Can intensity of long-term follow-up for survivors of childhood and teenage cancer be determined by therapy-based risk stratification?

OBJECTIVE: To determine the feasibility of therapy-based, risk-stratified follow-up guidelines for childhood and teenage cancer survivors by evaluating adverse health outcomes in a survivor cohort retrospectively assigned a risk category. DESIGN: Retrospective cohort study. SETTING: Tertiary level, single centre, paediatric cancer unit in South East Scotland. PARTICIPANTS: All children and teenagers diagnosed with cancer (<19 years) between 1 January 1971 and 31 July 2004, who were alive more than 5 years from diagnosis formed the study cohort. Each survivor was retrospectively assigned a level of follow-up, based on their predicted risk of developing treatment-related late effects (LEs; levels 1, 2 and 3 for low, medium and high risk, respectively). Adverse health outcomes were determined from review of medical records and postal questionnaires. LEs were graded using the Common Terminology Criteria for Adverse Event, V.3. RESULTS: 607 5-year survivors were identified. Risk stratification identified 86 (14.2%), 271 (44.6%) and 250 (41.2%) as levels 1, 2 and 3 survivors, respectively. The prevalence of LEs for level 1 survivors was 11.6% with only one patient with grade 3 or above toxicity. 35.8% of level 2 survivors had an LE, of whom 9.3%, 58.8%, 18.5%, 10.3% and 3% had grades 1, 2, 3, 4 and 5 toxicity, respectively. 65.2% of level 3 survivors had LE, of whom 5.5% (n=9), 34.4% (n=56), 36.2% (n=59), 22.1% (n=36) and 1.8% (n=3) had grades 1, 2, 3, 4 and 5 toxicity, respectively. CONCLUSIONS: Therapy-based risk stratification of survivors can predict which patients are at significant risk of developing moderate-to-severe LEs and require high-intensity long-term follow-up. Our findings will need confirmation in a prospective cohort study that has the power to adjust for all potentially confounding variables.