Patients were categorized according to level of seroresponse to E6 or E7 antigen (seronegative for both E6 and E7 vs tertiles of increasing HPV titer level for E6 or E7).
Patients were classified as HPV16 negative (seronegative for E6 and E7) vs positive (seropositive for E6 or E7).
Patients were categorized as having either a light (lowest tertile for both tobacco and alcohol consumption), moderate (middle tertile for at ≥1 exposure), or heavy (highest tertile for either tobacco or alcohol exposure history) exposure.
eTable. Independent and joint effects of seropositivity to E6 and E7 proteins
Customize your JAMA Network experience by selecting one or more topics from the list below.
Nelson HH, Pawlita M, Michaud DS, McClean M, Langevin SM, Eliot MN, Kelsey KT. Immune Response to HPV16 E6 and E7 Proteins and Patient Outcomes in Head and Neck Cancer. JAMA Oncol. 2017;3(2):178–185. doi:10.1001/jamaoncol.2016.4500
Copyright 2016 American Medical Association. All Rights Reserved.
Does immune response to human papillomavirus (HPV) E6 and E7 proteins predict patient outcomes for head and neck cancer?
Using a large, population-based study of patients with head and neck cancer, we investigated the relationship between patient titers for HPV E6 and E7 proteins at the time of diagnosis and patient survival. Any positive seroresponse, irrespective of level, was associated with a significantly improved outcome for all forms of head and neck cancer (oropharyngeal, oral cavity, and larynx.
Seropositivity to HPV E6 and E7 proteins may be a useful alternative or adjunct to pathology-based testing for HPV in head and neck cancers, particularly for tumors outside of the oropharynx.
Pathology-based measures of human papillomavirus (HPV) status are routinely obtained in the care of head and neck cancer and are clearly associated with patient outcome for cancers of the oropharynx. However, it is unclear if HPV status is of high value for cancers of the larynx and oral cavity. In addition, it is possible to assess HPV infection using serology-based methods; however, the suitability of this pathology-independent measure for predicting patient outcome in head and neck cancer is unknown.
To investigate whether immunologic response to HPV16 is associated with patient survival across anatomic sites, independent of smoking and drinking history.
Design, Setting, and Participants
This was a population-based study of 1054 patients with head and neck cancer in the greater Boston, Massachusetts, area (1999-2003, 2006-2011).
Main Outcomes and Measures
All-cause survival in relation to HPV16 E6 and E7 seropositivity.
The 1054 patients reflected the demographics of those treated in this timeframe (75% male; mean age, 59 years). Seropositivity was very strongly associated with improved survival overall (hazard ratio HR], 0.33; 95% CI, 0.24-0.45; P < .001), with no evidence that the magnitude of immune response, as assessed by titer levels, effected outcome. Seropositivity was associated with improved patient survival across all head and neck cancer sites: HR for oropharynx cancer, 0.26; 95% CI, 0.18-0.39; for oral cavity cancer, 0.45; 95% CI, 0.18-0.80; and for larynx cancer, 0.29; 95% CI, 0.10-0.85. In addition, the associations with seropositivity were similar across smoking and/or drinking exposure groups: HRfor low exposure, 0.52; 95% CI, 0.20-1.36; for moderate exposure, 0.42; 95% CI, 0.25-0.70; for heavy exposure, 0.51; 95% CI, 0.36-0.73. In a subset of 162 patients with both HPV serology and p16 immunohistochemical (IHC) measures available, both measures were similarly associated with survival in the oropharynx (HR for serology, 0.16; 95% CI, 0.03-0.47; for p16 measures, 0.16; 95% CI, 0.03-0.46), whereas only serology was associated with outcome when considering all head and neck cancer cases (HR for serology,0.49; 95% CI, 0.23-1.04; for p16, 0.65; 95% CI, 0.30-1.42).
Conclusions and Relevance
Collectively, these data suggest that a positive serologic response to HPV16 oncoproteins may be the best approach to assess HPV-disease for clinical outcome because it is associated with survival for all types of disease and is a marker that is not dependent on pathology material.
Human papillomavirus 16 (HPV16) is a well-recognized etiologic factor in oropharyngeal cancer, with serology-based data clearly demonstrating that infection precedes cancer occurrence.1 Oropharyngeal tumors that arise as a result of HPV infection have a distinct mutation landscape,2-4 and patients with HPV16-positive oropharyngeal cancer have markedly improved survival.5,6 Determination of HPV16 status is most often achieved using pathologic materials. However, the accuracy of this testing may be compromised depending on the biomarker used (ie, p16 staining is not highly specific to HPV16 infection outside the oropharynx7 or among African Americans8) and the sampling of the tumor (ie, if the presence of HPV is not clonal and homogeneous throughout the pathology specimen9). Recent work has demonstrated that seropositivity to HPV16 E6 and E7 proteins, which can be assessed independent of pathology material, is also strongly associated with patient survival in oropharyngeal cancer10 and can be used for monitoring cancer recurrence.11 This work raises the interesting question as to whether the magnitude of the serologic response to HPV16 is an important prognosticator for oropharyngeal cancer.
While the association of HPV with oropharyngeal cancer occurrence and outcome is established, there is continued controversy regarding the role of HPV in cancers of the oral cavity and larynx. A recent large, multisite study with multiple markers of infection suggests that active HPV infection is quite low in oral cavity (4.4%) and laryngeal (3.5%) tumors.12 An analysis of RNA-sequencing data in The Cancer Genome Atlas (TCGA) tumors detected HPV16 transcripts in approximately 10% of oral cavity and 5% of larynx tumors.13 Despite the low rates of active infection in these tumors, the prevalence of seropositivity to HPV16 E6 and E7 proteins is dramatically higher in patients with oral cavity and laryngeal cancers relative to healthy population controls.14 Furthermore, while HPV16 has been intensively investigated as a favorable prognostic factor for oropharyngeal cancer, less is known regarding the association of HPV16 with outcome for these other anatomic sites. The most robust data to date examined p16 expression (a surrogate for HPV status) in both oropharyngeal and nonoropharyngeal head and neck cancers, with a survival advantage observed for p16 (HPV)-positive cancer in both groups.7 Finally, there is a paucity of information on whether the survival benefit associated with HPV16 infection varies with patient history of smoking and alcohol consumption (as opposed to smoking at the time of and following diagnosis, which adversely affects outcome15,16). Herein, we examined whether the magnitude of seropositivity to HPV16 E6 and E7 proteins is important for head and neck cancer survival, whether the association between seropositivity and survival extends to oral cavity and laryngeal cancers, and whether the clinical benefit associated with HPV16 is influenced by history of smoking and alcohol consumption.
Incident cases of head and neck squamous cell carcinoma (HNSCC) were enrolled through major teaching hospitals located in Boston, Massachusetts (Brigham and Women's Hospital, Beth Israel Deaconess Medical Center, Boston Medical Center, Dana-Farber Cancer Institute, Massachusetts Eye and Ear Infirmary, Massachusetts General Hospital, and New England Medical Center), as part of a population-based case-control study of head and neck cancer in the greater Boston area.17 Patients with a confirmed incident diagnosis of HNSCC were included if they were residents in the study area and 18 years or older. Patients with carcinoma in situ, or tumors originating in the lip, salivary gland, or nasopharynx/nasal cavity were excluded. Recurrent cases and incident cases diagnosed more than 6 months prior to contact were also excluded. The study includes data collected from 2 periods of recruitment from the same population: phase 1 was conducted between December 1999 and December 2003 (533 cases) and phase 2 was conducted between October 2006 and June 2011 (509 cases). All patients enrolled in the study provided written informed consent as approved by the institutional review boards of the participating institutions.
Tumors were classified as oral cavity, oropharynx, or larynx based on recommendations by the American Joint Committee on Cancer (AJCC). According to the AJCC, tumors at the base of the tongue were classified as oropharyngeal, and those located at the anterior of the tongue were classified as those of the oral cavity. The pathology reports for patients with carcinoma of the tongue were reviewed and further classified as located at the anterior or the base of the tongue. This review was conducted blinded to history of alcohol consumption, tobacco use, and HPV16 status. Oral cavity tumors corresponded to International Classification of Diseases, Ninth Revision (ICD-9), codes 143, 144, 145, and, if located at the anterior of the tongue, 141; oropharyngeal tumors corresponded to ICD-9 codes 146, 148, 149, and, if at the base of the tongue, 141; and laryngeal tumors corresponded to ICD-9 code 161.
Patients completed a self-administered questionnaire that was reviewed with study staff to obtain data on demographic characteristics, medical history, family history of cancer, detailed smoking and alcohol consumption habits, occupational history, and residential history. Survival time and vital status were collected for all patients from the social security death index and from active search of public databases, including telephone confirmation of vital status, retirement records, and work history.
Serum was separated from venous blood within 12 to 24 hours of blood drawing and stored at −80°C. To detect antibodies against HPV16 E6 and E7 proteins, a glutathione S-transferase capture enzyme-linked immunosorbent assay was used in combination with fluorescent bead technology.18,19 This assay detects HPV antibodies with high type specificity and demonstrates assay sensitivity similar to the gold-standard for L1-serology that uses viruslike particles (VLPs) as antigens.19 Median fluorescence intensity (MFI) values were dichotomized as antibody positive or negative. Standardized cutoffs for HPV16 early proteins were determined using serum of 117 female, HPV16 DNA-negative, self-reported virgins from a cross-sectional study among Korean students.20 The mean + 5 SDs was calculated, and the resulting cutoff was doubled to stringently separate seropositive and seronegative reactions (HPV16 E6: 484 MFI, and HPV16 E7: 1096 MFI21). Individuals were classified as HPV-positive if they were seropositive for either the E6 or E7 proteins of HPV16. This assay has been reported as highly reproducible for both E6 and E7 antigens.1
Immunohistochemical staining for p16 was conducted using a purified mouse anti-human p16INK4 monoclonal antibody kit (clone E6H4, prediluted 1:2; mtm laboratories AG) with heat-induced antigen retrieval in citrate buffer (pH = 6.0) using pressure cooker pretreatment with positive and negative cell lines (and normal mouse serum) as controls stained in parallel, as described in detail previously.22 The staining intensity of positive cells was scored by 1 pathologist blind to patient serology status and other data.
For each patient, all-cause 5-year survival was calculated from the date of diagnosis to death or censoring event (loss to follow-up or on achieving 5 years of follow-up). Univariate survival differences were assessed by Kaplan-Meier 5-year survival functions, and the difference between strata was assessed using log-rank tests. Multivariable Cox proportional hazards models were applied to calculate the hazard ratios (HRs) and corresponding 95% CIs while controlling for potential confounders: age (continuous), race (white or nonwhite), sex, education (high school graduate or less vs some college or more), pack-years smoked (continuous), average alcoholic beverages per week (continuous), study enrollment phase (1 or 2), site (oropharynx, oral cavity, or larynx), and stage (I, II, III, IV). Missing values were imputed using multiple imputation.23 The proportional hazards assumption was met in both the overall survival analysis, and the anatomic site-specific analyses. A Pearson Φ coefficient was used to test the correlation between 2 binary variables. Statistical analyses were performed with R software, and all tests were 2-sided with a α = .05 significance level.
In total, 1054 patients with HNSCC were evaluated for HPV16 E6/E7 seropositivity and all-cause survival; 75% were male, and their mean age was 59 years. Demographic traits by HPV16 status are presented in Table 1. Patients who were HPV16 seropositive were statistically younger and more likely to be male, to be more educated, and to have experienced lighter smoking and alcohol consumption. In addition, HPV16 seropositive patients were more likely to have tumors in the oropharynx and present with advanced-stage disease.
HPV16 seropositivity was very strongly associated with improved 5-year survival of head and neck cancers (HR, 0.33; 95% CI, 0.24-0.45). To evaluate a potential dose response relationship, seropositivity was categorized according to tertile of HPV titer level. There was no association between titer level and outcome (Figure 1); each tertile indicating seropositivity was associated with improved survival odds in a Cox model adjusted for age, race, sex, education, pack-years smoked, average alcoholic beverages per week, study enrollment phase, anatomic site, and stage (HR for low titer, 0.47 [95% CI, 0.31-0.71]; HR for moderate titer, 0.42 [95% CI, 0.26-0.68]; HR for high titer, 0.37 [95% CI, 0.22-0.60]). For all remaining analyses, seropositivity was treated as a dichotomous (positive/negative) variable.
Given the well-established relationship between HPV16 and tumors arising in the oropharynx, we evaluated whether the association between HPV16 serostatus and patient outcome varied across anatomic site (Figure 2). The association between HPV16 and survival was most prominent for the oropharyngeal cancers. In addition, there were statistically significant associations between HPV16 serostatus and patient survival for tumors of the oral cavity and larynx. These associations remained significant after adjustment for potential confounding variables (Table 2): oropharyngeal cancer HR, 0.26 (95% CI, 0.18-0.39); oral cavity HR, 0.45 (95% CI, 0.18-0.80); and larynx HR, 0.29 (95% CI, 0.10-0.85). We also evaluated the independent and joint effects of E6 and E7 seropositivity (see the eTable in the Supplement). Positivity to either E6 or E7 alone was associated with improved outcome (HR, 0.46 and 0.55, respectively), while those seropositive for both E6 and E7 had the best outcomes (HR, 0.25; 95% CI, 0.13-0.48).
A subset of 162 cases had information on both HPV16 serology and p16 staining in the tumor with a modest association observed between the 2 measures (φ-coefficient, 0.73). Considering all cases, only the serology measure was associated with patient survival, although with limited precision (HR for serology, 0.49 [95% CI, 0.23-1.04]; HR for p16, 0.84 [95% CI, 0.44-1.58]). As expected, in 56 patients with oropharynx tumors, both measures were strongly associated with outcome (HR for serology, 0.16 [95% CI, 0.03-0.47]; HR for p16, 0.16 [95% CI, 0.03-0.46]).
Finally, we investigated whether the association between HPV serostatus and outcome differed by history of smoking and alcohol consumption. Patient exposure was categorized as having either a light (lowest tertile for both smoking and alcohol consumption), moderate (middle tertile for at ≥1 exposure), or heavy (highest tertile for either smoking or alcohol consumption) (Figure 3). Again, in all subgroups HPV16 serostatus was associated with better patient survival. This remained true after adjustment for possible confounding variables.
Using the resources of a large, population-based study, we have demonstrated that a positive serologic response to HPV16 oncoproteins is associated with improved patient survival, an association that persists across anatomic sites and patient history of tobacco and alcohol exposure. These data are among the first to demonstrate a convincing relationship between HPV16 and improved patient survival for tumors of the larynx and oral cavity, and are consistent with the analysis of p16 expression and patient outcomes in patients with nonoropharyngeal cancer in head and neck cancer treatment trials.7 Although HPV contributes to a smaller fraction of disease at these sites, our data would strongly suggest that immune response to HPV16 infection has clear prognostic clinical significance, similar to and consistent with the data for oropharyngeal cancers. In addition, our data are among the first to examine the potential interaction between HPV and patient smoking and drinking history in head and neck cancer survival. Finally, these data demonstrate that it will be possible to evaluate HPV in head and neck cancer using a blood-based, rather than tumor-based, measurement.
There are limited prior data examining the association between HPV16 and outcome in the oral cavity, with the most directly comparable data observing a nonstatistically significant association between HPV16 seropositivity and improved outcome for oral cancer.24 However, it must be noted that this prior study was underpowered (with only 13 seropositive oral cancer cases). More recently, Chung et al,7 using data from several clinical trials, demonstrated that p16 expression for nonoropharyngeal was inversely associated with patient outcome, consistent with our observations. However, results for tumors originating in the oral cavity were not statistically significant (HR, 0.7; 95% CI, 0.33-1.47). This may be attributable to the poor correlation between p16 overexpression and HPV status in the oral cavity.7,25 In contrast, and in support of our approach, Anderson et al26 have described a positive association between serology and HPV tumor status. These results suggest that estimates of association that rely on p16 overexpression likely bias results, with potential implications for use of this methodology in clinical practice.
There are 2 additional reports27,28 that describe HPV16 as a poor prognosticator for oral cavity cancers, in direct contrast to both our data and the well-established association of HPV infection and improved outcomes for oropharyngeal cancer. The differences in our data compared with Lee et al27 and Duray et al28 may be attributable to multiple factors, including study design. Lee et al27 reported a marked reduction in patient survival for patients with HPV16-positive advanced-stage oral cavity cancer in South Asia, a population with an elevated rate of HNSCC. The high prevalence of betel quid chewing in their study population (85%), as well as differences in the distribution of age, smoking, alcohol habit, and, potentially, clinical approach to the patients, may contribute to the counterintuitive associations observed in their study. Furthermore, they used a polymerase chain reaction (PCR)-based method in assessing HPV status, and this has not universally been proven to be a sensitive and specific approach.29-31 Duray et al28 studied a Western population without the presence of betel quid exposure but with a high incidence of HNSCC (in Belgium), with a very high prevalence of smoking and alcohol consumption both in the HPV-positive and HPV-negative patients.28 They also used a PCR-based approach for HPV 16 detection, and this yielded a high-risk HPV prevalence of 44% in the oral cavity, much higher than our serology-based measure (17%), the recent analysis of TCGA tumors,13 and the prevalence reported in other published studies.32 Our results derive from a prospectively collected population-based study, also distinctly different than these 2 prior reports. Furthermore, our results are highly consistent with the established paradigm in the oropharynx. While the etiologic fraction of oral cancer attributable to high-risk HPV is low,33 our observation of a significant association between infection and survival of oral cancer indicates that HPV16 seropositivity in patients with oral cancer is of high clinical importance.
Less well-studied is the question of HPV16 infection and clinical outcome in laryngeal cancers. In the analysis of clinical trial data (n = 140 laryngeal cancers), there was a nonstatistically significant association between p16 and improved survival (HR, 0.54; 95% CI, 0.27-1.10),7 consistent with our data. In a larger collection of laryngeal cancers (n = 324) p16 staining was not associated with outcome,34 although that analysis was limited to a 2-year survival time that may not be sufficient to detect HPV-driven differences. Other small studies have reported either no association,35-37 or that HPV is associated with worse outcome.38 The choice of marker to assess HPV may be particularly problematic in the larynx: HPV DNA has been detected at high frequency in pathologically normal tissue of the larynx,39 and there is a documented poor correlation of HPV DNA detection in serial sections of laryngeal tumors.9
There is an urgent need in the oral oncology field to arrive at consensus on HPV testing methodology. p16 Immunohistochemistry (IHC) has been used in many seminal studies of oropharyngeal cancer, but misclassification of tumors clearly occurs, as the p16 gene can be altered by nonviral mechanisms. Among African Americans there is low concordance between p16 expression and HPV detection.8 In addition, the positive predictive value of p16 IHC outside the pharynx is low, presumably because tobacco carcinogens also disrupt the Rb pathway, mimicking the action of HPV16.28,33 Our analysis, which included both HPV and smoking, demonstrates that smokers without immune evidence of HPV infection (who would potentially be misclassified using the p16 measure) have the worst survival outcomes, supporting our notion that p16 is not the best marker for clinical use. Another common metric, detection of HPV DNA in tumors, is problematic, as Halec et al9 have demonstrated a lack of reproducibility in DNA-based measures in serial tumor sections.
Our study is not without limitations, and we specifically note that the lack of direct comparison with HPV RNA or DNA detection within tumor specimens limits our ability to completely assess HPV status. Furthermore, the lack of assessment of non-HPV16 subtypes is also a limitation in that we do not have complete high-risk HPV characterization.
Finally, an alternative interpretation of our results is that serology is not necessarily capturing solely HPV-DNA status of the tumor but rather an immune phenotype that is beneficial for cancer outcome. Viral exposures shape immunity, and this has been well documented for cytomegalovirus.40 It is not yet known how an HPV infection may shape and immunity and how that relates to surveillance for cancer recurrence. We posit that HPV is not only important as an oncogenic driver of certain head and neck cancers, but also as a modulator of immunity that has an impact on the ability of the immune system to identify and remove residual cancer cells.
Using a large case series of head and neck cancers from a defined geographic region we have demonstrated that serologic evidence of high-risk HPV infection is associated with outcomes in not just oropharyngeal cancers but also oral cavity and laryngeal cancer. There is an urgent need for additional research that applies a consensus approach to HPV16 detection in similar large, population-based samples to improve the statistical precision of our estimates of association. Importantly, this will allow for the inclusion of patients with oral cavity and laryngeal cancer in the now developing deescalation trials.
Corresponding Author: Heather H. Nelson, MPH, PhD, Masonic Cancer Center, Division of Epidemiology and Community Health, University of Minnesota, 420 Delaware St SE, MMC 806, Minneapolis, MN 55455 (firstname.lastname@example.org).
Accepted for Publication: August 19, 2016.
Published Online: December 8, 2016. doi:10.1001/jamaoncol.2016.4500
Author Contributions: Drs Kelsey and Nelson had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Pawlita, Kelsey.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Nelson.
Critical revision of the manuscript for important intellectual content: Pawlita, Michaud, McClean, Langevin, Eliot, Kelsey.
Statistical analysis: Nelson, Eliot.
Administrative, technical, or material support: McClean, Kelsey.
Study supervision: Pawlita, McClean.
Other—planned and performed the serological analysis, contributed all serological data used in the manuscript: Pawlita.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by the National Cancer Institute grants R01CA121147, R01CA100679, and R01CA078609 to Dr Kelsey for design and conduct of the study; collection, management, analysis, and interpretation of the data; and National Institute of Environmental Health Sciences grant T32ES07272 to Dr Langevin for analysis, and interpretation of the data. Dr Nelson was supported by the Minnesota Masonic Charities for analysis and interpretation of the data.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.