PURPOSE:
Human papilloma virus (HPV) association remains one of the most important predictors of clinical outcome in oropharyngeal squamous cell carcinoma (OPSCC). We aimed to determine whether the relationship between HPV status and overall survival was influenced by socioeconomic factors.
MATERIALS AND METHODS:
Using the National Cancer Database, we examined the relationship between socioeconomic status and overall survival, controlling for demographics and socioeconomic variables (age at diagnosis, race, sex, clinical stage, facility type, facility location, insurance status, median-income quartiles, percent of no high-school education quartiles, rural-urban dwelling, Charlson-Deyo score, primary site, and treatment type).
RESULTS:
HPV-positive patients with private insurance have improved overall survival compared with HPV-positive patients who are uninsured (hazard ratio [HR], 0.51, 95% CI, 0.41 to 0.63, P < .001). HPV-negative patients with private insurance have improved overall survival compared with HPV-negative patients who were uninsured (HR, 0.62, 95% CI, 0.53 to 0.73, P < .001). HPV-positive patients living in the south had improved overall survival compared with HPV-positive patients living in the west (HR, 0.83, 95% CI, 0.72 to 0.96, P = .013). As assessed through interaction, relationships between survival and insurance (P = .004), rural-urban status (P = .009), and facility location (P = .021) statistically differed between HPV-positive and HPV-negative patients.
CONCLUSION:
HPV status impact on overall survival for patients with OPSCC is influenced by socioeconomic factors including insurance status and treatment facility. A deeper understanding of these interactions is needed to improve equity of care for patients with OPSCC.
INTRODUCTION
As the sixth most common cancer worldwide with nearly 650,000 people affected each year, squamous cell carcinoma of the head and neck is on the rise largely as a result of the increased incidence of human papilloma virus (HPV)-related oropharyngeal squamous cell carcinoma (OPSCC), with nearly 63,850 new cases a year in the United States alone.1-3 Oropharyngeal cancer (OPC) has now surpassed cervical cancer as the most prevalent HPV-related cancer in the United States.4 HPV continues to be a strong, independent risk factor determining prognosis.2,3 HPV positivity typically portends a better prognosis for OPSCC, with an observed greater sensitivity to chemotherapy and radiation compared with HPV-negative disease.3,5-7
Despite the role of biologic factors such as HPV in determining clinical outcomes, socioeconomic factors have differed between patients with HPV-related OPSCC and patients with HPV-unrelated disease.7-13 Socioeconomic disparities confer differing survival outcomes in a multitude of malignancies, and these aspects may contribute to a delay in diagnosis in OPC.14 Previous studies have reported an association between private insurance and high income in patients with OPC, leading to a more expedited initiation of therapy compared with patients with no insurance or lower baseline income.4 Moreover, HPV-positive OPC patients are more likely to be middle-aged or younger and have a moderate- to upper-income level,8,12-14 whereas HPV-negative OPC patients are generally older at diagnosis and are of lower socioeconomic status.8-14 OPSCC is a disease with one anatomic location, two distinct biologies (HPV-related and HPV-unrelated), and distinct socioeconomic circumstances, all interacting to influence patient outcome.
As OPSCC provides an excellent opportunity to examine the interaction of biologic and nonbiologic factors in defining inequities of care, we aimed to determine whether the relationship between socioeconomic factors and overall survival varied depending on HPV status and determine to which degree these factors influenced outcome for patients with HPV-positive versus HPV-negative OPSCC.
MATERIALS AND METHODS
Data Source
The National Cancer Database (NCDB), a joint project of the American Cancer Society and the Commission on Cancer (CoC), is a nationwide, facility-based oncology data set that currently captures 70% of all newly diagnosed cancers in the United States. Data reported from these hospital-based cancer registries include patient demographics, American Joint Committee on Cancer (AJCC) 7th Edition staging,15 tumor histopathology, and use of surgical treatment, radiation, and chemotherapy during the initial 12 months after diagnosis. Data on HPV status became available in 2010. Data are coded and reported according to nationally established protocols coordinated by the North American Association of Central Cancer Registries. All data within the NCDB are compliant with the privacy requirements of the Health Insurance Portability and Accountability Act. Institutional review board approval was not required for this study because no patient, provider, or hospital identifiers, or protected health information is included in the data files.
Study Population
Using the NCDB, we selected adult patients with squamous cell carcinoma of the oropharynx who were diagnosed from 2010 to 2014. The following head and neck PUF files were combined: NCDB oropharynx (C051—soft palate, NOS; C052—uvula), NCDB tongue (C019—tongue, base—NOS; C024—lingual tonsil), and NCDB tonsil (C090—tonsillar fossa; C091—tonsillar pillar; C098—overlapping lesion of tonsil; C099—tonsil, NOS [excluding lingual tonsil C024]). A total of 138,988 cases met these criteria.
After inclusion and exclusion criteria were applied, 18,164 cases remained. Inclusion criteria consisted of documented HPV status, vital status, overall clinical stage, and year of diagnosis from 2010 to 2014. We excluded metastatic disease or M stage. Any records missing survival time, overall clinical stage, race, facility type, facility location, insurance, income, education, rural-urban residence, T stage, N stage, or treatment type were excluded.
The primary end point was overall survival, defined as the time from date of diagnosis to the date of death or last follow-up, where those alive were censored at the last follow-up. The primary comparison was HPV status. HPV positivity was defined as being either type 16–positive or type 18–positive, positive for high-risk types, or NOS.
Other covariates included patient demographics and characteristics of race (White, Black, and Others), sex, age, clinical stage, and T- and N-staging (via AJCC 7th edition staging15), treatment type (surgery first and chemotherapy or radiation treatment first), year of diagnosis, primary site (tonsil v nontonsil), and Charlson-Deyo Score (a comorbidity index with values of 0, 1, 2, or greater). Socioeconomic status variables included facility location (Northeast, South, Midwest, and West reported by states of residence), facility type (community cancer program or integrated network cancer program, comprehensive community cancer program, and academic or research program), insurance status (private insurance, government insurance, and uninsured), income (the median value inpatient's census tract of residence on the basis of national quartiles of the 2000 US Census as < $30,000, $30,000-34,999, $35,000-45,999, and ≥ $46,000 in US dollars), education (proportion inpatient's census tract of residence without a high-school degree on the basis of national quartiles of the 2000 US Census as ≥ 29%, 20%-28.9%, 14%-19.9%, and < 14%), rural-urban residence (by the 2003 Rural-Urban Continuum Codes grouping codes 1-3 for metro, 4-6 for urban, and 7-9 for rural), great circle distance (the straight-line distance between two points) to treatment center, and/or distance to treatment facility (0-10, 10-50, and > 50 miles).
Statistical Analysis
Descriptive statistics such as frequencies and percentages were generated for categorical variables and means, and standard deviations were generated for numerical variables. Categorical patient characteristics were compared across HPV status using chi-squared tests, and numerical patient characteristics were compared using analysis of variance.
To determine the relationship between overall survival and any of the aforementioned socioeconomic status variables, univariate and multivariable analyses were performed. Multivariable analyses were stratified by HPV status while controlling for age at diagnosis, race, sex, clinical stage including T- and N-staging, facility type, facility location, insurance status, median-income quartiles, percent of no high-school education quartiles, rural-urban dwelling, Charlson-Deyo score, primary site, and treatment type. Additionally, interaction models using Cox regression were fit to control for age, race, sex, T stage, N stage, and treatment in each interaction model with HPV, SES, and the HPV 3 SES interaction. Model assumptions were checked and verified. The analysis was performed using SAS 9.4 (SAS Institute Inc., Cary, NC). Significance was assessed at the 0.05 level.
The median follow-up from diagnosis was 3.0 years.
RESULTS
Of the 18,164 remaining patients after applying inclusion and exclusion data, the majority or 63.9% were HPV-positive (n = 11,583) compared with those who were HPV-negative (n = 6,581); 90.8% of the entire cohort (n = 16,485) were Caucasian, and 83.1% (n = 15,095) were male. Of the 6,581 HPV-negative patients, 85.31% (n = 5,614) were Caucasian and only 12.48% (n = 821) were African American. The majority of HPV-negative patients had Medicaid, Medicare, or other government insurance (47.67%; n = 3,137). The remaining HPV status comparative data are presented in Table 1.
TABLE 1.
HPV Comparisons
HPV-positive patients (Table 2) living in the south had improved overall survival compared with HPV-positive patients living in the west (hazard ratio [HR]: 0.83, 95% CI, 0.72 to 0.96, P = .013). HPV-positive patients who have private insurance have improved overall survival compared with HPV-positive patients who are uninsured (HR, 0.51, 95% CI, 0.41 to 0.63, P < .001). HPV-positive patients living in census tracts where the median income is less than $30,000 based on the 2000 median-income quartiles have worse overall survival compared with HPV-positive patients living in tracts with a median income of $46,000 or more (HR, 1.42, 95% CI, 1.15 to 1.76, P < .001). HPV-positive patients living in census tracts where percentage without schooling was ≥ 29% have worse overall survival than HPV-positive patients whose tract percentage was < 14% (HR, 1.25, 95% CI, 1.02 to 1.52, P = .028).
TABLE 2.
Multivariate Analysis: Overall survival by HPV-Positive
HPV-negative patients (Table 3) who were African American had a slight risk of worse overall survival compared with HPV-negative patients who were Caucasian (HR, 1.17, 95% CI, 1.03 to 1.32, P = .015). HPV-negative patients who first received chemotherapy or radiation treatment had worse overall survival in contrast to HPV-negative patients who first received surgery (HR, 1.39, 95% CI, 1.24 to 1.55, P < .001). HPV-negative patients who had private insurance had improved overall survival compared with HPV-negative patients who were uninsured (HR, 0.62, 95% CI, 0.53 to 0.73, P < .001). HPV-negative patients living in census tracts where the median income is less than $30,000 had worse overall survival than HPV-negative patients making $46,000 or more (HR, 1.37, 95% CI, 1.16 to 1.63, P < .001).
TABLE 3.
Multivariate Analysis: Overall Survival by HPV-Negative
As assessed through interaction, the relationship between overall survival and facility location significantly differed between those who were HPV-positive and those who were HPV-negative (P = .021). In addition, the relationship between overall survival and insurance status significantly differed between those who were HPV-positive and those who were HPV-negative (P = .004), and the relationship between overall survival and rural-urban status significantly differed between those who were HPV-positive and those who were HPV-negative (P = .009).
DISCUSSION
Although HPV status represents the most significant biologic predictor of outcome in OPSCC, having HPV-related OPSCC confers a specific environmental and socioeconomic risk. The relation between all these factors is complex, and deciphering them is beyond the scope of this report. Our results, however, open the door for an important discussion as to better understanding of factors that affect equities in cancer care. If headway is to be made as far as equity in cancer care, a deeper understanding of the biologic and nonbiologic factors and their potential interactions is an important first step. HPV status alone is known to affect overall survival in OPSCC, yet this influence may rely on nonbiologic factors that distinguish the HPV-related versus HPV-unrelated groups. Our findings contribute to the limited existing literature on HPV status’ influence on overall survival and socioeconomic variables in patients with oropharyngeal head and neck cancer and may have significant implications for domestic policy improvement, expansion of access to healthcare, and preventive and behavioral interventions.
Although an association between lower socioeconomic status and advanced stage at presentation in OPC is documented, the literature is lacking as far as measures of insurance and income in all stages of the disease and their influence on survival by HPV status.5,16,17 Our results show that insurance status and patient income trended towards similar effects regardless of HPV status. Private insurance was largely protective in both HPV-positive and HPV-negative patients compared with uninsured patients of their respective HPV status. Both HPV-positive and HPV-negative patients with lower incomes had worse outcomes than their wealthier counterparts, respectively, in contrast to associations of a previous study.5
In contrast, HPV status does seem to affect overall survival depending on area of residence with facility location and rural-urban dwelling having significant associations with survival based on HPV status.5,17 In line with other studies, living in the south was somewhat protective in HPV-positive patients compared with living in the west.5 Additionally, HPV positivity itself worsened the overall survival of patients whose facilities were located in the west compared with that of patients whose facilities were located in the northeast for unclear reasons. HPV negativity slightly worsened overall survival in patients living in urban-dwelling environments compared with those living in metropolitan environments. However, HPV positivity significantly worsened overall survival in those living in rural environments compared with those living in metropolitan dwelling areas.
The remaining socioeconomic factors of race, education, and treatment conferred some influence on overall survival, although the results differed by HPV status after controlling for all aforementioned covariates. African American patients who were HPV-negative had somewhat worse overall survival compared with Caucasians. This finding did not hold for HPV-positive patients. Similarly, HPV-positive patients with lower amounts of schooling had slightly worse overall survival, which did not hold for HPV-negative patients.
Although interaction models for HPV status and median-income quartiles (P = .969), education (P = .813), and great circle distance (P = .599) were not statically significant, several of the individual hazard ratios were significant and are worth noting. Those living in census tracts where the median income is $46,000 or greater had a protective effect regardless of HPV status (HR, 0.58, 95% CI, 0.51 to 0.65, P < .001 and HR, 0.58, CI, 95% 0.50 to 0.68, P < .001, for HPV-positive and HPV-negative, respectively). Additionally, education measured via census tract percentage without high-school degree reduced the risk of death (< 14% v ≥ 29%: HR 0.64, 96% CI, 0.55 to 0.74, P < .001 and < 14% v ≥ 29%: HR 0.64, 95% CI, 0.56 to 0.72, P < .001, for HPV-positive and HPV-negative, respectively). Finally, the great circle distance of 10-50 miles versus ≤ 10 miles was somewhat protective (HR, 0.88, 95% CI, 0.77 to 1.02, P < .001 and HR, 0.87, 95% CI, 0.80 to 0.96, P < .004, for HPV-positive and HPV-negative, respectively).
Although specific causes for discrepancies by HPV status remain mostly elusive, our results shed some light on the possible contributing factors. Access to medical care, lack of patient mobility, and referral patterns by local physicians may be contributing factors.5,11,16-18 Our study has limitations, since the National Cancer Database does not represent population-based data, and additional studies will be needed to better determine the most influencing factors.
The need to address the social determinants of health inequities in patients with cancer is more apparent than ever. Increasing awareness of existing health disparities may promote further physician-patient discussions to address barriers and may target specific public health interventions.19,20 Not only do these findings call for advocacy for risk factor reduction in our most vulnerable patients, but also they provide a prospect to expand access to care in a multidisciplinary manner,18-21 be it through enhancing preventive efforts in these communities or possibly through other efforts such as telemedicine.
LIMITATIONS
Limitations of this study are similar to those inherent in other retrospective database evaluations. Given that definitive variables of income and education via NCDB are nonexistent and are technically zip code–associated variables, our study uses their data as surrogate, patient-associated variables. Similarly, Rural-Urban Continuum Codes, a classification schema via population density, were used as an estimation or surrogate marker of rural-urban residence. Routine HPV test reporting was not included in the NCDB until 2010, is considered a site-specific factor by the NCDB, and may not be submitted for all cases. This limitation may contribute to under-reporting of HPV for OPC in the NCDB and creating selection bias. History of tobacco use is generally not collected by NCDB, or included consistently in all medical records, although it represents a pertinent factor affecting outcomes of OPCs.5 Additionally, the NCDB does not specify the method by which HPV status was reported. This factor may lead to inherent reporting errors. The variability in HPV detection is limited by resources available at institutions; those institutions with more resources may more routinely perform and report HPV testing, further contributing to selection bias. Although univariate matching aims to minimize confounders, this method, unfortunately, cannot fully eliminate the risk of outstanding discrepancies, potentially allowing unforeseen, unbalanced covariates to influence the resulting outcomes.
Despite these limitations, our findings support the notion that socioeconomic factors including insurance status, rural-urban dwelling, and facility location modified the relation between tumor HPV status and overall survival.
In conclusion, although HPV association remains one of the most important biologic predictors of clinical outcome in OPSCC, our findings support the notion that the impact of HPV on outcome is influenced by nonbiologic factors such as insurance status, rural-urban dwelling, and facility location. Race, education, and treatment type exerted some influence on overall survival.
Although not directly assessed in this analysis, consideration for the impact these factors had on waiting time to treatment may serve as a vehicle for future exploration in OPSCC. Demographic and geographic discrepancies may serve as proxies for systemic barriers to care and can be addressed via quality or accountability metrics as advocated for by the American College of Surgeons Commission on Cancer accreditation program for Comprehensive Cancer Centers. Although the analysis of these variants is outside the scope of our investigation, these divergences remain areas of great interest. A deeper understanding of these interactions is needed to improve equity of care for patients with OPSCC.
SUPPORT
Research reported in this publication was supported in part by the Biostatistics and Bioinformatics Shared Resource of Winship Cancer Institute of Emory University and NIH/NCI under award number P30CA138292. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
AUTHOR CONTRIBUTIONS
Conception and design: Jennifer A. Marks, Conor E. Steuer, Nabil F. Saba
Administrative support: Dong M. Shin
Provision of study materials or patients: Mihir R. Patel, Dong M. Shin
Collection and assembly of data: Kristin Higgins
Data analysis and interpretation: Jennifer A. Marks, Jeffrey M. Switchenko, Conor E. Steuer, Mihir R. Patel, Johnathan J. Beitler, Dong M. Shin, Nabil F. Saba, Martha Ryan, Mark W. McDonald, Theresa W. Gillespie
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Socioeconomic Factors Influence the Impact of Tumor HPV Status on Outcome of Patients With OPSCC
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.
Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).
Conor E. Steuer
Honoraria: Merck
Consulting or Advisory Role: Abbvie, BerGenBio, ARMO BioSciences, Lilly
Research Funding: Vaccinex, Seattle Genetics, Daiichi Sankyo, Infinity Pharmaceuticals</string>
Kristin Higgins
Consulting or Advisory Role: AstraZeneca, Genentech, Varian Medical Systems, Presica
Speakers' Bureau: AstraZeneca
Research Funding: Reflexion Medical
Travel, Accommodations, Expenses: AstraZeneca
Nabil F. Saba
Honoraria: Merck, Lilly, Pfizer, Bristol-Myers Squibb, cue, GlaxoSmithKline, Aduro Biotech, Kura, Genentech/Roche
Consulting or Advisory Role: Pfizer, Bristol-Myers Squibb, Merck, Lilly, Bluprint, BioNTech AG
Research Funding: Bristol-Myers Squibb, Exelixis
Travel, Accommodations, Expenses: Bristol-Myers Squibb, Merck, Pfizer, Lilly, GlaxoSmithKline, Genentech/Roche, Bluprint
No other potential conflicts of interest were reported.
REFERENCES
- 1.Ferlay J, Colombet M, Soerjomataram I, et al. : Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods. Int J Cancer 144:1941-1953, 2019 [DOI] [PubMed] [Google Scholar]
- 2.Saba NF, Goodman M, Ward K, et al. : Gender and ethnic disparities in incidence and survival of squamous cell carcinoma of the oral tongue, base of tongue, and tonsils: A surveillance, epidemiology and end results program-based analysis. Oncology 81:12-20, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chaturvedi AK, Engels EA, Pfeiffer RM, et al. : Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol 29:4294, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cannon RB, Shepherd HM, McCrary H, et al. : Association of the Patient Protection and Affordable Care Act with insurance coverage for head and neck cancer in the SEER database. JAMA Otolaryngol Head Neck Surg 144:1052-1057, 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Liederbach E, Kyrillos A, Wang CH, et al. : The national landscape of human papillomavirus-associated oropharynx squamous cell carcinoma. Int J Cancer 140:504-512, 2017 [DOI] [PubMed] [Google Scholar]
- 6.Pytynia KB, Dahlstrom KR, Sturgis EM: Epidemiology of HPV-associated oropharyngeal cancer. Oral Oncol 50:380-386, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Ang KK, Harris J, Wheeler R, et al. : Human papillomavirus and survival of patients with oropharyngeal cancer. N Engl J Med 363:24-35, 2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bergmark RW, Sedaghat AR: Disparities in health in the United States: An overview of the social determinants of health for otolaryngologists. Laryngoscope Investig Otolaryngol 2:187-193, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ragin CC, Langevin SM, Marzouk M, et al. : Determinants of head and neck cancer survival by race. Head Neck 33:1092-1098, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Molina MA, Cheung MC, Perez EA, et al. : African American and poor patients have a dramatically worse prognosis for head and neck cancer: An examination of 20,915 patients. Cancer 113:2797-2806, 2008 [DOI] [PubMed] [Google Scholar]
- 11.Auluck A, Walker BB, Hislop G, et al. : Socio-economic deprivation: A significant determinant affecting stage of oral cancer diagnosis and survival. BMC Cancer 16:569, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Benard VB, Johnson CJ, Thompson TD, et al. : Examining the association between socioeconomic status and potential human papillomavirus-associated cancers. Cancer 113:2910-2918, 2008 [DOI] [PubMed] [Google Scholar]
- 13.Singh GK, Jemal A: Socioeconomic and racial/ethnic disparities in cancer mortality, incidence, survival in the United States, 1950-2014: Over six decades of changing patterns and widening inequalities. J Environ Public Health 2017:2819372, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Christophe V, Leroy T, Seillier M, et al. : Determinants of patient delay in doctor consultation in head and neck cancers (Protocol DEREDIA). BMJ Open 4:e005286, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Edge SB, Compton CC: The American Joint Committee on Cancer: The 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 17:1471-1474, 2010 [DOI] [PubMed] [Google Scholar]
- 16.Chen AY, Schrag NM, Halpern MT, et al. : The impact of health insurance status on stage at diagnosis of oropharyngeal cancer. Cancer 110:395-402, 2007 [DOI] [PubMed] [Google Scholar]
- 17.Amini A, Jasem J, Jones BL, et al. : Predictors of overall survival in human papillomavirus-associated oropharyngeal cancer using the National Cancer Data Base. Oral Oncol 56:1-7, 2016 [DOI] [PubMed] [Google Scholar]
- 18.Faraji F, Rettig EM, Tsai HL, et al. : The prevalence of human papillomavirus in oropharyngeal cancer is increasing regardless of sex or race, and the influence of sex and race on survival is modified by human papillomavirus tumor status. Cancer 125:761-769, 2019 [DOI] [PubMed] [Google Scholar]
- 19.Siegel RL, Miller KD, Jemal A: Cancer statistics, 2019. CA Cancer J Clin 69:7-34, 2019 [DOI] [PubMed] [Google Scholar]
- 20.Zandberg DP, Liu S, Goloubeva O, et al. : Oropharyngeal cancer as a driver of racial outcome disparities in squamous cell carcinoma of the head and neck: 10-year experience at the University of Maryland Greenebaum Cancer Center. Head Neck 38:564-572, 2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Conway DI, McMahon AD, Smith K, et al. : Components of socioeconomic risk associated with head and neck cancer: A population-based case-control study in Scotland. Br J Oral Maxillofac Surg 48:11-17, 2010 [DOI] [PubMed] [Google Scholar]