Factors Related to Human Papillomavirus Positivity among Oral Cavity and Pharynx Cancers from Surveillance, Epidemiology and End Results (SEER) Program Data

Oral cavity and pharynx cancer (OCPC) incidence and mortality rates have been on upward trends for the past decade (1). It is estimated that there will be 54,000 new OCPC cases identified in the United States and an estimated 11,230 OCPC deaths in 2022 (1). The 5-year relative survival rate for those diagnosed with OCPC is 68.0%, with earlier-stage diagnoses having more favorable survival rates (SEER*Explorer, 2022, SEER RRID: SCR_003293). OCPC disparities are apparent with males having an incidence more than double that of females (17.4 per 100,000 males vs. 6.4 per 100,000 females) and Blacks having a lower 5-year survival rate compared with Whites (52.4% vs. 69.2%, respectively; SEER*Explorer, 2022, SEER RRID: SCR_003293).

The human papillomavirus (HPV) has consistently been identified as a common OCPC risk factor (2–4). OCPC incidence, mortality, and survival rates vary by HPV status. HPV-positive (HPV+) OCPC diagnoses typically occur at later stages and need more drastic treatments as currently there is no standard screening protocol nor pharmacotherapeutic treatment for the earlier presence of an oral HPV infection (5). The number of HPV+ OCPC cases has increased by almost 50% since 1973 (6–9), occurring most often in males who are typically not heavy tobacco users (10). HPV+ OCPC mortality rates have been on a similar trajectory as HPV+ OCPC incidence rates, experiencing equivalent rate increases over time (8). The survival rate from an HPV+ OCPC diagnosis can be as high as 71% to 93% for those in lower risk groups (i.e., less tumor extension, no lymphnode involvement, non-smoking; ref. 5). Surprisingly, higher socioeconomic status (SES) has been associated with higher risk of HPV+ OCPCs (10). This unique disparity may offer some insight into the improved survival rate for HPV+ OCPCs, compared with those diagnosed with HPV negative (HPV−) OCPCs (10).

Only cancers from very well-defined specific OCPC anatomies have been routinely HPV tested and investigated (11–24). Specifically, HPV has been associated with more than 70% of all U.S. cancers found in the oropharyngeal region of the head and neck (25), predominately the tonsils and base of the tongue (3, 5, 6, 24, 26). However, the oral cavity and pharynx anatomic region is large with OCPCs occurring from the nasopharynx through the hypopharynx, with incidence, mortality, and survival rates also differing across anatomic sites (27). OCPC causal pathways are complex and multifactorial, with causes (e.g., HPV, tobacco/alcohol) potentially interacting with one another (i.e., independently, complementarily, or synergistically). Therefore, OCPC sites thought to be related to other risk factors (e.g., alcohol/tobacco) could also be HPV+ (28, 29). Additional OCPC anatomies also remain un(der)investigated and could be associated with HPV positivity (5, 30). For instance, a national study of cancer registry–representative tumors was HPV tested, and HPV was identified in 32.0% of oral cavity tumors—an OCPC site thought to not be HPV-related (24). Non-oropharyngeal OCPC cases may have different risk factors for HPV positivity (11, 18, 23) than those for oropharyngeal OCPCs (13–16, 19, 23), but research is limited as prior studies have concentrated on better understanding the latter.

Much of the existing national-level epidemiologic findings on OCPC cases with known HPV status are outdated and/or restricted in generalizability, relying on specific OCPC anatomic sites or populations (11–23). Moreover, sociodemographic factors reflecting neighborhood SES (nSES) and rurality/urbanity, potentially related to OCPC diagnoses, are untapped risk factors that may hold some of the keys to understanding differences in HPV+ and HPV− OCPCs. Given the significant increases in incidence rates of HPV+ OCPCs (3, 6–9, 31), this lack of relevant epidemiologic information is problematic as these risk factors could elucidate important causal pathways and identify high-risk populations which may aid disease control. Therefore, the objective of this investigation was to evaluate associations between known HPV status (i.e., from HPV testing) and demographic, census tract-level, and clinical factors among OCPC cases, overall, and for oropharyngeal and non-oropharyngeal sites independently, using data from a large, well-described collection of central cancer registries.

The Surveillance, Epidemiology and End Results (SEER) Program is a collection of population-based central cancer registries capturing cases from 22 registries across the United States. With permission from the SEER Program, the SEER Head and Neck with HPV Status and Census Tract-level SES/Rurality Combined Database was used to explore associations between known HPV status and demographic, census-tract level (nSES and rurality/urbanity), and clinical characteristics (SEER*Stat, December 2021, SEER RRID: SCR_003293). Institutional Review Board approval was not required, and ethical consent was not warranted because SEER data are publicly available and deidentified.

Invasive OCPC cases diagnosed and reported to SEER Program registries between 2014 and 2017 were included in this investigation (N = 25,357). Participants were excluded if they had missing socioeconomic data (i.e., nSES level, n = 1,535) or had a reporting source of death certificate only or autopsy only (n = 199). In addition, participants under 30 years old were excluded from analyses (n = 223) as their OCPCs are typically of a unique oral HPV-related disease presentation (i.e., juvenile onset recurrent respiratory papillomatosis) with a nonmodifiable transmission pathway (i.e., vertical; ref. 32).

SEER-defined sociodemographic variables were investigated as potential risk factors for HPV positivity in OCPC cases, overall, and more specifically in oropharyngeal and non-oropharyngeal OCPC sites. With HPV test status as the outcome of interest, OCPC cases were classified into three status groups based on their recorded HPV status—(i) HPV+; (ii) HPV−; and (iii) unknown. (As the SEER Program includes a collection of central cancer registries, HPV positivity was determined independently by reporting sources with varying HPV testing methods.) Sex was “Male” or “Female” and age at diagnosis was described using 10-year incremental categories (i.e., 30–39, 40–49, 50–59, 60–69, 70–79, 80+ years old). Race was described using the variable “Race Recode (W, B, AI, API)”, which classified race as “White”, “Black”, “Asian or Pacific Islander”, “American Indian or Alaskan Native”, or “unknown”. Ethnicity was dichotomized as “Spanish-Hispanic-Latino” or “non-Spanish-Hispanic-Latino”, using the variable “Origin Recode NHIA”. Marital status was defined as “divorced/separated/widowed”, “married/domestic partnered”, “single” or “unknown”. SEER's nSES, or Yost index, is a time-dependent composite score, including census tract-level information about education index, household income, percent below 150% of poverty line, median house value, percent unemployed, median rent, and percent working class variables (33–35). nSES scores were classified into quintiles, with the first and fifth quintiles describing the lowest and highest nSES groups, respectively. SEER census tract-level rurality data come from the U.S. Department of Agriculture's Rural Urban Commuting Area (RUCA) codes (36), using four levels (“All Urban”, “Mostly Urban”, “Mostly Rural”, “All Rural”).

SEER's clinical variables provided additional details on tumor site, histology, stage at diagnosis, first malignancy, diagnostic confirmation, and reporting source. The SEER variable “Site Recode”, based on International Classification of Disease for Oncology version-3 (ICD-O-3) topography codes (37), was used to determine specific anatomies within the category of OCPC (i.e., OCPC site), including “tongue” (ICD-O-3: C019), “tonsil” (ICD-O-3: C024, C090-099), “oropharynx” (ICD-O-3: C100-109), “nasopharynx” (ICD-O-3: C110-119), “pyriform sinus” (ICD-O-3: C129), “hypopharynx” (ICD-O-3: C130-139), “gum and other mouth” (ICD-O-3: C051, 052), and “other oral cavity and pharynx” (ICD-O-3: C140-148). To be comprehensive in the investigation of all potential OCPC risk factors for HPV positivity, all OCPC sites included in the dataset, as defined by the SEER “Site Recode” variable, were included in the current investigation. [It should be noted that SEER has restricted the dataset to only OCPC sites believed to be potentially associated with HPV instead of including all OCPC (sub)sites (e.g., “tongue” only included base of tongue (ICD-O-3: C019), not non-base subsites (ICD-O-3: C020-023, C025-029); therefore, the category is referred to as “base of tongue” hereafter]. For additional analyses, OCPC sites were later recategorized as oropharyngeal (i.e., base of tongue (ICD-O-3: C019), tonsil (ICD-O-3: C024, C090-099), oropharynx (ICD-O-3: C100-109), gum and other mouth (ICD-O-3: C051, C052), other oral cavity and pharynx (ICD-O-3: C140, C142, C148)) and non-oropharyngeal [i.e., nasopharynx (ICD-O-3: C110-119), pyriform sinus (ICD-O-3: C129), hypopharynx (ICD-O-3: C130-139)] (38). Histology was dichotomized as squamous cell neoplasms (ICD-O-3: 8050-8076, 8078, 8083, 8084, 8094) and non-squamous cell neoplasms (ICD-O-3: 8000-8049, 8090-8389, 8430-8579, 8800-8839, 8850-8999, 9120-9169, 9180-9249, 9350-9379, 9490-9529). Stage at diagnosis was characterized using “SEER Combined Summary Stage” and was classified as “Localized,” “Regional,” “Distant,” and “Unknown/unstaged.” As indicated by the name, SEER's “first malignant primary indicator” variable collected data on whether the current OCPC diagnosis was the individual's first malignant tumor or not. Diagnostic confirmation described how diagnoses of OCPC were verified, ranging from clinical diagnosis to positive cytology/histology/radiography [with(out) microscopic confirmation] to unknown methodology. The origin of OCPC diagnosis was described by the SEER variable “type of reporting source” which was dichotomized as either “hospital or clinic” or “other.”

Descriptive statistics [i.e., frequencies (percentages)] for sociodemographic and clinical characteristics were estimated for OCPC cases by HPV status (+, −, or unknown). Differences in missing data (i.e., HPV status) were determined by a 10% difference in the distribution of each characteristic between those with known and unknown HPV status. Unknown HPV status and histologically non-squamous cell neoplasm OCPCs were then removed such that the analysis set only included those with known (i.e., positive or negative) HPV status and squamous cell neoplasms to best assess factors associated with OCPC HPV positivity. Univariate logistic regression models were used to determine whether characteristics were associated with HPV+ status. All variables that were statistically significant (P < 0.05) univariately were included in a multivariable logistic regression model to identify characteristics and populations that continued to be associated with HPV positivity. This same approach was used to identify risk factors for HPV positivity for oropharyngeal and non-oropharyngeal OCPCs separately. ORs (univariate and adjusted) and 95% confidence intervals (CI) were calculated for all models. Variables in which almost all cases were concentrated within one category even after regrouping (e.g., diagnostic confirmation), there was no meaningful contribution to determining HPV status (e.g., year of diagnosis), or there was nonsignificance or collinearity with other variables of interest (e.g., RUCA) were omitted. Analyses were conducted using SAS 9.4 (SAS Institute Inc., SAS RRID: SCR_008567) and Stata 16 (StataCorp, Stata RRID: SCR_012763).

The data analyzed in this study were obtained from the SEER Program at www.seer.cancer.gov.

Of the 23,400 OCPCs’ meeting inclusion criteria, 42% (n = 9,832) were HPV+, 18% (n = 4,211) were HPV−, and 40% (n = 9,357) were of unknown HPV status. (Because SEER has restricted the dataset to only OCPC (sub)sites believed to be potentially associated with HPV instead of including all OCPC sites, prevalence of HPV positivity among all U.S. OCPC cases cannot be concluded.) Most tumors were diagnostically confirmed with positive histology (93%). Table 1 shows the breakdown of demographic, census tract-level, and clinical factors according to known HPV status. Overall, most OCPC cases were male, 50 to 69 years old, White, non-Spanish-Hispanic-Latino, and married/domestic partnered. They also tended to reside in urban census tracts and census tracts in the highest (i.e., fifth) nSES quintile. Clinically, most OCPCs were diagnosed on the base of tongue or tonsil, as squamous cell neoplasms, at the regional stage, by a hospital or clinic, and as first malignancies (Table 1). Cases of unknown HPV status had similar sociodemographic and clinical attributes as cases with known HPV status, except for stage at diagnosis which differed by more than 10% (Supplementary Table S1).

When OCPC cases with unknown HPV status (n = 9,357) and histology of non-squamous cell neoplasms (n = 308) were removed to more accurately assess factors associated with OCPC HPV positivity, 13,735 cases with a known HPV status remained [n = 4,009 (HPV−), n = 9,726 (HPV+)]. Because of the high prevalence of HPV+ OCPCs, ORs should not be interpreted as prevalence ratios as they can overestimate the strength of associations. Univariately, the odds of OCPCs being HPV+ were significantly higher for males, those of the five youngest age groups (ref = 80+ years old), those residing in census tracts in each of the four highest nSES quintiles [ref = lowest quintile (i.e., group 1)], those diagnosed with an OCPC in the sites of nasopharynx, oropharynx, other oral cavity and pharynx, base of tongue, and tonsil (ref = hypopharynx), those diagnosed at the regional, distant, and unknown/unstaged stage (ref = local stage), and those for whom this was the first malignancy (Table 2). Most notably, the odds of HPV positivity in OCPCs were two to three times higher for males, those ages 40 to 69 years (vs. those ages 80+ years), those diagnosed at the regional stage (vs. local), and those with first malignancies. The odds of being HPV+ was more than nine times higher for OCPCs on the base of tongue or tonsils than those of the hypopharynx (Table 2). Univariately, odds of OCPCs being HPV+ were significantly lower for American Indians/Alaska Natives and Asians or Pacific Islanders (ref = White), those of Spanish-Hispanic-Latino ethnicity, those with any of the three non-married/domestic partnered marital statuses, and those with an “other” reporting source (ref = hospital or clinic). Interestingly, odds of OCPCs being HPV+ were at least 60% lower among American Indians/Alaska Natives and Asians or Pacific Islanders as compared with Whites (Table 2). On the other hand, neither residence in a rural or urban census tract was significantly associated with OCPCs being HPV+ (Table 2).

Generally speaking, the univariate associations were maintained in the multivariable model for each sociodemographic and clinical characteristic, when holding all other variables constant. In other words, males, those of the five youngest age groups (i.e., 30–79 year olds), those residing in census tracts in each of the four highest nSES quintiles (i.e., groups 2–5), those diagnosed with an OCPC in the sites of nasopharynx, oropharynx, other oral cavity and pharynx, base of tongue, and tonsil (ref = hypopharynx), those diagnosed at regional, distant, and unknown/unstaged stage at diagnosis (ref = local stage), and those for whom this was the first malignancy retained higher odds of their OCPCs being HPV+. Younger cases (i.e., 30–59 year olds) had at least two times the odds of their OCPCs being HPV+ while older cases (i.e., 60–79 year olds) only had an odds for HPV positivity that was 70% to 92% higher (ref = 80+ year olds). OCPCs of the base of tongue and tonsils were still more than seven and nine times, respectively, likely to be HPV+ than those of the hypopharynx. Compared with those residing in census tracts in the lowest nSES quintile (i.e., group 1), the odds of OCPCs being HPV+ continued to increase through nSES quintiles, suggesting a dose–response relationship (Table 2). On the other hand, American Indians/Alaska Natives and Asians or Pacific Islanders (ref = White), those of Spanish-Hispanic-Latino ethnicity, those with any of the three non-married/domestic partnered marital statuses, and those with an “other” reporting source (ref = hospital or clinic) retained lower odds for their OCPCs being HPV+ (Table 2). More specifically, OCPCs among single and divorced/separated/widowed individuals had a significantly lower odds (32% and 22%, respectively) to be HPV+ than those of their married/partnered counterparts (Table 2).

There were 12,281 oropharyngeal cases [n = 3,021 (HPV−), n = 9,260 (HPV+)] and 1,454 non-oropharyngeal cases [n = 988 (HPV−), n = 466 (HPV+)]. All univariate and multivariable associations for HPV+ OCPCs in general, including directionality and strength of associations, were maintained for oropharyngeal cases (Table 3). For non-oropharyngeal OCPCs, only age (i.e., 40–69 year olds), stage at diagnosis (i.e., regional stage only), and first malignancy were univariately associated with higher odds of HPV positivity; race [i.e., American Indian/Alaska Native (borderline significant), Asian or Pacific Islander] was the only factor with a lower odds of HPV positivity (Table 4). In the multivariable model, age and race were the only significant risk factors. More specifically, non-oropharyngeal OCPC cases ages 40 to 69 years had approximately two times higher odds of HPV positivity than 80+ year olds with non-oropharyngeal OCPCs, after adjusting for all other variables in the model. American Indian/Alaska Native Asian or Pacific Islander cases had 34% and 47% lower odds of HPV positivity compared with Whites with non-oropharyngeal OCPCs, respectively (Table 4).

Both in the univariate and multivariable OCPC models, males, those of the five youngest age groups (i.e., 30–79 year olds), those residing in census tracts of every nSES quintile except the lowest quintile (group 1), those diagnosed with any OCPC site except hypopharynx or gum and other mouth, those diagnosed with any stage at diagnosis except local, and those with first malignancies had significantly higher odds of OCPCs being HPV+. American Indians/Alaska Natives or Asians or Pacific Islanders (vs. Whites), Spanish-Hispanics-Latinos, non-married/partnered cases, and those cases from “other” (vs. hospital or clinic) reporting sources had significantly lower odds of OCPCs being HPV+ compared with their respective referent groups. When investigating as oropharyngeal and non-oropharyngeal subgroups, OCPC risk factors for HPV positivity held for oropharyngeal OCPCs, while age and race were the only significant factors for non-oropharyngeal OCPCs. (Because of the high prevalence of HPV+ OCPCs, ORs should not be interpreted as prevalence ratios as they can overestimate the strength of associations.)

Sociodemographic and clinical factors found to be associated with OCPCs, and more specifically, oropharyngeal OCPCs, being HPV+ were in alignment with previous findings. Earlier investigations concluded that males (9, 10, 12–17, 19, 21–23, 26, 39), non-Hispanics (15–17, 40, 41), Whites (9–13, 15, 17, 21–23, 26, 39, 40, 42, 43), younger ages (13–16, 18, 19, 23, 26, 40), and individuals from higher socioeconomic environments (14, 15, 17, 40) were more likely to have HPV+ OCPCs than HPV− OCPCs. Prior research also determined that HPV+ head and neck cancers were less common among non-White races (14, 19, 20), non-married/single individuals (14, 16, 21, 23), and older ages (14, 15, 19, 22, 23). Clinically, other studies have found that HPV+ OCPCs were first identified at more advanced stages (23, 43) and more likely to occur in the base of tongue/oropharyngeal region (14, 15, 22, 23). Non-oropharyngeal OCPC sites being associated with HPV positivity herein is likely due in part to the (complementary, synergistic) interaction between HPV and tobacco/alcohol use [e.g., predominantly among low SES (10), rural residents (44)] risk factors, often whereby tobacco/alcohol use is an underlying risk factor for HPV-related OCPCs (3, 24, 45). Rurality has also been associated with HPV− OCPCs (46). HPV− OCPCs typically occur more frequently among those with lower SES (10), are diagnosed at earlier stages than HPV+ cases (5, 47), and identified in the oral cavity, mostly the front of the mouth (lips, buccal mucosa, gingiva) where tobacco/alcohol chemical exposure contact is greatest (26, 27).

Findings from the current non-oropharyngeal OCPC analysis were also mostly supported by prior research. Prior studies have also found that younger non-oropharyngeal OCPC cases (<70 years old) were more likely to have HPV+ OCPCs (11, 18, 23). As herein, minority racial status has been shown in the literature to be protective against an HPV+ status with Whites being more likely to have an HPV+ non-oropharyngeal OCPC (11, 18). Sex has generally not been identified as a risk factor for non-oropharyngeal OCPC cases, aligning with our findings (11, 18). Only when studies restricted their investigations to hypopharynx cases did their evidence counter our conclusions—age (no association; ref. 11), race (no association; ref. 18), sex (females: less common; ref. 23). Marital status has been investigated minimally with findings being inconclusive—one showing no association (ref. 11; like the current study) and one finding a positive association (23). Clinically, stage has not been related to HPV status of non-oropharyngeal OCPCs elsewhere either (11, 18).

Findings among sociodemographic and clinical factors being (potentially) associated with OCPC HPV positivity can be different depending on whether HPV status is known or assumed due to anatomic and histologic characteristics of the OCPC when HPV status is unknown. In a prior investigation using SEER datasets which did not contain HPV status, it also observed that HPV-related (vs. non–HPV-related) OCPCs tended to occur among those of younger ages, males, non-Hispanics, Whites, and at advanced stages at diagnosis (48). Without HPV status known and only HPV status anatomically/histologically assumed, the prior study (48) found no association with nSES and that HPV-related OCPCs were more common in non-urban census tracts. Such findings are in direct conflict with our current findings of a positive association between increasing nSES and HPV+ and no association between urban residence and HPV+, among OCPCs in general and more specifically, oropharyngeal OCPCs. Confirmatory HPV testing of OCPCs in the current SEER dataset supply strong confidence in the findings presented herein and reiterate the need for future studies to rely on datasets with known HPV status.

This study was strengthened by the utilization of SEER Program data, which covers a substantial racially and ethnically diverse U.S. population that aids in the generalizability of this investigation's findings to other U.S. OCPCs (49). OCPC HPV status was reported in the dataset, avoiding the need to make HPV status conjectures based on OCPC anatomic and/or histologic characteristics. Inclusion of all SEER-identified OCPCs and further refinement of analyses to oropharyngeal and non-oropharyngeal OCPC sites independently allowed for a more complete investigation on the topic of OCPC risk factors for HPV positivity. Residential census tracts of SEER cases were linked to various nSES measures, including primary SES elements of income, education, and employment. The Yost index, used to assess nSES, has face validity, is clear and easily replicated (50), and, compared with county-level data, census tract-level data more closely align with neighborhoods, allowing for more precise nSES estimations [SEER*Stat, SEER Head and Neck with HPV Status and Census Tract-level SES/Rurality Combined Database (December 2021), SEER RRID: SCR_003293]. Reliance on RUCA Codes, also census tract-level data (vs. county-level data), for rurality/urbanity determination also contributed to a more complete assessment of investigated OCPC cases.

The primary limitation was the amount of missing and/or unknown HPV status data, especially in the earlier years of the dataset. To ensure the accuracy of our analyses, we restricted cases to the diagnosis years of 2014 to 2017 whereby more than 60% of HPV statuses were known. Assessing missing data, stage at diagnosis was the only factor substantively different between cases with known and unknown HPV status. The multivariable model was adjusted for all known possible confounders, including stage at diagnosis; however, bias might be introduced because of unknown confounder(s). Additional cases were excluded because of missing nSES data, but post hoc analyses indicated there were no significant changes in OCPC HPV positivity associations as a relatively large sample size was maintained.

This study also had other limitations. OCPC HPV positivity was evaluated by sources reporting information to the SEER Program; therefore, there was likely heterogeneity in HPV testing approaches across reporting sources. However, SEER has strict reporting guidelines in place and is known for complete and quality data so there is minimal concern with the validity of the HPV status data herein (49). Another restriction of the dataset is SEER's inclusion of only OCPC (sub)sites believed to be potentially associated with HPV instead of including all OCPC (sub)sites. To maximize use of the dataset and produce the most comprehensive results possible, we included all OCPC (sub)sites included in the dataset in our analyses whereas prior studies further limited their investigations to OCPC anatomic sites of interest (11–13, 16–20, 22, 23). Together, categories of combined anatomic sites of the SEER “Site Recode” variable (e.g., other oral cavity and pharynx, gum, and other mouth) coupled with unknown HPV testing methodologies of these sites potentially created an additional weakness of the dataset. In addition, SEER Program data do not include details concerning comorbidity and patient behavior, including OCPC screening behavior, and this is another limitation of the dataset, impeding the adjustment for confounding and/or effect modification by these factors. nSES and RUCA categories were assessed at one point in time; neighborhood status may vary over time. Furthermore, as area-based measures, it is possible that nSES and RUCA categories may not accurately describe individual SES and rurality/urbanity statuses. Finally, reliance on prevalence ORs with a high prevalence of HPV+ OCPCs suggests that ORs should not be interpreted as prevalence ratios. [Prevalence values are contained within (prevalence) ORs and prevalence ratios and therefore would be inaccurate to interpret in place of ORs or prevalence ratios.] In such situations of high disease prevalence, (prevalence) odds ratios can overestimate the prevalence ratio by at least 10%, especially when the intent is to approximate a risk ratio by an OR (51).

In conclusion, there are distinct sociodemographic and clinical differences in OCPC cases by HPV status, especially among oropharyngeal OCPCs, but less so among non-oropharyngeal OCPCs. An accurate understanding of epidemiologic differences in OCPC cases, namely oropharyngeal OCPCs, by HPV status is imperative so that high-risk populations are not ignored and OCPC incidence and mortality rates do not go unchecked. An HPV vaccine exists and can prevent much of the oral HPV infections (i.e., 88%) that cause OCPCs (52), but the prevention measure is underutilized (i.e., Healthy People 2020 goal for 13–17 years old adolescents: 80% vaccinated; United States: 59% up-to-date; refs. 53–55). Given that HPV-related OCPCs tend to occur later in life, and we have not reached herd immunity with current vaccination rates, HPV-related OCPCs may continue to increase until at least 2050, assuming HPV vaccination rates can be rapidly scaled up between now and then (56). In the meantime, health care providers and public health professionals can use this (oropharyngeal) OCPC risk factor for HPV positivity information to more accurately and effectively direct their messages and impact behaviors to reduce the OCPC risk within specific populations. For instance, public health efforts to minimize the oropharyngeal OCPC burden may target males, Whites, younger ages (especially 30–59 year olds), higher nSES groups, and those at higher risk for tonsil and base of tongue OCPCs while also encouraging more frequent OCPC screenings to prevent primary malignancies and later stage diagnoses. They should also consider intervening with awareness messages with Whites and/or those who are 40 to 69 years old to keep non-oropharyngeal OCPCs from increasing significantly. Health care providers can formulate patient risk assessments by evaluating significant sociodemographic and clinical risk factors identified in this study within their patient populations. And although a preliminary finding, clinicians may want to consider testing non-oropharyngeal (e.g., nasopharynx) OCPCs for HPV instead of assuming them to be HPV− as traditionally assumed, especially if patients exhibit additional OCPC risk factors for HPV positivity identified herein (i.e., 40–69 years old). Collectively, public health targeting and provider tailoring of educational efforts on these [(non)oropharyngeal] OCPC risk factors for HPV positivity may work together with repetitive consistent messaging at both the community and individual levels to maximize our potential efforts at preventing these cancers from reaching epidemic proportions in the United States.

K.H. Jordan reports grants from NCI/NIH during the conduct of the study. E.D. Paskett reports grants from Merck Foundation, Pfizer, Genentech, and Guardant Health outside the submitted work. No disclosures were reported by the other authors.

K.H. Jordan: Conceptualization, funding acquisition, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing, and data interpretation. J.L. Fisher: Conceptualization, data curation, formal analysis, investigation, methodology, writing–original draft, writing–review and editing, and data interpretation. J.A. Stephens: Data curation, formal analysis, investigation, methodology, writing–review and editing, and data interpretation. E.D. Paskett: Conceptualization, resources, supervision, funding acquisition, writing–review and editing.

The NIH's NCI generously supported this research (P01 CA229143-S1 for K.H. Jordan), but the funders did not have any other role in this study.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.