Comparison of Survival Outcomes Among Human Papillomavirus–Negative cT1-2 N1-2b Patients With Oropharyngeal Squamous Cell Cancer Treated With Upfront Surgery vs Definitive Chemoradiation Therapy: An Observational Study

Key Points

Question  Does upfront surgery improve overall survival in patients with cT1-2 N1-2b human papillomavirus (HPV)-negative oropharyngeal squamous cell carcinoma (OPSCC)?

Findings  Upfront surgical intensification of treatment does not improve overall survival in patients with cT1-2 N1-2b HPV-negative OPSCC, including a subset analysis of patients with margin-negative surgery. Nearly 60% of surgical patients received trimodal therapy with adjuvant chemoradiation.

Meaning  Given that upfront surgical intensification is not associated with improved overall survival, further research should focus on better selection of surgical patients who are less likely to require adjuvant CRT, as trimodal therapy increases treatment-related toxic effects.

Importance  Human papillomavirus (HPV)-negative oropharyngeal squamous cell carcinoma (OPSCC) has shown resistance to conventional concurrent chemoradiation (CRT) therapy and carries a relatively poor prognosis in comparison with HPV-positive disease, with decreased locoregional control and overall survival (OS). In the present analysis, we examine whether upfront surgical resection improves overall survival in a large national sample.

Objective  To compare survival outcomes among patients with newly diagnosed cT1-2 N1-2b HPV-negative OPSCC when treated with primary surgical resection vs CRT.

Design, Setting, and Participants  This was an observational study of factors associated with primary treatment modality were identified using multivariable logistic regression. Overall survival was compared using Kaplan-Meier analysis with log-rank tests, multivariable Cox regression, and propensity score matching. Statistical tests were 2-sided. Patients newly diagnosed as having cT1-2 N1-2b pathologically confirmed HPV-negative OPSCC in 2010 to 2012 were identified using the National Cancer Data Base, which includes more than 70% of patients newly diagnosed as having cancer in the United States.

Exposures  Primary surgical resection vs definitive CRT.

Main Outcomes and Measures  Overall survival.

Results  We identified 1044 patients, among whom 460 (44.1%) received upfront surgery and 584 (55.9%) received CRT. Median age was 59 years (range, 25-90 years); 812 patients were male (77.8%), 232 were female (22.2%). Median follow-up was 30 months. Approximately 59% of surgical patients received adjuvant CRT. On multivariable Cox regression, upfront surgery was not associated with increased OS when compared with CRT (adjusted hazard ratio [HR], 1.01; 95% CI, 0.74-1.39; P = .93). Propensity score-matching identified a cohort of 822 patients and redemonstrated equivalent OS (HR, 1.14; 95% CI, 0.81-1.62; P = .46). Lack of OS benefit with upfront surgery persisted in a subset analysis of patients with margin-negative resection (HR, 0.97; 95% CI, 0.66-1.45; P = .88).

Conclusions and Relevance  In this observational study, OS was similar for patients with HPV-negative OPSCC when treated with primary surgery vs CRT. Most surgical patients received trimodal therapy with adjuvant CRT. Our data may have implications for future research focusing on optimal patient selection for surgery.

Human papilloma virus (HPV)-negative oropharyngeal squamous cell cancer (OPSCC) presents a therapeutic challenge, as survival has not improved using treatment intensification with induction chemotherapy,¹ concurrent chemoradiation (CRT) with accelerated fractionation,² or the addition of cetuximab to CRT.³ Treatment-related toxic effects may limit further intensification using chemotherapy or radiation therapy (RT). The advent of minimally invasive transoral robotic surgery (TORS) renders treatment intensification with surgery an enticing option. RTOG 1221 (NCT01953952) thus attempted to randomize patients with cT1-2 N1-2b HPV-negative OPSCC to TORS or CRT to explore whether surgical treatment intensification improved outcomes. While the trial was closed owing to a lack of accrual, this question remains salient.

In the absence of randomized prospective data, carefully controlled, population-based analyses can provide important evidence regarding comparative effectiveness. We therefore used the National Cancer Data Base (NCDB) to determine if primary surgical resection would improve overall survival (OS) compared with definitive CRT among patients with T1-2 N1-2b HPV-negative OPSCC.

In this observational study, using the NCDB, we identified patients with cT1-2 N1-2b OPSCC pathologically negative for high-risk HPV (eFigure 1 in the Supplement). Chemoradiation was defined as chemotherapy initiation within 30 days of RT initiation. The institutional review board of Yale University granted this study exempt status. The requirement for informed consent was also waived.

Patient clinicodemographic factors were categorized and compared by treatment group (upfront surgery vs CRT) using the χ² test. A multivariable logistic regression model was constructed using backward elimination with a univariable inclusion criterion of P < .10. The variance inflation factor was estimated to exclude colinearity, and goodness-of-fit was ensured using the Hosmer-Lemeshow test.

Overall survival was evaluated using the time from diagnosis until death or the date the patient was last contacted. Univariable Cox regression was used to calculate unadjusted hazard ratios (HRs) for survival using clinicopathologic factors potentially affecting survival. Variables trending toward significance on univariable analysis (P < .10) were included in a Cox multivariable regression to identify factors significantly associated with improved OS. Schoenfeld residuals were calculated to ensure the proportional hazards assumption was not violated. The Kaplan-Meier estimator and the log-rank test were used to compare OS between treatment groups. Shared frailty analysis was performed to examine heterogeneity between treatment centers.

Propensity score matching was performed using bootstrapping with 1-to-1 nearest-neighbor matching without replacement (caliper distance of 25% of the standard deviation of the pooled propensity scores) to identify matched cohorts representing the 2 treatment modalities. Matching was performed using variables significantly associated with receipt of CRT on multivariable analysis. Covariate balance was evaluated using standardized differences of means. We then repeated survival analyses in the matched groups as described for the unmatched cohort, adjusting for propensity quintile. For this paired sample, the robust variance estimator was used for our Cox proportional hazards model.

All tests were 2-sided, and a P <.05 was considered statistically significant. All analyses were performed using Stata SE statistical software (version 13.0; StataCorp).

We identified 1044 patients with HPV-negative T1-2 N1-2b OPSCC diagnosed from 2010 to 2012. Median age was 59 years (range, 25-90 years); 812 patients were male (77.8%), 232 were female (22.2%). Median follow up was 30 months. Clinical and demographic characteristics are presented in Table 1.

On multivariable regression, the odds of receiving CRT increased with nonprivate insurance status, higher T and N stage, absence of comorbid conditions, proximity to treatment facility and unknown histologic grade (eTable in the Supplement).

In the entire cohort, 3-year OS was 79.2% (95% CI, 74.5%-82.9%) for patients receiving CRT and 81.4% (95% CI, 76.9%-85.2%) for upfront surgery (log-rank P = .65; Figure, A). Of the factors associated with OS on univariable analysis, higher T stage (HR, 1.85; 95% CI, 1.27-2.57; P = .001) and higher comorbidity score (HR, 2.14; 95% CI, 1.53-2.99; P < .001) remained significantly associated with OS on multivariable analysis (Table 2). In comparison with CRT, primary surgery was not associated with improved OS on univariable (HR, 0.93; 95% CI, 0.68-1.26; P = .65) or multivariable (HR, 1.01; 95% CI, 0.74-1.39; P = .93) analysis. Results were unchanged when shared frailty was incorporated.

Propensity score matching produced a well-matched cohort of 822 patients (eFigure 2 in the Supplement) and redemonstrated similar OS with surgical management compared with CRT (HR, 1.14; 95% CI, 0.81-1.62; P = .46). The 3-year OS was 81.3% (95% CI, 76.6%-85.2%) for the upfront surgery group and 82.5% (95% CI, 77.4%-86.6%) for the matched CRT group (Figure, B).

We performed a subset analysis among patients with negative margins (microscopic or macroscopic) after surgery. A lack of OS benefit with upfront surgery persisted on univariable analysis (3-year OS, 83.6% vs 79.3%; log-rank P = .23; Figure, C), multivariable analysis (HR, 0.97; 95% CI, 0.66-1.45; P = .88), and propensity score–matched analysis (HR, 0.8; 95% CI, 0.53-1.28; P = .39) in this cohort.

To our knowledge, this study is the first large-scale analysis to compare upfront surgery to CRT in patients with HPV-negative OPSCC. No significant difference in OS was found in our cohort. This finding persisted in a propensity score–matched analysis and a sensitivity analysis of patients with margin-negative surgery, suggesting that lack of OS benefit is not a reflection of poor surgical quality. This finding is more notable because CRT patients were more likely to have poorer prognostic features (eg, higher T/N stage, nonprivate insurance). Older studies, prior to the era of HPV testing, similarly demonstrated a lack of OS benefit but were conducted in the general population of patients with head and neck cancer using outdated treatment paradigms, and showed higher rates of severe toxic effects with upfront surgery (32% vs 4%; P < .001).^4,5

Our finding that 59.1% of patients received adjuvant CRT is comparable with rates in other series, including those using TORS.^6-8 Adjuvant CRT is indicated for patients with positive margins and/or extracapsular extension (ECE)⁹ but increases toxic effects.^10-12 That nearly 60% of patients treated with upfront surgery received trimodal therapy is notable given the lack of OS benefit and probable increased toxic effects. To spare patients the morbidity of trimodal therapy, further research is needed to identify preoperative predictors of patients who will require adjuvant CRT.¹³

The findings of this analysis do not negate the role of surgery in HPV-negative OPSCC. Rather, patients undergoing TORS who do not require adjuvant CRT may have favorable quality of life in comparison with those receiving definitive CRT, a question currently under investigation (NCT01590355). However, our results suggest upfront surgery should not be recommended solely because it is felt to intensify therapy or because it is hypothesized to overcome perceived resistance mechanisms of tumors to CRT, which is particularly important as primary surgery for early-stage OPSCC is rising.¹⁴ Rather, treatment recommendations should be based on factors such as patient preference, extent of disease, and the presence of radiographic ECE.

A limitation of our study lies in the selection bias inherent in retrospective analyses. Our methods incorporated propensity score matching and adjustment for multiple sources of bias including T stage and comorbidity, although unmeasured covariates could be unaccounted for. Moreover, because the NCDB does not contain information regarding disease recurrence, we cannot determine the impact of surgical intervention on outcomes other than OS.

Our study found no significant OS difference between patients with HPV-negative OPSCC receiving upfront surgery and definitive CRT, suggesting that upfront surgery does not aid in overcoming the resistance mechanisms of HPV-negative disease to CRT. That nearly 60% of patients treated with upfront surgery received adjuvant CRT is notable, as trimodal therapy likely worsens toxic effects without a discernable OS benefit. Our data may have implications for future research focusing on optimal patient selection for surgery, as TORS shows promise in improving quality of life in patients requiring minimal adjuvant therapy.

Corresponding Author: Zain A. Husain, MD, Smilow Cancer Hospital, Yale Cancer Center, PO Box 208040, New Haven, CT 06510 (zain.husain@yale.edu).

Accepted for Publication: October 14, 2016.

Published Online: January 5, 2017. doi:10.1001/jamaoncol.2016.5769

Author Contributions: Dr Kelly had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Kelly, Yarbrough, Decker, Husain.

Acquisition, analysis, or interpretation of data: Kelly, Park, An, Contessa, Yarbrough, Burtness, Decker, Husain.

Drafting of the manuscript: Kelly.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kelly, Park.

Administrative, technical, or material support: An, Decker.

Conflict of Interest Disclosures: Dr Park has received honoraria and travel expenses from Varian Medical Systems, Inc. Dr Husain has received research funding from Merck pharmaceuticals. No other disclosures are reported.