Outcomes of Salvage Therapy for Oropharyngeal Cancer Recurrence Following Upfront Radiation Therapy and Prognostic Factors

Purpose This study aimed to investigate the oncologic outcomes and prognostic factors of salvage treatments in patients with recur-rent oropharyngeal squamous cell carcinoma (OPSCC) after radiotherapy (RT)-based treatment. Materials and Methods A cancer registry was used to retrieve the records of 337 patients treated with definitive RT or concurrent chemoradiotherapy (CRT) from 2008 to 2018 at a single institution. The poor-responder group (PRG) was defined as patients with residual or recurrent disease after primary treatment, and the oncologic outcomes for each salvage treatment method were analyzed. In addition, prognostic indicators of recurrence-free survival (RFS) and overall survival (OS) were identified in patients who underwent salvage treatment. Results After initial (C)RT, the PRG comprised 71 of the 337 patients (21.1%): 18 patients had residual disease, and 53 had recur-rence after primary treatment (mean time to recurrence 19.5 months). Of these, 63 patients received salvage treatment (surgery 57.2%, re-(C)RT 23.8%, and chemotherapy 19.0%), and the salvage success rate was 47.6% at the last follow-up. The overall 2-year OS for salvage treatments was 56.4% (60.8% for the salvage surgery group and 46.2% for the salvage re-(C)RT). Salvage surgery patients with negative resection margins had better oncologic outcomes than those with close/positive resection margins. Using multivariate analyses, locoregional recurrence and residual disease after primary surgery were associated with poor outcome after salvage treatment. In Kaplan-Meier analyses, p16 status was significantly associated with OS in the initial treatment setting but not in the salvage setting. Conclusion In recurrent OPSCC after RT-based treatment, successful salvage was achieved in 56.4% patients who had undergone salvage surgery and radiation treatment. Salvage treatment methods should be selected carefully, given recurrence site as a prog-nostic factor for RFS.