The Risk of Developing a Neutron-induced Second Malignancy for Pediatric Proton Therapy Patients

Reviewer: Nathan Jones DO
Abramson Cancer Center of the University of Pennsylvania
Last Modified: September 24, 2008

Presenter: Harald Paganetti
Presenter's Affiliation: Massachusetts General Hospital, Boston, MA
Type of Session: Scientific


  • Second malignancies and other late toxicities are a major concern when treating pediatric cancer patients with radiation therapy. 
  • The physical properties of the proton beam offer the potential for improved normal tissue sparing, as the exit dose is essentially eliminated.
  • High energy protons interact with matter to produce neutron contamination.
  • Neutrons are largely produced by interactions of the beam within the treatment head, including the scattering foil; they are produced to a lesser extent through neutron scatter within the patient.
  • The relative biologic equivalence (RBE) of neutrons is believed to be up to 20 times greater than photons, with some papers suggesting 60-70 times greater
  • With increasing use of proton therapy in pediatric cancer, it is important better understand the extent and potential consequences of neutron contamination
  • BEIR report suggests that “epidemiologic studies should be based on accurate individual dose preferentially to the organ of interest”

Materials and Methods

  • Computer modeling was used to calculate the neutron equivalent dose to specific organs at risk that would result from 80 Gy to the brain through 8 different fields
  • 6 whole body virtual phantoms were developed of various ages and both genders (4 male, 2 female) ranging from 9 months old to adult
  • Virtual body phantoms included body and organ geometry and organ-specific composition by age and gender; for example, skeletal carbon content is known to vary significantly with age and was modeled accordingly
  • Accurate modeling was performed of the entire treatment head in a scattering beam configuration
  • Monte carlo simulations were performed using the modeled treatment head and virtual body phantoms to simulate 8 different proton therapy fields treating 70 Gy to the brain
    • These fields were varied by both aperture size and width of the spread out Bragg peak
  • Neutron RBE varied by energy, with a maximum of 20 used for calculations
  • Specific epidemiological risks to each organ were determined according to the BEIR VII risk models to calculate the lifetime attributable risk (LAR) of developing second malignancy for each organ based on neutron equivalent dose as it varied by field configuration, age and gender


  • Overall risks for developing second malignancy from neutron contamination in proton therapy is not higher, on average, than that obtained by scattered photons from IMRT
  • As a frame of reference, the risk was calculated to some organs (i.e. thyroid) to be equivalent to ~30 CT scans
  • Risks from neutrons generated in the treatment head appear to be roughly 10-fold higher (80-92% of total) than those from neutron production within the patient
  • Larger field size decreased neutron dose to lung, for example, as fewer neutrons were produced by the aperature, despite increased internal neutron production due to field size
    • Target size varied neutron dose by up to a factor of 2
  • Overall, almost all LARs were less than the baseline risk, with a few exceptions, i.e. LAR for thyroid in 8-year-old female is double the baseline risk
  • Most lifetime risks of developing second malignancies attributable to these proton brain fields were <1%
    • Exceptions included breast, thyroid, and lung cancer for females
  • For a 4-year-old female, the LAR for breast cancer is up to 5% (17% baseline)
  • Risks vary by age, with the main risk for an adult male being leukemia
  • Main risks for males were leukemia, thyroid and lung cancers

Author's Conclusions

  • Proton therapy produces a relatively small, but non-trivial risk of second malignancy due to neutron contamination
  • Risk is presumably greatly outweighed by risk associated with current malignancy
  • Neutron production is primarily from the treatment head, which is encouraging as it can be modified with improved engineering, unlike internal neutron production which is inherent to the treatment modality and field arrangement
  • Age at the time of treatment plays a critical role in the lifetime risk of developing a second malignancy
  • Uncertainties in RBE for neutron equivalent doses are much greater than for IMRT or CT imaging
  • Areas for future study include differences in passive versus scanning beam proton therapy, differences between IMRT, and effect when different body sites and field arrangements are used

Clinical/Scientific Implications

  • These data provide reassurance that the risk of second malignancy from neutron contamination in proton therapy to the brain is relatively small
  • It must be emphasized that while this study uses robust methodology in determining the neutron dose to organs (Monte carlo calculations using detailed virtual phantoms and accurate treatment head modeling), the extrapolation of this data to the risk of second malignancy is less reliable, and includes estimation of RBE for the neutrons and how this RBE corresponds to the risk modeling from the BEIR report 
  • This study is limited by the risk modeling as developed for BEIR VII, which is primarily based on late effects of atomic bomb survivors
  • The relevancy of these models to patients with known malignancy at the onset and with varying fractionation and distribution of treatment is unclear
  • As many pediatric patients are treated to the extracranial CNS or other parts of the body with a wide variety of field arrangements, additional analysis is necessary to better understand potential effects of these treatments
  • Additional engineering advancement to lessen neutron dose from the treatment head is justified by these data, potentially including improved shielding and use of a scanning beam rather than passive scattering
  • Taken in context, this is a very useful addition to increasing the understanding of neutron-related risks in patients of various ages receiving radiation therapy to the brain
  • Rigorous longitudinal data collection is imperative to understanding the real-world correlation of these models