[Skip to Content]
[Skip to Content Landing]
Figure.
Overall Survival Kaplan-Meier Curves
Overall Survival Kaplan-Meier Curves

Plus signs indicate censored data. CRT indicates chemoradiotherapy; and PORT, postoperative radiotherapy.

Table 1.  
Patient and Disease Characteristics of the Cohort
Patient and Disease Characteristics of the Cohort
Table 2.  
Prevalence and Predictors of Adjuvant Chemoradiotherapy (CRT) Use
Prevalence and Predictors of Adjuvant Chemoradiotherapy (CRT) Use
Table 3.  
Univariable and Multivariable Predictors of Postoperative Radiotherapy (RT) Doses Greater Than 60 Gy
Univariable and Multivariable Predictors of Postoperative Radiotherapy (RT) Doses Greater Than 60 Gy
Table 4.  
Univariable and Multivariable Predictors of Overall Survival (OS)
Univariable and Multivariable Predictors of Overall Survival (OS)
1.
Sinha  P, Piccirillo  JF, Kallogjeri  D, Spitznagel  EL, Haughey  BH.  The role of postoperative chemoradiation for oropharynx carcinoma: a critical appraisal of the published literature and National Comprehensive Cancer Network guidelines.  Cancer. 2015;121(11):1747-1754.PubMedGoogle ScholarCrossref
2.
Cooper  JS, Pajak  TF, Forastiere  A,  et al.  Precisely defining high-risk operable head and neck tumors based on RTOG #85-03 and #88-24: targets for postoperative radiochemotherapy?  Head Neck. 1998;20(7):588-594.PubMedGoogle ScholarCrossref
3.
Cooper  JS, Pajak  TF, Forastiere  AA,  et al; Radiation Therapy Oncology Group 9501/Intergroup.  Postoperative concurrent radiotherapy and chemotherapy for high-risk squamous-cell carcinoma of the head and neck.  N Engl J Med. 2004;350(19):1937-1944.PubMedGoogle ScholarCrossref
4.
Bernier  J, Domenge  C, Ozsahin  M,  et al; European Organization for Research and Treatment of Cancer Trial 22931.  Postoperative irradiation with or without concomitant chemotherapy for locally advanced head and neck cancer.  N Engl J Med. 2004;350(19):1945-1952.PubMedGoogle ScholarCrossref
5.
Bernier  J, Cooper  JS, Pajak  TF,  et al.  Defining risk levels in locally advanced head and neck cancers: a comparative analysis of concurrent postoperative radiation plus chemotherapy trials of the EORTC (#22931) and RTOG (#9501).  Head Neck. 2005;27(10):843-850.PubMedGoogle ScholarCrossref
6.
Pfister  DG. National Comprehensive Cancer Network (NCCN) Guidelines Discussion. https://www.ahns.info/presentations/2010/NCCN/NCCN.swf. Accessed July 20, 2015.
7.
Cooper  JS, Zhang  Q, Pajak  TF,  et al.  Long-term follow-up of the RTOG 9501/intergroup phase III trial: postoperative concurrent radiation therapy and chemotherapy in high-risk squamous cell carcinoma of the head and neck.  Int J Radiat Oncol Biol Phys. 2012;84(5):1198-1205.PubMedGoogle ScholarCrossref
8.
Sinha  P, Lewis  JS  Jr, Piccirillo  JF, Kallogjeri  D, Haughey  BH.  Extracapsular spread and adjuvant therapy in human papillomavirus-related, p16-positive oropharyngeal carcinoma.  Cancer. 2012;118(14):3519-3530.PubMedGoogle ScholarCrossref
9.
Peters  LJ, Goepfert  H, Ang  KK,  et al.  Evaluation of the dose for postoperative radiation therapy of head and neck cancer: first report of a prospective randomized trial.  Int J Radiat Oncol Biol Phys. 1993;26(1):3-11.Google ScholarCrossref
10.
Ang  KK, Trotti  A, Brown  BW,  et al.  Randomized trial addressing risk features and time factors of surgery plus radiotherapy in advanced head-and-neck cancer.  Int J Radiat Oncol Biol Phys. 2001;51(3):571-578.Google ScholarCrossref
11.
Sanguineti  G, Richetti  A, Bignardi  M,  et al.  Accelerated versus conventional fractionated postoperative radiotherapy for advanced head and neck cancer: results of a multicenter phase III study.  Int J Radiat Oncol Biol Phys. 2005;61(3):762-771.Google ScholarCrossref
12.
Harari  PM, Harris  J, Kies  MS,  et al.  Postoperative chemoradiotherapy and cetuximab for high-risk squamous cell carcinoma of the head and neck: Radiation Therapy Oncology Group RTOG-0234.  J Clin Oncol. 2014;32(23):2486-2495.PubMedGoogle ScholarCrossref
13.
Maxwell  JH, Ferris  RL, Gooding  W,  et al.  Extracapsular spread in head and neck carcinoma: impact of site and human papillomavirus status.  Cancer. 2013;119(18):3302-3308.PubMedGoogle ScholarCrossref
14.
Ang  KK, Harris  J, Wheeler  R,  et al.  Human papillomavirus and survival of patients with oropharyngeal cancer.  N Engl J Med. 2010;363(1):24-35.PubMedGoogle ScholarCrossref
Original Investigation
August 2016

Patterns of Care and Comparative Effectiveness of Intensified Adjuvant Therapy for Resected Oropharyngeal Squamous Cell Carcinoma in the Human Papillomavirus Era

Author Affiliations
  • 1Division of Outcomes and Health Services Research, Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas
  • 2Head and Neck Disease Oriented Team, Simmons Cancer Center, University of Texas Southwestern, Dallas
  • 3Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas
  • 4Department of Medical Oncology, University of Texas Southwestern Medical Center, Dallas
  • 5Department of Otolaryngology–Head and Neck Surgery, University of Texas Southwestern Medical Center, Dallas
  • 6Department of Radiation Oncology, University of Chicago, Chicago, Illinois
JAMA Otolaryngol Head Neck Surg. 2016;142(8):777-788. doi:10.1001/jamaoto.2016.1162
Abstract

Importance  There is a growing debate on the relative benefits of adjuvant chemoradiotherapy (CRT) and boost doses of postoperative radiotherapy (B-PORT) in oropharyngeal squamous cell carcinoma (OPSCC) treated with primary surgery, especially for patients with human papillomavirus (HPV)-driven disease.

Objective  To characterize the recent patterns of care in and overall survival (OS) outcomes following the use of adjuvant CRT and B-PORT after primary surgery for OPSCC.

Design, Setting, and Participants  Retrospective analysis of patients in the National Cancer Database with stage III to IVA-B OPSCC treated with surgery and adjuvant radiotherapy between 2010 and 2012 at Commission on Cancer–accredited facilities. The data analysis was performed between June 15, 2015, and May 4, 2016.

Main Outcomes and Measures  The primary outcomes were prevalence of CRT and B-PORT, and OS. The primary predictors were HPV positivity and high-risk pathologic features (HRPFs) (extracapsular extension and positive surgical margins).

Results  Of the 1409 patients (1153 [82%] male; median age, 57 [interquartile range {IQR}, 51-63] years), 873 (62%) and 789 (56%) patients received CRT and B-PORT, respectively; most patients (n = 583 [79%]) with HRPFs received CRT, and many patients (n = 227 [40%]) without HRPFs received CRT. Multivariable predictors of CRT included adverse pathologic features (extracapsular extension [OR, 6.99; 95% CI, 5.22-9.35], positive surgical margins [OR, 2.07; 95% CI, 1.50-2.87], ≥6 involved nodes [OR, 2.34; 95% CI, 1.39-3.92], or low-neck disease [OR, 1.52; 95% CI, 1.01-2.28]), and treatment at a nonacademic institution (OR, 1.59 [95% CI, 1.21-2.10] for comprehensive community cancer center vs academic program). Patients with HPV-positive disease (OR, 0.47; 95% CI, 0.33-0.68) were less likely to receive CRT; this decrease was limited to absent HRPF treated at academic institutions (n = 173, 44 [25%] received CRT). With a median follow-up of surviving patients of 27 (IQR, 21-33) months, the 2-year OS probability was 92% (95% CI, 90%-94%). Multivariable analysis including age, sex, pathologic T stage, 6 or more positive nodes, and educational status confirmed the prognostic impact of HPV positivity (hazard ratio [HR], 0.41; 95% CI, 0.21-0.80) and HRPFs (positive surgical margins [HR, 2.15; 95% CI, 1.27-3.66] and ≥6 involved nodes [HR, 2.11; 95% CI, 1.13-3.93]), but neither CRT (HR, 1.27; 95% CI, 0.70-2.30) nor B-PORT (HR, 1.04; 95% CI, 0.63-1.73) was associated with improved OS.

Conclusions and Relevance  Postoperative CRT and B-PORT following resection of OPSCC were dependent on factors beyond HRPFs, including HPV status and treatment at an academic institution. No benefit was seen with intensified adjuvant therapy, supporting enrollment of the HPV-positive population into deintensification trials.

Introduction

The optimal intensity of postoperative radiotherapy (PORT) for oropharyngeal squamous cell carcinoma (OPSCC) is controversial.1 Although PORT has been established as an important component of the treatment paradigm following most surgical resections, historically the locoregional failure risk has still been substantial, particularly in the context of high-risk pathologic features, such as positive surgical margins (PSM) and extracapsular extension (ECE).2

The need to improve locoregional control in the postoperative setting generated 2 landmark randomized trials, EORTC study 22931 and Intergoup/RTOG 9501, both of which compared PORT with PORT plus 3 cycles of concurrent bolus cisplatin.3,4 The former study showed a significant survival benefit to chemoradiotherapy (CRT), whereas the original report for RTOG 9501 only showed a disease-free survival gain. An influential combined analysis suggested that the therapeutic gain was limited to patients with PSM and/or ECE,5 and by 2010 the National Comprehensive Cancer Network (NCCN) guidelines recommended adjuvant CRT only for patients with PSM or ECE, with “consideration” of CRT for others with adverse characteristics such as multiple nodes or low-neck disease.6 An unplanned subset analysis in the final 2012 report from RTOG 9501 further suggested that any survival gain from concurrent treatment was restricted to patients with these high-risk characteristics.7 This latter trial highlighted the toxic effects of combined therapy, as 3% of patients died from treatment, and this risk of toxic effects may not be warranted in the current era of treatment-sensitive, human papillomavirus (HPV)-driven disease.1 In fact, recent retrospective analyses have suggested that the survival benefit of CRT is lost in patients with HPV-positive OPSCC with ECE.8

The other mechanism to intensify adjuvant locoregional therapy is through radiation dose escalation, but prospective studies of PORT dose have been lacking. An early study of PORT alone showed that patients should receive at least 57.6 Gy, with a boost to 63 Gy for patients with ECE, to maximize locoregional control.9 Two subsequent studies found that accelerating PORT to less than 6 weeks was not associated with a survival benefit.10,11 These studies, in combination with respected prospective trials using roughly 60 Gy for standard-risk and 66 Gy for high-risk pathologic features,3,4,12 have established 60 to 66 Gy as the standard dose regimen for PORT.

This uncertainty in the utility of CRT and dose escalation may lead to varying postoperative practice patterns, and there are limited data characterizing recent adjuvant treatment approaches. Moreover, there are few data analyzing the real-world effectiveness of adjuvant CRT and dose escalation following surgery for OPSCC, which is particularly relevant given the increasing prevalence of HPV-positive cancer. Randomized trials are generally small with a restricted patient population, and large database analysis may provide insight into practice patterns and the comparative effectiveness of different intensities of treatment. The purpose of this study was to evaluate the recent patterns of care and survival outcomes with postoperative CRT and dose escalation for patients with OPSCC.

Box Section Ref ID

Key Points

  • Question What are the recent patterns of care in the postoperative treatment of patients with oropharyngeal squamous cell carcinoma, and to what extent does intensified locoregional therapy affect survival?

  • Findings In this analysis of patient data from the National Cancer Database, high-risk pathologic features, human papillomavirus negativity, and treatment at a nonacademic institution were associated with the use of chemoradiotherapy. Treatment with chemoradiotherapy and boost doses of radiotherapy did not improve overall survival.

  • Meaning The variation in care deserves further investigation to optimize outcomes in this population; these data support enrollment of the human papillomavirus–positive population into deintensification trials.

Methods
Database

The National Cancer Database (NCDB) is a combined program of the Commission on Cancer (CoC) of the American College of Surgeons and the American Cancer Society. There are more than 1500 CoC-accredited institutions, and the NCDB includes more than 70% of patients with a new cancer diagnosis. The data used in the study are derived from a deidentified NCDB file. This study received exempt status from the University of Texas Southwestern institutional review board.

Cohort Definition

Eligible patients received a diagnosis of pathologic stage III to IVB OPSCC between 2010 and 2012, when HPV status became available. This malignant neoplasm was their first or only cancer, and information on the date of surgery, radiotherapy (RT), and/or chemotherapy was mandatory, as was RT dose information; patients without these data were excluded. All patients were required to have had treatment with initial surgery to the primary tumor and neck and to have received PORT within 10 weeks of surgery. The surgical procedure must have included resection of the primary tumor (excluding excisional biopsy) and neck surgery; patients with lymph node aspiration only or neck “sampling” were excluded. Concurrent chemotherapy must have started between 1 week before the first fraction of RT to 3 weeks after the first fraction. A total dose between 59.4 and 66.6 Gy was necessary for eligibility into the analysis. The strict dose limits were imposed to ensure that treatment was delivered with radical, adjuvant intent.

Determination of Predictor and Outcome Variables

Predictors were divided into clinical, geographic, socioeconomic, and institutional variables. Human papillomavirus status was recorded in 837 (59%) of the cohort, and positivity was considered if HPV type 16 and/or HPV type 18 were found; the method of testing for HPV was not specified (eg, polymerase chain reaction). The categorical variable of HPV status was recorded as negative, positive, or unknown. The p16 status was not recorded. The multivariable analyses included all patients, regardless of HPV status, and unknown HPV status was included as its own variable. High-risk disease was defined as the presence of PSM or ECE. The presence of these high-risk features was extracted from the pathology report at each institution, and so there was no formal standardization of the definition of a positive margin; per the coding criteria, we defined PSM when the margin status was “residual tumor not otherwise specified,” “microscopic residual tumor,” or “macroscopic residual tumor.” A negative margin was defined as “all margins are grossly and microscopically negative.” Similarly, there was no formal grading of ECE, which was coded in the NCDB as “microscopic,” “macroscopic,” or “unknown microscopic/macroscopic.” Additional pathologic variables included low-neck disease (level 4 or below) and the presence of 6 or more positive lymph nodes; this latter number (ie, 6) was chosen because of its inclusion as a subgroup analysis in the update of RTOG 9501.7 Several of the ordinal variables were categorized by the NCDB as quartiles relative to the US population, and age, number of examined lymph nodes, and distance from the facility were stratified into quartiles.

There were 3 main outcome variables: treatment with CRT, delivery of boost dose RT (defined as treatment >60 Gy), and overall survival (OS), the latter of which was only analyzed in patients treated in 2010 through 2011, for whom survival data were available.

Statistical Analyses

Differences in patient characteristics between patients who did or did not receive CRT and boost PORT were tested using the χ2 test. The 95% confidence interval of the difference (CID) in the use of CRT or boost PORT in the presence or absence of a given characteristic was calculated and is presented in the text, with the P value in the tables. We then performed separate stepwise multivariable logistic regressions with CRT and boost PORT as the end points, with all covariates included in the initial regression, with preservation in the model if the adjusted P value was .05 or less.

Univariable survival analyses were performed with the log-rank test. Multivariable survival analysis was also performed, using the same stepwise selection algorithm. Treatment with CRT, PORT dose, and HPV status were forced into the model given their primary importance in the analysis. Propensity score matching was performed to reduce bias due to confounding factors that may have affected the choice of treatment. The propensity models were developed using all potential covariates; 1-to-1 propensity matching was then carried out using the caliper match algorithm described by Coca Perraillon, with the caliper width set to 0.05, and these cohorts (ie, adjuvant CRT vs PORT alone) were compared using a log-rank test, and the hazard ratio (HR) was derived using univariable Cox regression.12 All analyses were performed with SAS, version 9.4.

Results
Patient Characteristics

Patient, disease, and treatment characteristics are presented in Table 1. Most patients were male and relatively young. Patients with stage III disease made up 311 (22%) of the cohort, and most patients had stage IVA disease (1020 [72%]), consistent with the T- and N-stage distribution of the cohort. Neck dissections were robust, with a median of 31 (interquartile range, 20-45) nodes dissected. A nontrivial percentage of the population had some degree of unfavorable pathologic features, ranging from PSM (n = 282 [20%], of whom only 3 patients had macroscopic residual tumor) and ECE (n = 746 [53%]) to low-neck disease (level 4 or 5, n = 200 [14%]). Only 67 patients (9% of those with ECE, 4.8% of total cohort) had macroscopic ECE. Most patients had a favorable comorbidity score and socioeconomic status. Most were treated with adjuvant concurrent CRT, and 56% (n = 789) of patients received a boost. The majority of patients (n = 1002 [71%]) were treated with intensity-modulated RT. Of the patients with known HPV status (n = 837 [59%]), 599 (72%) were HPV positive. Testing for HPV was significantly more common in academic centers (64% vs 53% in academic and comprehensive community cancer cancers, respectively; 95% CID, 5.0%-16.4%).

Predictors of Adjuvant Concurrent Chemoradiotherapy

Table 2 presents the predictors of concurrent chemotherapy with PORT. Univariable analyses revealed that patients with HPV-positive OPSCC were significantly less likely to receive CRT (58% vs 69%; 95% CID, 4.2%-18.9%). Subset analysis of patients with known HPV status revealed that CRT was used less often only in those without adverse pathologic features (33% HPV-positive vs 55% HPV-negative; 95% CID, 9.3%-33.9%), and not in individuals with high-risk disease (77% HPV-positive vs 82% HPV-negative; 95% CID, −3.4% to 14.3%).

Patients treated at academic institutions were also less likely to receive chemotherapy (CRT delivered in 59% of patients treated in academic vs 67% nonacademic institutions; 95% CID, 2.3%-12.9%). However, this difference was also specific to patients without high-risk features (CRT delivered in 34% academic vs 51% nonacademic; 95% CID, 9.2%-26.7%), as there was no difference in CRT use in patients with high-risk features (80% academic vs 77% nonacademic; 95% CID, −3.5% to 9.3%). When looking at both HPV status and academic status of the facility, the difference in CRT use between academic and nonacademic institutions was driven by HPV-positive patients without high-risk features, as these patients were significantly less likely to receive CRT in an academic facility (25% vs 54%; 95% CID, 13.8%-42.0%).

On multivariable regression (Table 2), patients with HPV-positive cancer were significantly less likely to receive adjuvant CRT (odds ratio [OR], 0.47; 95% CI, 0.33-0.68). The other covariates independently associated with CRT implementation reflected the motivation to intensify therapy in higher-risk patients perceived to tolerate the treatment, including not only those with PSM (OR, 2.07; 95% CI, 1.50-2.87) and ECE (OR, 6.99; 95% CI, 5.22-9.35) but also low-neck disease (OR, 1.52; 95% CI, 1.01-2.28), 6 or more positive nodes (OR, 2.34; 95% CI, 1.39-3.92), and higher nodal stage. There were no socioeconomic characteristics that predicted for concurrent CRT delivery, and patients treated at a comprehensive community cancer center were significantly more likely to receive CRT than those treated at academic institutions (OR, 1.59; 95% CI, 1.21-2.10).

When this model was developed with an interaction term between academic institution (ie, academic or not) and high-risk features, it was significant (P < .001), indicating that the reduction in the odds of receiving CRT at academic vs nonacademic institutions in patients without ECE and PSM (OR, 0.43; 95% CI, 0.31-0.60) was significantly different than the odds increase (OR, 1.32; 95% CI, 0.90 -1.93) of receiving CRT at academic vs nonacademic institutions in patients with ECE or PSM. The interaction term between HPV status and high-risk features was not statistically significant (P = .07).

Predictors of Radiotherapy Dose Escalation

Predictors of boost RT are presented in Table 3. On univariable analysis, in contrast to the use of adjuvant CRT, the difference in boost prevalence by HPV status was not significantly associated with the presence of adverse pathologic features. The boost was delivered in 59% and 62% of HPV-positive and negative patients with high-risk features, respectively (95% CID, −0.08% to 13.9%), and 40% and 49% of positive and negative patients without high-risk features, respectively (95% CID, −0.03% to 21.7%).

The most influential independent variables predicting for the boost were associated with more advanced disease (ie, PSM, ECE, and T4 status). In addition, Northeast location significantly increased the odds of receiving the boost, as patients treated there were approximately twice as likely to receive the boost as those in the Midwest and West regions. The use of intensity-modulated RT was less common in patients receiving higher doses of RT. Finally, HPV-positive patients were significantly more likely to receive lower doses (OR, 0.80; 95% CI, 0.69-0.94).

Predictors of Overall Survival

The median follow-up for surviving patients was 27 months (interquartile range, 21-33 months). The 2-year OS probability for the whole cohort was 92% (95% CI, 90%-94%) (Figure, A). Patients with HPV-positive disease experienced excellent survival outcomes (2-year OS, 95%; 95% CI, 93%-98%) (Figure, B), which were superior to those in patients with HPV-negative cancer (2-year OS, 85%; 95% CI, 79%-91%). The 2-year survival for HPV-positive patients without high-risk features was 100% after treatment with PORT alone (n = 86) and 98% (95% CI, 94%-100%) with CRT (n = 54), and the 2-year survival in HPV-positive patients with high-risk features was 89% (95% CI, 76%-100%) with PORT alone (n = 41) and 93% (95% CI, 88%-97%) with CRT (n = 134). Only 9 HPV-negative patients with high-risk features were treated with PORT alone, so no conclusions can be drawn on that population.

Univariable and multivariable predictors of OS are presented in Table 4. Several classic adverse pathologic predictors of survival maintained significance, including PSM, T stage, and presence of multiple positive lymph nodes. Patients with HPV-positive disease experienced improved OS (HR, 0.41 vs HPV-negative; 95% CI, 0.21-0.80). Older patients, women, and individuals with a lower educational status also faced higher mortality risks.

Neither treatment with adjuvant concurrent CRT (HR, 1.27; 95% CI, 0.70-2.30) nor higher doses of PORT (HR, 1.03; 95% CI, 0.63-1.72) was associated with improved OS, even in the subset of patients with high-risk features (HR, 1.06 for CRT; 95% CI, 0.46-2.43, and HR, 1.47 for boost dose; 95% CI, 0.76-2.85). In the propensity-matched cohorts of 220 each, there was no statistically significant difference in OS (2-year OS, 93% for CRT vs 94% for RT) (Figure, C) in patients treated with CRT (HR, 1.35 favoring PORT alone; 95% CI, 0.66-2.76). Propensity matching was also used to generate a cohort of patients equally likely to receive RT doses above 60 Gy (301 in each arm). There was no significant OS benefit with higher doses of RT (2-year OS, 91% for boost vs 94% for standard dose).

Discussion

In this analysis using a large population-based cohort of patients treated with primary surgery and adjuvant RT for OPSCC, we found substantial variability in the intensity of postoperative treatment. In addition to the more conventional and expected high-risk pathologic factors such as ECE and PSM, other adverse pathologic factors such as low-neck disease and multiple positive nodes were also associated with the use of adjuvant CRT, and the presence of lymphovascular invasion increased the odds of higher PORT dose.

We also found that HPV-positive status and treatment at an academic institution significantly influenced the decision to deliver adjuvant CRT. Indeed, although individuals with HPV-positive disease were significantly less likely to receive chemotherapy, closer analysis suggested that this difference was restricted to patients without ECE and PSM. This treatment approach was consistent with NCCN guidelines, and so it should not be considered deintensification, yet it does reflect how physicians were more willing to deliver less therapy to patients with improved prognoses. In fact, the finding that HPV status did not influence treatment decisions in patients with more advanced pathologic factors is reassuring that deintensification for HPV-positive individuals has not yet diffused into routine practice.

There was evidence that academic institutions have different practice patterns than nonacademic centers. For example, HPV testing was significantly more likely in these centers. Academic facilities were less likely to deliver CRT to HPV-positive patients without ECE and PSM, and because this recommendation is most consistent with the national guidelines, this result does highlight the benefit of consultation at tertiary care centers for complicated head and neck cancer treatment decisions. However, it is important to note that patients seeking care at academic centers may have different characteristics than those who seek care at community facilities, and so unmeasured confounders in this population (such as differential smoking status, or patient aversion to receive concurrent chemotherapy) may have contributed to this finding as well.

Although we had relatively modest follow-up, our OS results provide important insight into the modern postoperative management of OPSCC. First, we confirmed that HPV-positive status confers a significantly improved prognosis in routine practice, and HPV positivity was easily the most powerful prognostic variable in the model. Postoperative radiotherapy alone for HPV-positive patients without high-risk features led to 100% survival at 2 years, supporting treatment with PORT alone in this population. Other established prognostic factors such as low-neck disease, multiple positive lymph nodes, advanced T stage, ECE, and PSM were all negatively associated with survival. Because the margin width was not specified in the database (simply negative, microscopically positive, or macroscopically positive), no conclusions can be drawn on the OS effect of a close but negative margin (eg, 1-5 mm). On the other hand, we saw no signal that CRT or higher PORT dose was associated with improved OS, which was further supported by the propensity score analysis.

The finding that adjuvant CRT did not improve OS may not be surprising. The final report from RTOG 9501 had a finding of insignificant OS benefit in the high-risk pathologic subset (HR, 0.76; 95% CI, 0.57-1.00),7 but this older study presumably had a higher percentage of patients with HPV-negative tumors than the present analysis. In addition, several careful retrospective analyses from Washington University in St Louis and University of Pittsburgh found that CRT was not associated with an OS advantage for patients with ECE for p16-positive oropharyngeal cancer.8,13 However, these retrospective studies were limited because most patients with ECE received concurrent chemotherapy, restricting the impact of any conclusion on the benefit of intensified adjuvant therapy.

Similarly, we found no improvement in OS with higher doses of RT. Although 66 Gy has become the de facto dose for patients with high-risk disease, there are limited prospective data on RT dose in the postoperative setting; perhaps it is this uncertainty that explains why more favorable HPV-positive patients were less likely to receive the boost. An early randomized trial suggested that a minimum of 63 Gy (in 35 fractions, roughly equal to 60 Gy in 30 fractions) was needed to improve locoregional control with ECE, but this higher dose did not lead to a survival benefit.9 Two randomized trials of accelerated PORT alone for patients with poor-prognosis pathologic features showed no improvement with altered fractionation beyond 63 or 60 Gy.10,11 Given evidence in the transoral surgery setting that doses above 60 Gy may lead to a higher risk of soft-tissue necrosis, the preponderance of data suggests restricting the upper end of PORT dose to 60 Gy.

Our study has several limitations inherent in large database analysis that must be considered in its interpretation. First, we could not adjust for several important confounding variables, such as pretreatment imaging, smoking status, and the patient’s presurgery and postsurgery performance status. The comorbidity score helps mitigate the lack of this information, but not entirely. Importantly, the quality of the pathologic information was dependent on facility coders reading through their own pathology reports, and so its accuracy and completeness require validation. There were no standard definitions of positive margin status or ECE (and certainly no formal grading of the extent of ECE), which does make cross-institutional comparisons more difficult. Similarly, HPV testing was not standardized and did not incorporate p16 status, although we believe that it is unlikely that testing heterogeneity would have meaningfully affected treatment decision making. Although p16 positivity may still be an unknown confounder for survival, it is improbable that a sufficient number of discordant HPV-p16 patients in the database would have changed the results because HPV status was in the model. That said, we were able to adjust for many potentially confounding clinical, pathologic, demographic, and socioeconomic variables, and thus we are confident in the generalizability of our results. In fact, the reduction in mortality (HR, 0.42) seen with the HPV-positive cohort is similar to the HR of 0.44 seen in RTOG 0129, lending further support for the validity of these results.14 Finally, there are intrinsic selection biases in retrospective database studies, such that patients treated in a CoC-affiliated facility may have associated unknown confounders that affected treatment decisions or OS outcomes. Nevertheless, because more than 70% of newly diagnosed cancers are included in the NCDB, we do not think that unadjusted selection bias meaningfully influenced the results.

Importantly, the NCDB only includes OS data, and disease-free survival and locoregional control—important end points in head and neck cancer—were not assessable. We therefore could not determine whether CRT improved locoregional disease control in patients with high-risk disease. Finally, the limited follow-up time must be acknowledged in order to interpret the survival data, and the possibility of late recurrences with inadequate adjuvant therapy cannot be discounted.

Conclusions

We have characterized contemporary patterns of care in the adjuvant treatment of patients with OPSCC, revealing that the intensity of therapy was dependent on HPV status and the academic nature of the facility, as well as pathologic factors beyond the established PSM and ECE. More time is necessary to determine whether the final results of RTOG 9501 further reduce the implementation of CRT for the lower-risk patient population. The favorable OS outcomes for HPV-positive patients without ECE and PSM argue against CRT in this population, and the impressive OS results with and without these factors support enrollment in randomized trials deintensifying treatment, either through radiation dose deescalation or sparing concurrent chemotherapy.

Back to top
Article Information

Accepted for Publication: May 5, 2016.

Corresponding Author: David J. Sher, MD, MPH, Department of Radiation Oncology, UT Southwestern Medical Center, 5810 Forest Park Dr, Dallas, TX 75390 (david.sher@utsouthwestern.edu).

Published Online: June 30, 2016. doi:10.1001/jamaoto.2016.1162.

Author Contributions: Dr Sher 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: Sher, Khan, Koshy.

Acquisition, analysis, or interpretation of data: Sher, Nedzi, Hughes, Sumer, Myers, Truelson, Koshy.

Drafting of the manuscript: Sher, Nedzi, Koshy.

Critical revision of the manuscript for important intellectual content: Sher, Khan, Hughes, Sumer, Myers, Truelson.

Statistical analysis: Sher, Koshy.

Administrative, technical, or material support: Khan, Myers.

Study supervision: Sumer, Myers.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Disclaimer: The American College of Surgeons and the Commission on Cancer have not verified and are not responsible for the analytic or statistical methodology used, or our conclusions drawn from these data.

References
1.
Sinha  P, Piccirillo  JF, Kallogjeri  D, Spitznagel  EL, Haughey  BH.  The role of postoperative chemoradiation for oropharynx carcinoma: a critical appraisal of the published literature and National Comprehensive Cancer Network guidelines.  Cancer. 2015;121(11):1747-1754.PubMedGoogle ScholarCrossref
2.
Cooper  JS, Pajak  TF, Forastiere  A,  et al.  Precisely defining high-risk operable head and neck tumors based on RTOG #85-03 and #88-24: targets for postoperative radiochemotherapy?  Head Neck. 1998;20(7):588-594.PubMedGoogle ScholarCrossref
3.
Cooper  JS, Pajak  TF, Forastiere  AA,  et al; Radiation Therapy Oncology Group 9501/Intergroup.  Postoperative concurrent radiotherapy and chemotherapy for high-risk squamous-cell carcinoma of the head and neck.  N Engl J Med. 2004;350(19):1937-1944.PubMedGoogle ScholarCrossref
4.
Bernier  J, Domenge  C, Ozsahin  M,  et al; European Organization for Research and Treatment of Cancer Trial 22931.  Postoperative irradiation with or without concomitant chemotherapy for locally advanced head and neck cancer.  N Engl J Med. 2004;350(19):1945-1952.PubMedGoogle ScholarCrossref
5.
Bernier  J, Cooper  JS, Pajak  TF,  et al.  Defining risk levels in locally advanced head and neck cancers: a comparative analysis of concurrent postoperative radiation plus chemotherapy trials of the EORTC (#22931) and RTOG (#9501).  Head Neck. 2005;27(10):843-850.PubMedGoogle ScholarCrossref
6.
Pfister  DG. National Comprehensive Cancer Network (NCCN) Guidelines Discussion. https://www.ahns.info/presentations/2010/NCCN/NCCN.swf. Accessed July 20, 2015.
7.
Cooper  JS, Zhang  Q, Pajak  TF,  et al.  Long-term follow-up of the RTOG 9501/intergroup phase III trial: postoperative concurrent radiation therapy and chemotherapy in high-risk squamous cell carcinoma of the head and neck.  Int J Radiat Oncol Biol Phys. 2012;84(5):1198-1205.PubMedGoogle ScholarCrossref
8.
Sinha  P, Lewis  JS  Jr, Piccirillo  JF, Kallogjeri  D, Haughey  BH.  Extracapsular spread and adjuvant therapy in human papillomavirus-related, p16-positive oropharyngeal carcinoma.  Cancer. 2012;118(14):3519-3530.PubMedGoogle ScholarCrossref
9.
Peters  LJ, Goepfert  H, Ang  KK,  et al.  Evaluation of the dose for postoperative radiation therapy of head and neck cancer: first report of a prospective randomized trial.  Int J Radiat Oncol Biol Phys. 1993;26(1):3-11.Google ScholarCrossref
10.
Ang  KK, Trotti  A, Brown  BW,  et al.  Randomized trial addressing risk features and time factors of surgery plus radiotherapy in advanced head-and-neck cancer.  Int J Radiat Oncol Biol Phys. 2001;51(3):571-578.Google ScholarCrossref
11.
Sanguineti  G, Richetti  A, Bignardi  M,  et al.  Accelerated versus conventional fractionated postoperative radiotherapy for advanced head and neck cancer: results of a multicenter phase III study.  Int J Radiat Oncol Biol Phys. 2005;61(3):762-771.Google ScholarCrossref
12.
Harari  PM, Harris  J, Kies  MS,  et al.  Postoperative chemoradiotherapy and cetuximab for high-risk squamous cell carcinoma of the head and neck: Radiation Therapy Oncology Group RTOG-0234.  J Clin Oncol. 2014;32(23):2486-2495.PubMedGoogle ScholarCrossref
13.
Maxwell  JH, Ferris  RL, Gooding  W,  et al.  Extracapsular spread in head and neck carcinoma: impact of site and human papillomavirus status.  Cancer. 2013;119(18):3302-3308.PubMedGoogle ScholarCrossref
14.
Ang  KK, Harris  J, Wheeler  R,  et al.  Human papillomavirus and survival of patients with oropharyngeal cancer.  N Engl J Med. 2010;363(1):24-35.PubMedGoogle ScholarCrossref
×