Second neoplasms after megavoltage radiation for pediatric tumors

BACKGROUND. Because ionizing radiation is a known carcinogen, diligent longterm follow-up in children exposed to therapeutic radiation is required. The authors updated an earlier study on the nature and risk of second neoplasms (SN) in patients treated with megavoltage radiotherapy as children. METHODS. The authors followed 446 children who were treated for a primary malignancy with megavoltage radiotherapy between January 16, 1954 and December 31, 1980. These children survived a minimum of 5 years. The nature and incidence of SNs were evaluated in this population. Patients with bilateral retinoblastoma or neurofibromatosis were excluded from the study due to their large genetically based predisposition for developing an SN. RESULTS. The Kaplan-Meier actuarial survival rate was 80% (95% confidence interval [CI] 74-85%) at 30 years for all patients. Thirty-seven (8.3%) patients developed SNs, most of which occurred within the original radiation treatment field, 3.8-31.8 years (median, 15.5 years) after radiotherapy. The cumulative risk of developing an SN was 13% (95% CI 9-19%) at 30 years and the standardized incidence ratio for the development of any SN was 5.2 (95% CI 3.4-7.6%). The most common SNs were breast carcinoma (n = 8), skin carcinoma (five basal cell carcinomas, two malignant melanomas, and one dermatosarcoma), and meningiomas (n = 6). All eight breast carcinomas occurred after the treatment of childhood Hodgkin disease. Of the 37 patients with SNs, 12 died of either the SN (n = 10) or of recurrent disease (n = 2). Risk factors associated with developing a SN included initial diagnosis of Hodgkin disease (P = 0.0003), female gender (P = 0.008), and an initial diagnosis of acute lymphoblastic leukemia (P = 0.02). CONCLUSIONS. Patients in the radiation-treated cohort experienced increased mortality, were at an increased risk of developing an SN, and should undergo increased medical surveillance as adults. The cumulative probability of developing an SN has increased substantially at 30 years, largely due to an increase in follow-up time. In addition, the cumulative probability curve does not show evidence of plateau after increased duration of follow-up. Finally, the emergence of secondary breast carcinoma in the current study was not noted in the previous analysis.