J. S. Sung1, H. G. Wu1, J. H. Kim1, S. H. Ahn2, E. J. Chung2, B. Keam3, S. E. Hyun4, and J. H. Lee1; 1Department of Radiation Oncology, Seoul National University Hospital, Seoul, Korea, Republic of (South), 2Otorhinolaryngology-Head and Neck Surgery, Seoul National University Hospital, Seoul, Korea, Republic of (South), 3Department of Internal Medicine, Seoul National University Hospital, Seoul, Korea, Republic of (South), 4Department of Rehabilitation Medicine, Seoul National University Hospital, Seoul, Korea, Republic of (South)
Purpose/Objective(s):In laryngeal and hypopharyngeal cancers (LHC), radiation therapy (RT) has been widely used for organ-preserving definitive treatments, However, long-term survivors of LHC experience radiation-associated dysphagia (RAD). We sought to identify dosimetric parameters associated with late RAD in LHC.Materials/
Methods: We conducted a single institutional retrospective study on a cohort of patients who were histologically confirmed laryngeal or hypopharyngeal squamous cell carcinoma and treated by definitive RT from 2012 to 2019. Patients with initial dysphagia at diagnosis, progression of disease during follow-up, and an insufficient follow-up period of fewer than 6 months were excluded. We used the Common Terminology Criteria for Adverse Events (CTCAE) system to evaluate dysphagia occurring later than six months after RT. The substructures for the segmentation and dosimetric analysis were the inferior, middle, and superior constrictors (IPC, MPC, and SPC), cricopharyngeal inlet (CPI), cervical esophagus (CE), supraglottic and glottic larynx (SGL and GL), medial pterygoids (MPM), anterior digastrics (ADM), and base of tongue (BOT). Random forest algorithm was used to identify dose-volumetric parameters highly associated with RAD. Univariate and multivariate analyses were performed with clinical and dosimetric factors. Receiver operator characteristic (ROC) curve analysis was used to identify the optimal threshold point and evaluate its predictive ability. Results: A total of 73 patients were included, of which 12 (16.4%) developed late RAD. Patients with hypopharyngeal and laryngeal cancer were 28 (38.4%) and 45 (61.6%), respectively. Among the clinical factors, late RAD was highly correlated with hypopharyngeal cancer (p=0.003), cervical lymph node irradiation (p=0.02), and stage IV (p=0.048) in the univariate analysis. Dose-volumetric parameters of SPC, MPC, IPC, CE, MPM, ADM, and BOT showed a high correlation with late RAD in the random forest algorithm. On multivariate analysis, hypopharyngeal cancer (p=0.007) and the volume receiving 60Gy or more (V60) of BOT (p=0.028) were most critically associated with a higher possibility of late RAD. The ROC curve of BOT V60 showed a cutoff value of 11.5, and the area under the curve (AUC) was 0.795. Conclusion: Critical substructures for late RAD in hypopharyngeal and laryngeal cancers were identified. V60 of BOT was newly identified as a surrogate and dose constraint for late RAD in LHC patients. Dose to BOT should be monitored and limited when feasible in LHC patients undergoing RT.
