Incidence of Second Malignancies Among Patients Treated With Proton Versus Photon Radiation

Abstract

Proton radiation, when compared with photon radiation, allows delivery of increased radiation dose to the tumor while decreasing dose to adjacent critical structures. Given the recent expansion of proton facilities in the United States, the long-term sequelae of proton therapy should be carefully assessed. The objective of this study was to compare the incidence of second cancers in patients treated with proton radiation with a population-based cohort of matched patients treated with photon radiation. We performed a retrospective cohort study of 558 patients treated with proton radiation from 1973 to 2001 at the Harvard Cyclotron in Cambridge, MA and 558 matched patients treated with photon therapy in the Surveillance, Epidemiology, and End Results (SEER) Program cancer registry. Patients were matched by age at radiation treatment, sex, year of treatment, cancer histology, and site. The main outcome measure was the incidence of second malignancies after radiation. We matched 558 proton patients with 558 photon patients from the Surveillance, Epidemiology, and End Results registry. The median duration of follow-up was 6.7 years (interquartile range, 7.4) and 6.0 years (interquartile range, 9.3) in the proton and photon cohorts, respectively. The median age at treatment was 59 years in each cohort. Second malignancies occurred in 29 proton patients (5.2%) and 42 photon patients (7.5%). After we adjusted for sex, age at treatment, primary site, and year of diagnosis, proton therapy was not associated with an increased risk of second malignancy (adjusted hazard ratio, 0.52 [95% confidence interval, 0.32-0.85]; P=.009). The use of proton radiation therapy was not associated with a significantly increased risk of secondary malignancies compared with photon therapy. Longer follow-up of these patients is needed to determine if there is a significant decrease in second malignancies. Given the limitations of the study, these results should be viewed as hypothesis generating.