Second cancer after radiotherapy, 1981-2007.

Abstract

Today, there is growing concern about radiotherapy induced secondary malignancies. We analysed the incidence and dose dependence of second cancer. The study includes 12,000 one-year survivors of radiotherapy, treated between 1981 and 2007. For risk estimates a public databank on cancer in Germany served as reference. Contralateral second breast cancer, second oesophageal and colorectal cancer were analysed with retrospective dosimetry. GI-tract data were used for risk modelling. The incidence rate of second cancers (493 cases) was ~1% per year. Contralateral breast cancer was the most frequent entity (relative risk RR=2.8). The scatter-dose gradient (2-3 Gy) across the contralateral breast did not cause a detectable risk gradient. There was an increased risk for second head and neck cancer (RR=5.1) and for male oesophageal cancer (RR=5.8). For both entities, dose response modelling with case-control data predicted maximum curves with peak induction at 1-5 Gy and positive excess absolute risk values at high doses. A survey of second cancer after radiotherapy requires follow-up over decades. Preliminary dose response modelling albeit with low case numbers suggests an increased risk from multiportal techniques. To improve risk assessment, prospective out-of-field dosimetry and long-term multicentre data collection are recommended.