The impact of adjuvant chemoradiotherapy timing on survival of head and neck cancers.

Abstract

Delays in postoperative head and neck (HN) radiotherapy have been associated with decreased overall survival; however, the impact of delays in postoperative HN chemoradiotherapy remains undefined. All patients with nonmetastatic HN cancer (oral cavity, oropharynx, larynx, hypopharynx) who underwent curative intent surgery and received adjuvant chemoradiotherapy were identified from the National Cancer Database (2005-2012). Overall treatment time (OTT) was defined as the time from surgery to the end of radiation therapy. Statistical methods included Cox proportional hazards modeling, which adjusted for clinicopathologic, demographic, and socioeconomic factors. Recursive partitioning analysis (RPA) identified the optimal threshold of OTT via conditional inference trees to estimate the greatest differences in overall survival (OS) on the basis of randomly selected training and validation sets. A total of 16,733 patients were included, with a median follow-up of 37 months. Median OS for OTT in a predefined threshold of ≤ 13 weeks was 10.1 years (95% confidence interval [CI], 9.8 years; not reached) compared with 8.7 years (95% CI, 8.2-9.2 years) in > 13 weeks. On multivariate analysis, OTT of > 13 weeks versus ≤ 13 weeks independently increased mortality risk (hazard ratio, 1.10; 95% CI, 1.04-1.17; P = < 0.001). RPA identified an optimal OTT threshold of 97 days (interquartile range: 96-98 days). The OTT threshold of 97 days was confirmed in a full Cox regression model estimating the risk of death according to overall treatment time as a continuous variable. In this large hospital-based national data, an OTT of greater than approximately 14 weeks most consistently increased the risk of death. 4. Laryngoscope, 128:2326-2332, 2018.