Second primary cancer after radiotherapy for prostate cancer--a seer analysis of brachytherapy versus external beam radiotherapy.

PURPOSE To determine the incidence of second primary cancers (SPCs) and radiotherapy-induced SPCs (RTSPCs). PATIENTS AND METHODS The incidence of SPCs and RTSPCs was compared among four treatment groups with locoregional prostate adenocarcinoma in the 1973-2002 Surveillance, Epidemiology, and End Results database. These groups were no radiotherapy (RT), no surgery (Group 1); external beam RT (EBRT) (Group 2); brachytherapy (Group 3); and a combination of EBRT and brachytherapy (Group 4). RESULTS The age-adjusted estimates of SPCs were greater with EBRT than with brachytherapy (2,178 vs. 1,901 SPCs/100,000; p = 0.025) or with the no RT, no surgery group (1,971 SPCs/100,000; p <0.0001). The age-adjusted rate of late SPC (>or=5 years) for EBRT (2,425 SPCs/100,000) was only significantly greater (p <0.0001) than that for no RT, no surgery (1,950 SPCs/100,000). The hazard ratio adjusted for age, race/ethnicity, and grade was constant at 1.263 for EBRT compared with no RT, no surgery (p <0.0001) but varied with the length of follow-up in both the brachytherapy (0.721 at 5 years to 1.200 at 9 years) and combination (0.920 at 5 years to 1.317 at 9 years) groups. The incidence of RTSPCs was only significantly different between the no RT, no surgery group and the EBRT group, with an increase of 162 cases/100,000 or a 0.16% increased SPC risk (p = 0.023). No significant differences in the incidence of RTSPC were seen between the RT groups. CONCLUSION No significant differences were seen in the incidence of RTSPCs between the RT groups. The initial smaller relative risk of overall SPCs in the brachytherapy group increased with time until the curves converged, suggesting that the effect had resulted from patient selection bias.