An analysis of SEER data of increasing risk of secondary malignant neoplasms among long-term survivors of childhood brain tumors.

Background: Advances made in treatment of a childhood brain cancer have extended the lives of many children and adolescents. Treatment success, however, brings the opportunity to assess late effects; most worrying among these are secondary malignant neoplasms (SMN). Even though the cumulative incidence is quite small, long-term follow-up is required because treatment-induced cancers can occur years after initial treatment.

Procedure: The purpose of this project was to determine what treatments and what host characteristics of children treated for a primary brain cancer are associated with an increase in the risk of a SMN in long-term survivors. Data were analyzed from 2,056 5-year survivors, of primary brain cancer in the surveillance, epidemiology, and end results (SEER) database between 1973 and 1998. Thirty-nine patients developed a SMN. Cox regression models were used to evaluate the independent contribution of a number of risk factors.

Results: The most important risk factor for developing a SMN in 5-year survivors was the era in which the primary cancer was treated. Compared to treatment prior to 1979, patients treated between 1979 and 1984 had a 4.7-fold increase in risk (P = 0.001), while those treated after 1985 had a 6.7-fold increase in risk. (P = 0.002). Patients treated most recently carry the greatest risk of SMN development even after controlling for radiotherapy. This could be due to the increase in intensive treatment compared to earlier years.

Conclusion: Although the absolute excess risk of SMN remains quite low, continued surveillance is needed to evaluate long-term effects of new therapies for primary brain tumors.