Long-term medical imaging use in children with central nervous system tumors

Background Children with central nervous system (CNS) tumors undergo frequent imaging for diagnosis and follow-up, but few studies have characterized longitudinal imaging patterns. We described medical imaging in children before and after malignant CNS tumor diagnosis. Procedure We conducted a retrospective cohort study of children aged 0-20 years diagnosed with CNS tumors between 1996-2016 at six U.S. integrated healthcare systems and Ontario, Canada. We collected computed topography (CT), magnetic resonance imaging (MRI), radiography, ultrasound, nuclear medicine examinations from 12 months before through 10 years after CNS diagnosis censoring six months before death or a subsequent cancer diagnosis, disenrollment from the health system, age 21 years, or December 31, 2016. We calculated imaging rates per child per month stratified by modality, country, diagnosis age, calendar year, time since diagnosis, and tumor grade. Results We observed 1,879 children with median four years follow-up post-diagnosis in the U.S. and seven years in Ontario, Canada. During the diagnosis period (+/- 15 days of diagnosis), children averaged 1.10 CTs (95% confidence interval [CI] 1.09-1.13) and 2.14 MRIs (95%CI 2.12-2.16) in the U.S., and 1.67 CTs (95%CI 1.65-1.68) and 1.86 MRIs (95%CI 1.85-1.88) in Ontario. Within one year after diagnosis, 19% of children had >= 5 CTs and 45% had >= 5 MRIs. By nine years after diagnosis, children averaged one MRI and one radiograph per year with little use of other imaging modalities. Conclusions MRI and CT are commonly used for CNS tumor diagnosis, whereas MRI is the primary modality used during surveillance of children with CNS tumors.