Background: The use of computed tomography (CT) surveillance imaging as a part of post-cancer treatment follow-up schedules is increasing rapidly. However quantification of radiation-attributable cancer risk associated with long-term CT surveillance imaging, is lacking.
Methods: A simulation model was built to calculate life-long radiation exposure in asymptomatic adults treated for AJCC stage IIIA-D melanoma aged 20, 40, or 60-years at diagnosis. Imaging protocols included CT chest-abdomen-pelvis and CT brain imaging at intervals of three, six or 12-months, over a three, five or 10-year surveillance duration. Organ doses were estimated using Monte-Carlo-simulation-based dosimetry software (CT Expo) with parameters from national delivered CT dose data (2015), calculated using methods from the Biological Effects of Ionizing Radiation (BEIR VII) report, and the National Cancer Institute's Radiation Risk Assessment Tool (RadRAT); with adjustment for stage III melanoma life-expectancy. New cancers attributable to CT radiation were presented as cases per 100,000 population.
Findings: Mean estimated lifetime attributable risk of new cancers from cumulative CT imaging for males was 250 (90%CI 107 to 443) to 6,310 (90%CI 2550 to 11,230) cases per 100,000 population depending on age, frequency of imaging and surveillance duration. For females, this risk was 252 (90%CI 135 to 406) to 7,710 (90%CI 2,550 to 11,230) cases per 100,000 population. Risk was highest for both genders at 20-years-old undergoing three-monthly imaging for 10-years (n=40 scans), and for those with Stage IIIA compared to IIIB/C/D melanoma. Cancer risk from radiation exposure decreased markedly as a function of age and fewer total CTs.
Interpretation: In cancer survivors, frequent and prolonged CT surveillance may increase the absolute risk of new cancers, particularly among young people. This harm needs to be considered alongside potential benefits including earlier detection of treatable disease.