Reduction in late mortality among 5-year survivors of childhood cancer: A report from the Childhood Cancer Survivor Study (CCSS).

Pediatric Oncology
Session Type and Session Title: 
Plenary Session, Plenary Session Including the Science of Oncology Award and Lecture
Abstract Number: 


J Clin Oncol 33, 2015 (suppl; abstr LBA2)
Gregory T. Armstrong, Yutaka Yasui, Yan Chen, Wendy M. Leisenring, Todd M. Gibson, Ann Mertens, Marilyn Stovall, Melissa M. Hudson, Kevin C. Oeffinger, Smita Bhatia, Kevin R. Krull, Paul C. Nathan, Joseph Philip Neglia, Daniel M. Green, Leslie L. Robison; St. Jude Children's Research Hospital, Memphis, TN; University of Alberta, Edmonton, AB, Canada; Fred Hutchinson Cancer Research Center, Seattle, WA; Emory University, Atlanta, GA; Department of Radiation Physics, The University of Texas M. D. Anderson Cancer Center, Houston, TX; Memorial Sloan Kettering Cancer Center, New York, NY; University of Alabama at Birmingham, Birmingham, AL; The Hospital for Sick Children, Toronto, ON, Canada; University of Minnesota, Minneapolis, MN

Abstract Disclosures


Background: Over the past four decades, treatment of many childhood cancers has been modified with the aim of achieving high survival rates while reducing the risk of life-threatening late-effects, and promoting risk-based follow-up care of survivors. Methods: Late mortality was evaluated in 34,033 5-year survivors (diagnosed < 21 years of age from 1970-1999, median follow-up 21 years, range 5-38) using cumulative incidence and Poisson regression models adjusted for demographic and disease factors to calculate relative risk (RR) and 95% confidence intervals (CI). Mortality due to non-recurrence/non-external (NR/NE) causes, which includes deaths that reflect late-effects of cancer therapy, was evaluated. Results: 1,622 (41%) of the 3,958 deaths were attributable to NR/NE causes, including 751 subsequent neoplasm (SN), 243 cardiac, and 136 pulmonary deaths. Changes in therapy by decade included reduced rates of: cranial radiotherapy (RT) for acute lymphoblastic leukemia (ALL, 86%, 54%, 22%), RT for Wilms tumor (WT, 77%, 54%, 49%) and RT for Hodgkin lymphoma (HL, 96%, 88%, 77%). Reductions in 15 year cumulative NR/NE mortality were observed across treatment eras for ALL (p < .001), HL (p = .005), and WT (p = .005). Cardiac deaths decreased in ALL (p = .002), HL (p = .06), and WT (p = .04), and SN deaths decreased in WT (p < .001). Year of diagnosis (adjusted for age, sex, diagnosis, follow-up time) was significantly associated with a reduced risk of all-cause mortality (RR = 0.85, CI 0.83-0.87), NR/NE death (RR = 0.87, CI 0.84-0.91), death from SN (RR = 0.84, CI 0.80-0.89), cardiac death (RR = 0.78, CI 0.69-0.87) and pulmonary death (RR = 0.79, CI 0.68-0.91). Conclusions: The CCSS cohort provides evidence that the strategy of modifying therapy to reduce the occurrence of late-effects, and promotion of early detection, is successfully translating into a significant reduction in observed late mortality.

Cumulative incidence (%) of death at 15 years from diagnosis.

Treatment eraAll-CauseNR/NE CausesSNCardiacPulmonary
P-value< 0.001< 0.001< 0.0010.0010.02