Involved Site Radiotherapy in HPV Positive Oropharyngeal Cancer: Patterns of Failure Analysis Across Prospective De-Escalation Trials

Abstract

HPV-associated oropharyngeal cancers (HPV+OPC) have a favorable prognosis with ongoing efforts to reduce long term toxicity while maintaining oncologic outcomes. One method under investigation includes lowering the elective radiation dose or in some cases omitting radiation to elective lymphatic nodal stations. Furthermore, pre-clinical evidence demonstrates that elective nodal irradiation blunts the anti-tumor immune response in head and neck cancer. This is a pooled secondary analysis reporting patterns of failure in patients (pts) with HPV+OPC enrolled on consecutive induction chemo- or chemoimmunotherapy (IC) based response-adaptive de-escalation trials and treated with involved-site radiotherapy (ISRT). Pts treated on two prospective phase II trials as well as on a prospective cohort registry were included for analysis. Pts with ≥ 50% response to IC based on RECIST 1.1 who received de-escalated definitive radiotherapy (RT) or concurrent chemoradiation (CRT) with ISRT were evaluable. Pts with locally advanced low risk or high-risk HPV+OPC (LR and HR, respectively) were eligible for enrollment. Pts were considered to have HR if at least one of the following criteria was met: T4 primary, N2c-N3 disease (AJCC 7th ed.), or > 10-20 pack years smoking. In the first trial, pts with ≥ 50% response to IC received RT to gross disease plus a 1.5 cm margin (PTV1) and to the next echelon of uninvolved nodes (PTV2). Pts with LR received 50 Gy in 2 Gy daily fractions without chemotherapy; pts with HR received 30 Gy in 1.5 Gy BID fractions to PTV2 with a 15 Gy sequential boost to PTV1 with CRT. In the subsequent trial, pts with ≥ 50% response received RT to PTV1 alone to 50 Gy in 2 Gy daily fractions; concurrent CRT was included if pts had HR. Survival was estimated using the Kaplan Meier method for progression free survival (PFS), locoregional PFS (LRPFS), and overall survival (OS). Patterns of failure analysis was performed by comparing RT plans to radiographic surveillance scans. Of 172 evaluable pts, 119 (69.2%) achieved a ≥ 50% response to IC and received definitive ISRT. 45 (37.8%) pts evaluated received RT to gross disease only plus margin without the next nodal echelon included. With a median follow up of 46 (IQR 34-65) months, 3-year PFS, LRPFS, and OS with their 95% confidence intervals were 96.2% (90.1-98.5%), 97.1% (91.1-99.0%), and 96.2% (90.3-98.6%), respectively. All locoregional failures were in-field and in the high dose region. No failures were observed in the RT omitted neck. This prospective experience demonstrates feasible volume de-escalation using IC response-based selection with progressively smaller elective volumes over time. Notably, despite a marked reduction in elective treatment volume, there were no regional out-of-field failures. IC may allow for selection of pts with favorable tumor biology and microscopic disease sterilization in the regional nodes. Further efforts at elective nodal de-escalation are needed.