Outcomes Using TORS or Definitive Chemoradiation for T3/T4 HPV+ Oropharynx Cancer

Abstract

Patients with HPV+ oropharyngeal squamous cell carcinoma (OPSCC) generally have a good prognosis, though those with locally advanced (LA) disease are at increased risk of locoregional failure (LRF) and distant metastases (DM). A recent SEER study suggested improvement in overall survival and cancer-specific mortality with trimodality therapy compared to definitive chemoradiation (dCRT) for LA HPV+ OPSCC, but is prone to confounding and lacks pattern of failure data. We therefore performed this retrospective analysis of oncologic outcomes for patients with T3/T4 HPV+ OPSCC treated at a high-volume TORS center with surgery +/- adjuvant therapy compared to dCRT alone.Included were patients with cT3/T4 HPV+ OPSCC treated with dCRT or upfront TORS +/- adjuvant therapy from 2010-2019 with ≥1 year of follow-up. Patients with DM at diagnosis were excluded. Primary outcomes were freedom from LRF (FF-LRF) and LRF leading to death, defined as LRF directly resulting in death or occurring at the same time or before DM resulting in death. Secondary outcomes were freedom from DM (FF-DM) and cancer-specific survival (CSS). Patients were risk stratified based on criteria in the secondary analysis of RTOG 0129 by Ang et al (NEJM 2010): Low-risk: ≤10 pack-year (py) smoking, or > 10 py and N0-N2a disease (AJCC 7th ed.); Intermediate-risk: > 10 py and N2b-N3 disease. Fisher's exact test was used to compare demographics and Kaplan-Meier estimates to compare endpoints.78 patients were treated with dCRT and 44 with upfront TORS. Significantly more dCRT patients were Intermediate-risk (per RTOG 0129 criteria) compared to TORS patients (35.9% vs 6.8%, P < 0.001). ECOG performance status was similar between groups with most 0-1. 87% of TORS patients received adjuvant therapy (32% CRT and 55% RT alone). The median follow-up was 43 months. There was no difference in 4-year FF-LRF for dCRT vs TORS (87% vs 93%, P = 0.33) and no difference in freedom from LRF leading to death (93% vs 98%, P = 0.30). There was also no difference in CSS (92% vs 95%, P = 0.14) or FF-DM (80% vs 86%, P = 0.45). On univariate analysis, RTOG 0129 risk group predicted for LRF (OR 3.4, P = 0.05) but not LRF leading to death (OR 3.14, P = 0.175). Multivariate analysis of LRF as a function of 0129 risk group and treatment modality showed neither were statistically significant (OR for risk group 3.22, P = 0.08; OR for surgery 0.85, P = 0.833), though risk group trended towards significance.This large, single-institution series of patients with T3/T4 HPV+ OPSCC treated with either dCRT or upfront TORS +/- adjuvant therapy shows high FF-LRF and low rates of LRF leading to death regardless of treatment. Even in patients with advanced T stage, dCRT was able to obtain excellent disease outcomes. Future prospective studies are warranted to determine which combinations of multimodality therapy are optimal for LA HPV+ OPSCC, incorporating both oncologic and functional outcome endpoints.