor effects, memory, and processing speed.
  • Patients had baseline (BL) neuropsychological testing starting in 9/02. Of these, 56 patients received follow-up testing two years from BL (M=2.12, SD=1.30).


  • A total of 56 patients receiving proton radiation for treatment of brain tumors were enrolled on this study. Type of cancer included: medulloblastoma (50%), craniopharyngioma (16%), ependymoma (13%), and other (21%) tumors.
  • Demographics: 30 males (54%) and 26 females (46%). Mean age was 8.23 years (SD=4.47) at BL (range: 1 to 19). 82% of patients were right-handed.
  • 68% had complete tumor resection; 36% had treatment complications, and 9% had posterior fossa syndrome.
  • 31 patients had CSI and 25 patients did not have CSI. Mean total radiation dose was 52.77 GyE (SD=4.14) ,and 73% received chemotherapy as part of treatment.
  • Patients with follow-up testing did not differ in age at baseline or gender from those without follow-up testing.
  • 5% of patients had recurrences at 5 year follow up.
  • No significant decline in the whole cohort was seen between baseline and follow-up testing for most areas, including IQ, receptive/expressive vocabularies, working memory, manual dexterity, attention, reading, and arithmetic, and memory.
  • However, significant decline was seen in scores for processing speed, graphicomotor skills, and visual organization (p < 0.05 for change from baseline to follow-up).
    • Age, gender, location of tumor, and extent of radiation (CSI or no CSI) did not significantly impact the presence or degree of this decline.
  • Patients who received CSI did demonstrate a higher rate of decline in IQ from BL at follow-up testing that those who did not receive CSI.
    • In terms of histology, a larger IQ decline was seen for patients with medulloblastoma compared to those with other brain tumors, likely because medulloblastoma patients received CSI.

Authors' Conclusions

  • The authors concluded that at 2-year follow-up after proton radiation for brain tumors, they did not see significant changes in most of the neurocognitive areas which were assessed. Neurocognitive performances were overall stable from BL testing to follow-up testing.
  • The only declines seen were in processing speed, graphicomotor skills, and visual organization.
  • They note that these results compare favorably to previous reports from photon radiation treatment in pediatric brain tumor patients.
  • Partial brain radiation resulted in sparing of IQ decline when compared to whole brain irradiation with CSI, likely due to smaller treatment fields.

Clinical Implications

  • Neurocognitive decline is a toxicity that is associated with radiation treatment in pediatric brain tumor patients, and which may have lifelong repercussions.
  • There is currently very little clinical evidence showing that protons are superior to photons in the pediatric population.
  • The present study is a well-designed and thorough longitudinal analysis of various neurocognitive side effects in children treated with proton radiation for brain tumors, and used standardized age-based assessment tools for measurement.
  • The results suggest that the greater conformality that is possible with proton therapy compared with IMRT may decrease neurocognitive decline, without decreasing efficacy of treatment.
  • This data has important implications for patients:
    • If neurocognitive toxicity is, in fact, decreased with the use of proton therapy, this could lead to better QOL for these patients. Neurocognitive sparing would also decrease societal costs associated with radiation of the pediatric brain. Finally, it might allow escalation of doses of chemotherapy or radiation, which could potentially improve outcomes.
    • If areas of neurocognitive decline (such as decline in processing speed) are identified early, enrichment tools or occupational therapy may be initiated promptly to deter these effects.
  • The significant difference in decline in IQ between the CSI and non-CSI group suggests that IQ changes may be related to extent of radiation fields. This argument would also favor the use of protons in these patients as opposed to 3D conformal photon treatment or IMRT, which would increase total dose to the brain due to exit dose.
  • Limitations of this study:
    • No valid control groups (IMRT or 3D conformal photon) were presented. However, it is difficult to perform randomized studies of photons vs. protons in these patients due to ethical considerations and accrual issues.
    • This cohort may be a selected subset of patients who may have greater access to resources to help delay neurocognitive decline.
    • There is a need for longer follow-up to better assess late and long-term outcomes.
  • In conclusion, this study provides important information to add to the growing, but scant body of literature on clinical outcomes in proton patients. We need more well-designed studies, such as this one, to further investigate the magnitude of clinical benefit with the use of proton therapy in children.