Background: Whether the waiting time for radical radiotherapy (WRT) detrimentally impacts nasopharyngeal carcinoma (NPC) prognosis is unclear. We estimated the influence of WRT on overall survival (OS) and disease-specific survival (DSS) of NPC.
Patients and methods: Patients were identified from prospectively maintained database. WRT was calculated from histological diagnosis to initiation of radiotherapy (RT). Survival analysis was estimated using Weibull parametric model and propensity score analysis (PSA). Recursive partitioning analysis (RPA) identified optimal WRT threshold via conditional inference trees to estimate the greatest survival differences based on randomly selected training and validation sets, and this process was repeated 1000 times to ensure threshold robustness. Sensitivity analysis estimated effects of potential unmeasured confounders.
Results: A total of 9896 patients were included. In multivariate analysis, WRT of 31-60°d, of 61-90°d and of greater than 90°d independently increased mortality risk compared to less than 30°d. Upon RPA, ranges of 30-35°d with the peak of 30°d were confirmed with 89% of simulations validating optimal thresholds. In threshold-based groups, adjusted hazard ratios (HRs) for WRT of greater than 30°d by both Weibull model and PSA were significantly higher than for WRT of less than 30°d [OS: HR = 1.13, 95% confidence interval (CI) 1.04-1.23, P = 0.003; DSS: HR = 1.15, 95% CI 1.05-1.26, P = 0.002]. Sensitivity analysis revealed robustness of results.
Conclusions: WRT independently affects survival. Increasing WRT beyond 30°d was most consistently detrimental to survival. WRT of NPC should be as short as reasonably achievable (ASARA).
Keywords: ASARA; Nasopharyngeal carcinoma; Propensity score analysis; Radical radiotherapy; Recursive partitioning analysis; Sensitivity analysis; Survival; Waiting time; Weibull analysis.
Copyright © 2017 Elsevier Ltd. All rights reserved.