Trends in Use and Clinical Outcomes Associated with Post-Operative Radiation Therapy for HPV-Associated Oropharyngeal Squamous Cell Carcinomas

Abstract

<h3>Purpose/Objective(s)</h3> Intermediate and high-risk features (IRFs/HRFs) for locoregional recurrence following initial surgery for oropharyngeal SCCs (OP-SCCs) were defined prior to the known association of HPV with OP-SCC. There are limited reports on practice patterns and outcomes associated with post-operative radiation therapy (RT) or chemoradiation (CRT). We aimed to address these questions with the National Cancer Database (NCDB). <h3>Materials/Methods</h3> The NCDB was queried for patients with HPV-associated OP-SCCs managed initially with surgery with IRFs or HRFs. IRFs were defined as T3/T4 disease, node positive disease, and the presence of lymphovascular space invasion (LVSI), and HRFs as positive margins and extranodal extension (ENE). Patients were stratified into groups if they received no adjuvant therapy, RT, or CRT. Patients treated with RT or C-RT were required to have documentation of receipt of 50-70 Gy. Kaplan-Meier analysis was utilized for comparison of overall survival (OS) between treatment arms followed by a Cox multivariate (MVA) proportional-hazards model and propensity score analyses performed using inverse probability treatment weighting (IPTW). <h3>Results</h3> We identified 6,301 patients; 1,319 patients had positive margins, 2,266 patients had ENE, and 3,074 patients had one or two HRFs. The median number of RFs was 2 (range: 1-5). Regarding treatment, 1,606, 2,407, and 2288 patients received no RT, RT, and CRT, respectively. The median RT dose was 60 Gy (range: 50-70 Gy). Median age was 58 years (range: 24-90) and the median Charlson-Deyo comorbidity score was 0 (range: 0-3). Among patients with IRFs, 8-year OS with no RT (84.3%) was significantly lower vs. RT (90.1%) or CRT (88.8%; <i>p</i> < 0.0001) with no difference with RT vs. CRT (<i>p</i> = 0.13). On Cox MVA, RT was associated with significantly improved OS vs. no RT (hazard ratio (HR) = 0.57 (95% CI: 42.7-75.2%); <i>p</i> < 0.0001) but not CRT (HR = 1.37 (95% CI: 0.97-1.93). On IPTW analyses, receipt of RT or CRT was associated with significantly improved OS on Cox MVA (HR = 0.72 (95% CI: 0.57-0.91; <i>p</i> = 0.006). Among patients with HRFs, 8-year OS with no RT (69.1%) was significantly lower vs. RT (79.3%) or CRT (82.1%; <i>p</i> < 0.001) with no difference with RT vs. CRT (<i>p</i> = 0.22). On sub-group analysis of patients with positive margins, 8-year OS was not significantly different between RT (80.5%) and CRT (81.2%; <i>p</i> = 0.72). On Cox MVA comparing RT to CRT, CRT was not associated with significantly improved OS (HR = 0.94 (95% CI: 0.74-1.18)) for patients with HRFs. On IPTW analyses, CRT was not associated with improved 8-year OS (81.8%) vs. RT (81.5%) for patients with HRFs or on Cox MVA (HR = 0.97 (95% CI: 0.76-1.24); <i>p</i> = 0.81)). <h3>Conclusion</h3> A significant proportion of HPV-associated OP-SCC patients with IRFs or HRFs did not receive RT, which was associated with inferior OS. CRT did not result in superior outcomes compared to RT for patients with HRFs. Traditional HRFs as indications for CRT require re-evaluation for HPV-associated OP-SCC patients in prospective studies.