McIlwain WR, Sood AJ, Nguyen SA, Day TA. Initial Symptoms in Patients With HPV-Positive and HPV-Negative Oropharyngeal Cancer. JAMA Otolaryngol Head Neck Surg. 2014;140(5):441-447. doi:10.1001/jamaoto.2014.141
This study addresses the most common initial symptoms of oropharyngeal squamous cell carcinoma (OPSCC) and investigates differences between human papillomavirus (HPV)–positive vs HPV-negative tumors.
To analyze the most common initial symptoms in patients with OPSCC and to determine if any differences in initial symptoms occur between HPV-positive and HPV-negative tumors.
Design, Setting, and Patients
Retrospective single-institution review of medical records of previously untreated patients with OPSCC diagnosed from January 1, 2008, to May 20, 2013, who were evaluated by 1 physician (the senior author, T.A.D.) at the Medical University of South Carolina.
Main Outcomes and Measures
We determined the most common initial symptoms of OPSCC and analyzed differences between HPV-positive and HPV-negative tumors.
Neck mass (in 39 patients [44%]) and sore throat (in 29 patients [33%]) comprised the most common initial symptoms in OPSCC. Patients who were HPV-positive were more likely to initially notice a neck mass than HPV-negative patients (51% vs 18%; P = .02), whereas HPV-negative patients were more likely to notice sore throat (53% vs 28%; P = .09), dysphagia (41% vs 10%; P = .05), or odynophagia (24% vs 6%; P = .04).
Conclusions and Relevance
This study provides preliminary evidence supporting neck mass and sore throat as the initial symptoms of patients with OPSCC. Patients who were HPV-positive more commonly complained of a neck mass as the initial symptom, whereas HPV-negative patients more commonly had symptoms related to the primary tumor site, including sore throat, dysphagia, and/or odynophagia.
The incidence of oropharyngeal squamous cell carcinoma (OPSCC) has been steadily increasing, in contrast to the steady decline of other head and neck cancers.1- 5 This trend has been attributed to the increased incidence of human papillomavirus (HPV)–positive OPSCC,6,7 comprising 40.5% of cases of OPSCC before 2000 and up to 70% since 2009 (P < .001).3 The differing etiologies for OPSCC has been shown to affect 2 different patient populations with differing overall prognosis. Human papillomavirus–positive OPSCC affects younger, nonsmoking, white males and patients with more extensive sexual history.1,8- 12 Human papillomavirus–negative OPSCC tends to affect older patients with a history of heavy tobacco smoking and alcohol use, similar to squamous cell carcinomas (SCCs) affecting other locations.5,13,14 Patients with HPV-positive OPSCC have been proven to have a more favorable prognosis6,8,15,16 and quality of life17 than those with HPV-negative OSPCC.
Given that T stage in OPSCC correlates with prognosis, the best opportunity for cure is when tumors are asymptomatic and small (<2 cm) but visible or palpable, allowing recognition by an examining clinician. Early-stage carcinomas (stages I or II) typically require single-modality therapy, including surgery or radiation, whereas late-stage carcinomas (stages III or IV) often require multimodality treatment with combinations of surgery, radiation, and/or chemotherapy. As primary tumors enlarge (higher T stage), patients often experience increased morbidity with larger surgical resections requiring advanced reconstructive techniques. Historically, OPSCC has been diagnosed at an advanced stage, thus increasing morbidity from multimodality treatment and compromising survival. Many studies have reported on reasons that oral and oropharyngeal cancers are diagnosed at a late stage. For example, Rogers et al18 surveyed patients with oral and oropharyngeal cancer, and most considered their symptoms to be insignificant and/or something that would get better with time. Patients without insurance or those who had Medicaid coverage were associated with advanced cancer stage.19 Overall, a recent meta-analysis20 demonstrated that a delay in diagnosis is a moderate predictor of advanced stage at diagnosis.20- 22 Because early diagnosis and earlier stage tumors seem to be important for overall survival, patients and clinicians should be aware of opportunities for early presentation, biopsy, and treatment.
Many studies have investigated the clinical presentation of oral and oropharyngeal cancer with efforts to improve earlier diagnosis of this cancer. Mashburg and Meyers23 demonstrated that most asymptomatic oral cavity and oropharyngeal SCCs present with tumors smaller than 2 cm (stage T1), and the presence of symptoms (pain, referred otalgia, dysphagia, odynophagia, sore throat) suggested more advanced disease. Furthermore, Guggenheimer et al24demonstrated that patients with oral cavity or oropharyngeal cancer would not endorse symptoms until the primary tumors became large enough (>2 cm, or stage T2) to invade nearby structures. By this point, many of these tumors have already metastasized to regional lymph nodes, thus presenting at an advanced stage.5,16 Other studies have reported common oral cavity and oropharyngeal cancer symptoms.24- 27 For example, Pitchers and Martin21 reported the frequency of cumulative OPSCC symptoms, which most commonly consisted of the presence of a neck lump (49.3%) or sore throat (33.3%). Similarly, Gorsky et al26 reported that the most common base of tongue OPSCC symptoms included neck mass (25.7%), dysphagia (25%), ear pain (17.6%), sore tongue (15.4%), lump on tongue (11.8%), voice changes (3.7%), and bleeding (0.7%). However, many of these studies do not distinguish between oral cancer and oropharyngeal cancer symptoms,18,21,26- 30 and there has been increasing evidence that SCCs of these 2 anatomic locations are distinctly different in etiopathogenesis, behavior, and even responses to therapy.
To our knowledge, no study to date has reported the initial symptoms of oropharyngeal cancer, or differences in clinical presentation according to HPV status. The primary objectives of this study were to determine the most common initial symptom of patients with OPSCC and to determine if any differences in presentation occur based on HPV status.
From January 1, 2008, to May 20, 2013, we collected consecutive medical records of patients evaluated by 1 of us (the senior author, T.A.D.) at the Head and Neck Tumor Center Multidisciplinary Clinic, Hollings Cancer Center, Medical University of South Carolina, Charleston, with the following International Classification of Diseases, Ninth Revision (ICD-9) codes: 146.0-9, 141.0-9, 147, 145.3, and 145.5. Inclusion criteria included newly diagnosed, previously untreated OPSCC with documented HPV tumor status via p16 immunohistochemical analysis. Patients were excluded if they had recurrence or persistence of previously treated OPSCC, unknown HPV status, non-SCC of the oropharynx, oral cavity carcinomas coded with the ICD-9 codes noted herein, or insufficient medical records. The institutional review board at Medical University of South Carolina, Charleston, gave approval for the study.
Patient information was extracted from our electronic medical record system through health care providers’ notes on the patients’ initial encounter, referral letters, and a universal new-patient symptom checklist forms. The patients’ demographics extracted included age, sex, race/ethnicity, and tobacco and alcohol use history. Tobacco and alcohol use was defined as never (<1 year of use in lifetime), former (quit ≥1 year prior to cancer diagnosis), and current (within 1 year of diagnosis). The degree of tobacco use was defined in pack-years and defined as moderate (<30 pack years) and heavy (≥30 pack-years). Alcohol use was defined by average number of drinks per week (“light” indicated <1 drink or “social drinking,” “moderate,”<21 drinks; “heavy,” ≥21 drinks). Tumor characteristics include the location (tonsillar complex, base of tongue, posterior pharyngeal wall, or soft palate), HPV status, and clinical TNM stage. Human papillomavirus positivity was defined by the presence of the surrogate marker, p16, via immunohistochemical analysis. The diagnosis date was based on biopsy-proven SCC of the oropharynx.
The initial symptoms were extracted from all notes in the patient’s medical history. We defined the initial symptoms as the symptoms of longest duration when patient presented to his or her primary care physician. We used following terms synonymously when analyzing initial symptoms: neck mass/lump, difficulty swallowing/dysphagia, sore throat, painful swallowing/odynophagia, ear pain/otalgia, lump in throat/globus sensation, bleeding/blood-tinged sputum, change in voice/muffled voice/“hot potato voice”/dysphonia/dysarthria, and direct visualization/visualized mass. We defined visualized mass as a mass that the patient noticed initially. We recorded all symptoms from the most detailed health care provider’s note if specific delineation of which symptom came first was not apparent. For example, if a patient complained of “sore throat and neck mass for 3 months” on all health care provider notes, we recorded both “sore throat” and “neck mass.” We attempted to limit these discrepancies by looking at all health care providers’ notes and universal symptom checklist forms.
The difference between initial symptoms and HPV status were initially analyzed with χ2 testing with Yates correction or a Fisher exact test when appropriate. In addition, a Spearman correlation among variables (symptoms and HPV status) with the assumption of a monotonic relationship was used to show association between symptoms and HPV status. P ≤ .05 was considered to indicate a statistically significant difference for all tests.
A database consisting of 430 patients evaluated by the senior author was generating using the ICD-9 codes listed in the Methods section. From January 1, 2008, through May 20, 2013, 88 patients met inclusion criteria for newly diagnosed, previously untreated OPSCC with known HPV tumor status. Excluded patients (n = 342) included those with unknown HPV status (n = 57); tumor surveillance, persistence, recurrence, or prior treated HPV (n = 94); oral cavity involvement by SCC (n = 151); non-SCCs of the oral cavity and oropharynx (n = 28); and insufficient medical records (n = 12). As shown in Table 1, 71 of the patients (81%) were HPV-positive, and 17 (19%) were HPV-negative. The average age at diagnosis, sex, HPV-positive OPSCC, HPV-negative OPSCC, race/ethnicity, smoking severity, and alcohol use history are also described in Table 1.
The tumor characteristics, including location, T stage, N stage, and overall TNM stage, are shown in Table 2. Overall, the primary tumors were located in the tonsillar complex (46 tumors [52%]), base of tongue (34 [39%]), soft palate (6 [7%]), and the posterior pharynx (2 [2%]). In the HPV-positive cohort, the location encompassed the tonsillar complex (41 tumors [58%]), base of tongue (26 [37%]), soft palate (4 [6%]), and none in the posterior pharynx. In the HPV-negative cohort, the location encompassed the base of tongue (8 tumors [47%]), tonsillar complex (5 [29%]), soft palate (2 [12%]), and posterior pharynx (2 [12%]).
The most common initially reported symptoms (ie, neck mass, sore throat, dysphagia, visualized mass, globus sensation, odynophagia, otalgia) of patients diagnosed as having OPSCC are listed in Table 3. In the HPV-positive vs HPV-negative cohort, initial presenting symptoms were neck mass (51% vs 18%; P = .02), sore throat (28% vs 53%; P = .09), dysphagia (10% vs 41%; P = .005), visualized mass (14% vs 6%; P = .60), globus sensation (9% vs 12%; P = .81), odynophagia (6% vs 24%; P = .04), otalgia (8% vs 0; P = .48), nonspecific pain (5% vs 12%; P = .32), bleeding (1% vs 12%; P = .09), weight loss (1% vs 12%; P = .09), change in voice (2% vs 6%; P = .48), asymptomatic (2% vs 0; P = .99), and fatigue (1% vs 0%; P = .99).
In addition, further analysis was done with a Spearman rank correlation model to determine association between variables (Table 4). The following symptoms were positively or negatively associated with HPV positivity: neck mass (ρ= 0.263; P = .01), sore throat (ρ = −0.208; P = .05), dysphagia (ρ = −0.338; P = .001), odynophagia (ρ = −0.246; P = .02), bleeding (ρ = −0.225; P = .03), and weight loss (ρ = −0.225; P = .03).
The incidence of head and neck cancer has been steadily declining over time,1,3 likely associated with the decline in per-capita smoking consumption in the United States.4 However, the incidence of OPSCC is steadily increasing,31 a phenomenon explained by the rise in HPV-positive OPSCC.1- 4,6 Mehanna et al3 demonstrated that the prevalence of HPV-positive OPSCC has increased from 40% before 2000 to up to 72.2% from 2005 through 2009. Our patient cohort demonstrates similar incidences: 81% of our patients had HPV-positive OPSCC. Human papilloma virus–positive OPSCC has also been shown to affect a younger population.1,2,5,8 In 2005, Shiboski et al2 demonstrated an increase in tonsillar and tongue SCC incidence in younger white patients aged 20 to 44 years from 1973 to 2001. Our HPV-positive cohort contained a higher percentage of whites than nonwhites (93% vs 47%, respectively) compared with the study by Shiboski et al2; however, our study detected no statistically significant age difference between HPV-positive and HPV-negative patients with OPSCC (59 vs 60 years, respectively).
In a cross-sectional study as part of the National Health and Nutrition Examination Survey in 2009, the overall prevalence of oral HPV infection in the United States was 6.9%. Men had a significantly higher prevalence of oral HPV infection than women (10.1% vs 3.6%; P < .001),32 perhaps correlating with the higher incidence of men affected by HPV-positive OPSCC.1,5,31 Data from our patient cohort agree with this suggestion, as our HPV-positive cohort had a greater percentage of men than our HPV-negative cohort (94% vs 59%, respectively). In a case-control study, the odds of oral HPV infection among college-aged men increased with the number of oral-sex partners (P = .046 for trend) or open-mouthed kissing partners (P = .02 for trend), suggesting that oral HPV infection is likely sexually acquired.10In another case-control study, D’Souza et al9 demonstrated that OPSCC is associated with a higher lifetime number of oral-sex partners (≥6) (odds ratio [OR], 3.4 [95% CI, 1.1-8.8]) and a higher number of vaginal-sex partners (≥26) (OR, 3.1 [95% CI, 1.5-6.5]). Data on sexual history were not consistently available from our retrospective review. Regardless, extensive sexual history is an associated risk factor for HPV-positive OPSCC.
Several studies have mentioned incidence of presenting symptoms in oropharyngeal cancer. Pitchers and Martin21 reported the frequency of cumulative OPSCC symptoms, which consisted of the presence of a neck lump (49.3%), sore throat (33.3%), direct visualized mass (5.9%), incidental finding at an outside health care provider (2.9%), otalgia (2.9%), dysphagia (2.9%), and globus sensation (1.4%). Gorsky et al26 reported that the most common base of tongue OPSCC symptoms included neck mass (25.7%), dysphagia (25%), ear pain (17.6%), sore tongue (15.4%), lump on tongue (11.8%), voice changes (3.7%), and bleeding (0.7%). However, to our knowledge, no study has reported on the most common initial presenting symptoms of OPSCC. In our cohort, 77% of patients reported a neck mass and/or sore throat initially, with neck mass more frequent in HPV-positive patients and sore throat more common in HPV-negative patients. Other common initial complaints include dysphagia (14 patients [16%]), visualized mass (11 [13%]), globus sensation (9 [10%]), odynophagia (8 [9%]), and otalgia (6 [7%]) (Table 3). By the time patients were referred to a tertiary care academic center, they often had multiple complaints potentially coinciding with the extent of disease (data not shown).
According to HPV tumor status, neck mass was significantly more common in patients with HPV-positive OPSCC vs HPV-negative OPSCC (51% vs 18%, respectively; P = .02). Sore throat was more common in HPV-negative OPSCC vs HPV-positive OPCSS (53% vs 28%; P = .09). Although less common initial complaints, dysphagia and odynophagia were significantly more associated with HPV-negative OPSCC vs HPV-positive OPSCC (41% vs 10%; P = .005; and 24% vs 6%; P = .04, respectively). Although not statistically significant, patients were more likely to self-report and notice a visible mass initially with HPV-positive OPSCC than HPV-negative OPSCC (14% vs 6%; P = .60). Using a Spearman rank order correlation test to further analyze symptoms associated with HPV positivity, Table 4 shows a positive correlation of neck mass with HPV-positive OPSCC (ρ = 0.263; P = .01), whereas sore throat, dysphagia, odynophagia, bleeding, and weight loss were less likely to present as HPV-positive patients’ initial symptoms. Cantrell et al33 demonstrated that HPV-negative tumors were more likely to show radiographical evidence of invasion into adjacent muscle (26% vs 6%; P = .01), perhaps correlating with symptoms of HPV-negative tumors that seem to be associated with the primary tumor site in this study. Therefore, we were surprised that otalgia was not more commonly associated with HPV-negative tumors, and we hypothesize that this may be a “later” presenting symptom than odynophagia or dysphagia.
Oropharyngeal squamous cell carcinoma is often diagnosed at advanced stage,1,16,31 likely owing to the prolonged asymptomatic phase of these tumors. Only 12 of our patients with OPSCC (14%) were diagnosed at an early stage (stages I-II), whereas 76 (86%) were diagnosed at advanced stages (stages III-IVc) (Table 2). Mashberg and Myers23 demonstrated that 84% of asymptomatic patients with oral and OPSCC presented with stage T1 tumors (≤2 cm). Likewise, Guggenheimer et al24 demonstrated that most of these tumors are often asymptomatic until they grow large enough (> 2 cm) to ulcerate, become infected, cause pain, or encroach on nearby structures affecting their function.
The discrepancies of initial symptoms based on HPV status may be inherent to the nature of the 2 differing etiologies of OPSCC. Paz et al34 demonstrated that patients with HPV-positive OPSCC have a higher incidence of lymph node metastasis (P = .003). In our cohort, 49 (69%) of HPV-positive patients presented with N2 or N3 disease. Patients who were HPV-negative were more likely to present without nodal disease (N0) compared with HPV-positive patients (35% vs 17%, respectively). Furthermore, HPV-positive patients were more likely to present with T1 disease than HPV-negative patients (24% vs 6%, respectively). A study with a larger HPV-negative cohort would be necessary to validate these trends.
In our HPV-positive cohort, the extent of nodal disease and initial noticeable symptom of neck mass may relate to the current theory of the pathophysiologic characteristics of HPV-positive cancer. Human papillomavirus typically infects the crypts of the lingual and palatine tonsils.35,36 These crypts are lined with reticulated epithelium composed of an interrupted basal cell layer and basement membrane. This structure facilitates passage of antigens, lymphocytes, and antigen-presenting cells for defense of pathogens entering the upper aerodigestive tract. Consequently, HPV has a direct passage to bind to the heparin sulfate modalities of the basement membrane and infect nearby basal cells. Furthermore, the interrupted basement membrane facilitates cellular migration, possibly contributing to early regional nodal metastasis and perhaps earlier symptoms of neck mass in HPV-positive OPSCC.35,37 Cantrell et al33 recently demonstrated that HPV-negative tumors were more likely to demonstrate invasion of adjacent muscle (26% vs 6%; P = .01). In our study, symptoms of primary tumor invasion (sore throat, dysphagia, odynophagia, bleeding, weight loss) were associated with HPV-negative tumors. These clinical presentation findings perhaps are associated with the more locally invasive nature of HPV-negative tumors seen on imaging.33
With the significantly better survival rate in HPV-positive OPSCC, current trials are investigating “deintensification” strategies that reduce radiation and/or chemotherapy therapeutic regimens in an effort to decrease morbidity associated with multimodality therapy. Moreover, larger primary tumors and more extensive nodal disease require difficult and more involved surgical resections with subsequent advanced reconstructive techniques. This study provides preliminary evidence of the first symptoms of OPSCC with hopes of detecting these tumors at earlier stages.
The retrospective design of our study is its major inherent limitation. A detailed history for data accrual relied on proper and complete documentation of symptoms in the health care providers’ notes. To help alleviate this limitation, we used information from multiple health care providers’ notes and a universal symptom checklist form, and we excluded any potential patients with inconsistent reporting of symptoms. Of course, the patients’ awareness of symptoms is extremely variable, as some patients notice symptoms earlier than others and may or may not attribute a symptom to their diagnosis. Symptom overlap is also a concern. For example, “painful swallowing or odynophagia” is a more specific description than “sore throat,” but we recorded the data based on what the patient described as his or her first symptom. Although accurately describing specific symptoms, such as odynophagia vs dysphagia, relies on both the interviewer and patients’ understanding, the presence of a neck mass as a chief complaint is less likely to have overlap with other symptoms. Statistical analysis was limited owing to our small sample size of HPV-negative OPSCC, and a larger study with more HPV-negative tumors would be ideal now that HPV status testing is standard practice. To eliminate most of these biases, a prospective study with universal surveys during new patient encounter visits would render the most accurate results. Consideration should be given as clinical practice guidelines for primary care providers and dental health professionals for referral of patients with concerning symptoms of OPSCC if studies reveal this opportunity in the future.
It has been well documented that OPSCC effects 2 different populations with different prognostic indications based on HPV status. To our knowledge, this study is the first to report the most common initial symptoms in OPSCC and illustrates the differences in clinical presentation between HPV-positive and HPV-negative OPSCC. The most common initial symptoms of OPSCC are neck mass and sore throat. Patients with HPV-positive OPSCC are more likely to present with neck mass indicative of regional disease, whereas patients with HPV-negative OPSCC will most likely present with symptoms associated with primary tumor site, including sore throat, dysphagia, and/or odynophagia. The differences in clinical presentation between HPV-positive and HPV-negative OPSCC may better support the ongoing theory of how HPV-positive OPSCC tends to originate in the tonsillar crypts resulting in early cervical metastasis, and that HPV-negative OPSCC is more likely to be locally invasive correlating with initial symptoms at the primary tumor site. With the rapidly increasing incidence in OPSCC, it may be important to consider education of the general public about the early symptoms of OPSCC and to encourage primary care providers and dental health care professionals to have a high index of suspicion in patients with symptoms suggestive of OPSCC.
Submitted for Publication: September 6, 2013; final revision received December 4, 2013; accepted January 6, 2014.
Corresponding Author: Terry A. Day, MD, Department of Otolaryngology–Head and Neck Surgery, Medical University of South Carolina, 135 Rutledge Ave, Charleston, SC 29425 (firstname.lastname@example.org).
Published Online: March 20, 2014. doi:10.1001/jamaoto.2014.141.
Author Contributions: Dr Day 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: McIlwain, Sood, Day.
Acquisition of data: McIlwain, Sood, Day.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: McIlwain, Sood, Day.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: McIlwain, Sood, Nguyen.
Study supervision: Day.
Conflict of Interest Disclosures: None reported.
Additional Information: The abstract for this article was submitted to the Fifth World Congress–International Federation of Head and Neck Oncologic Societies and the Annual Meeting of the American Head and Neck Society; July 26-30, 2014; New York, New York.