Abstract
BACKGROUND: The aim of the study was to determine the interactive effect of age on overall survival (OS) and relative survival (RS) benefits of radiotherapy (RT) in early-stage diffuse large B-cell lymphoma (DLBCL). METHODS: Data for 10,841 adults with early-stage DLBCL from the Surveillance, Epidemiology, and End Results database between 2002 and 2015 were retrospectively analyzed. Primary therapy was classified as combined-modality treatment (CMT; n = 3,631) and chemotherapy alone (n = 7,210). Inverse probability of treatment weighting was used to balance covariate distribution between the treatment groups. Survival was estimated and compared using the Kaplan-Meier method and log-rank test, respectively. Age-RT interactive effect on survival was examined through Cox regression multiplicative interaction analysis. RESULTS: Using age of 60 years as the reference, older age was an independent predictor of shorter OS in the multivariable Cox model (hazard ratio (HR), 1.07; 95% confidence interval (CI), 1.06-1.07; P < 0.001). After controlling for background mortality, older age was not an independent predictor of RS (HR, 1.00; 95% CI, 0.99-1.00; P = 0.842). Across all age groups, patients treated with CMT had better OS and RS than those who received chemotherapy alone. A significant interaction between age and RT was identified for both OS (P(interaction) = 0.020) and RS (P(interaction) = 0.038), indicating greater RT benefit in young patients. A linear correlation existed between RS and OS at the treatment arm level. CONCLUSIONS: RT was associated with improved net survival across all ages, particularly for young adults. RS was a valid alternative endpoint for prognostication and benefit evaluation.