Neighborhood Socioeconomic Status and Racial and Ethnic Survival Disparities in Oral Cavity and Laryngeal Cancer

◥ Background: Oral cavity cancer (OCC) and laryngeal cancer are among the most common cancers worldwide. This study investigated survival in non-Hispanic (NH) Black, NH White, Asian, and Hispanic patients with OCC and laryngeal cancer of low, intermediate, and high neighborhood socioeconomic status (nSES). Methods: WeuseddatafromtheSEER18CensusTract-levelSES and Rurality Database of the National Cancer Institute to create cohorts of OCC and laryngeal cancer patients from 2013 to 2018. Univariate survival analysis was performed with Kaplan – Meier curves and log-rank P values by nSES and then the cross-classi ﬁ cation of race, ethnicity, and nSES. We used Cox proportional hazards regression model for multivariable analysis. Results: HighernSESwasassociatedwithbetterOCCsurvivalfor NH White, NH Black, and Asian patients, and better laryngeal cancer survival for NH White, NH Black, Hispanic, and Asian patients. In the multivariable analyses of both OCC and laryngeal cancer survival, NH Black patients had worse survival than NH WhitepatientsinthehighnSEStertile.NHBlackpatientswithOCC were at higher risk of death than NH White patients at all nSES levels. Conversely, Asian patients with laryngeal cancer demonstrated better survival than other races within the high nSES. Conclusions: Overall survival differs between racial and ethnic groups of similar nSESs. These health disparities in patients with OCC and laryngeal cancer re ﬂ ect broader inequities in the cancer control continuum. Impact: The cross-classi ﬁ cation of race, ethnicity, and nSES revealed disparities in the 5-year overall survival of patients with OCC and laryngeal cancer and highlights the importance of inter-sectionality in the


Introduction
Although head and neck cancers are considered together due to proximal anatomy, risk factors differ dramatically by tumor site.Oropharyngeal cancer is primarily driven by human papillomavirus (HPV), while oral cavity cancer (OCC) and laryngeal cancer are associated with smoking and drinking.In the United States alone, there will be an estimated 12,620 new cases and 3,770 deaths from laryngeal cancer in 2021 (1).OCC affects even more people, with estimates of about 54,010 new cases and 10,850 deaths from oral cavity or oropharyngeal cancer in 2021 (1).
Black Americans have particularly poor survival in both OCC and laryngeal cancer, with 5-year survival of 50% and 52.7%, respectively (2).Previous research exploring the contributors to this survival disparity highlighted the stage at diagnosis and history of tobacco and alcohol use (3,4).Black patients with OCC were almost four times as likely as White patients to be diagnosed with advanced stage disease (4).While tobacco and alcohol use are known risk factors for worse OCC and laryngeal cancer survival, their relationship and contribution to racial disparities in survival are unclear (5,6).White and Black patients with head and neck cancer have been found to have similar tobacco usage rates, yet among smokers and nonsmokers, White patients have better survival than Black patients (7).
Socioeconomic status (SES), race, and ethnicity are also inextricably linked in the United States with effects on perceived discrimination in health, which translates to measured outcomes that differ by race and ethnicity and are further stratified by SES.While higher SES in White patients is associated with less perceived discrimination, the opposite was true among Black patients, highlighting the complex interplay between SES and race (8,9).There is a paucity of literature analyzing the impact of SES and race on cancer sites not associated with HPV.Thus, the objective of this study is to estimate the effect of race, ethnicity, and neighborhood SES (nSES) in head and neck cancers not associated with HPV.We limited our study to investigate these cancers because of differences in pathophysiology and outcomes with HPVassociated head and neck cancers.

Exposure
Our primary exposures of interest were nSES and race and ethnicity.The nSES index was provided by SEER.It was constructed from a factor analysis of seven census-tract variables: median household income, median house value, median rent, percent below 150% of the poverty line, education index, percent working class, and percent unemployed (11).Patients with cancer were linked with the Decennial Census 2010 census tract boundaries.The nSES indices were estimated from the American Community Survey (ACS) 5-year estimates.The indices were linked to cancer cases at the census tract level by matching the survey year with the cancer diagnosis year.Tumors diagnosed in 2006 to 2007 are linked with SES quintiles calculated using ACS 2006-2010 data.Tumors diagnosed in 2008 to 2017 are linked to SES quintiles based on ACS 2006-2010, 2007-2011, 2008-2012, 2009-2013, 2010-2014, 2011-2015, 2012-2016, 2013-2017, 2014-2018, and 2015-2019 data, respectively.Finally, tumors diagnosed in 2018 are linked to the index estimated from 2015 to 2019 ACS data.The census tract-level SES index was used to stratify subjects into tertiles of low, intermediate, and high socioeconomic levels.Tertiles were provided by SEER, which defined three groups of equal number within a state.For race, we grouped race and ethnicity into non-Hispanic (NH) Black, NH White, NH Asian, NH Pacific Islander, NH American Indian/Alaska Native, and Hispanic (all races).Throughout the remainder of the article, references to NH Asian American, NH Pacific Islander, and NH American Indian/Alaska Native race omit the "NH" prefix but are understood to include only non-Hispanic cases.

Covariates
All covariate variables were available through SEER.We selected age, sex, primary subsite, stage, and rural-urban context a priori as variables of interest.The census tract-level rurality variables were the US Department of Agriculture (USDA)'s Rural Urban Commuting Area (RUCA) codes.We categorized RUCA into the Census Bureau's four-category classification as All Rural, All Urban, Mostly Rural, and Mostly Urban.Rural urban context was updated every 10 years.

Outcomes
The primary outcome was death from any cause provided by SEER, which receives mortality data from CDC's National Center for Health Statistics.We censored patients after five years from the date of diagnosis.

Statistical analyses
Because OCC and laryngeal cancer have different survival outcomes, we conducted survival analyses separately for patients with OCC and laryngeal cancer.Univariate survival analysis was performed with Kaplan-Meier analyses and log-rank P values by nSES and then the cross-classification of race and nSES.We used the Cox proportional hazards regression model for multivariable analysis, adjusting for age in 10-year groups, sex, stage, and subsite.We tested for Cox proportional hazards.We tested the proportion hazards assumption using the cox.zph function in the survival package to test for independence between Schoenfeld residuals and time.In addition, given the large sample size, we also graphed scaled Schoenfeld residuals against the transformed time.For variables that violated the proportional hazards assumption, we used underlying stratification, which allows for different hazard functions across the strata of the variable.In the oral cavity models, site and stage violated the hazards assumption.In the larynx models, site, age, and stage violated the hazards assumption.Race and ethnicity, nSES and the cross-classification of race and ethnicity and nSES met the proportional hazard assumption, and hazard ratios (HR) were presented.We assessed interaction multiplicatively with an interaction term in the Cox proportional hazards regression.All analyses were performed with R (version 4.0.3).We used an alpha of 0.05.

Data availability
The data analyzed in this study were retrieved from the SEER (Surveillance, Epidemiology, and End Results) 18 Census Tract-level SES and Rurality Database (2000-2016) database of the National Cancer Institute at https://seer.cancer.gov/data/.

Patient characteristics-opposing trends in nSES distribution between races and ethnicities in patients with OCC and laryngeal cancer
Patient characteristics are displayed in Table 1 for patients with OCC and Table 2 for patients with laryngeal cancer.Male patients comprised a larger proportion of the laryngeal cancer patients than the OCC patients.In both laryngeal cancer and OCC patients, more NH Black and Hispanic patients belonged to the lowest nSES tertile than the middle and highest tertiles.In contrast, Asian patients with OCC and laryngeal cancer were more likely to be in the highest nSES tertile, followed by middle and low tertiles.Similarly, most NH White patients with OCC were in the high nSES tertile.Pacific Islander and American Indian/Alaska Native patients with OCC were most likely to be in the middle nSES tertile while American Indian/Alaska Native patients with laryngeal cancer were most likely to be in the lowest nSES tertile.
Univariable analyses-nSES was associated with OCC and laryngeal cancer five-year survival Among patients with OCC (Supplementary Fig. S1; Supplementary Table S1) and laryngeal cancer (Supplementary Fig. S2; Supplementary Table S2), living in low nSES census tracts was associated with poor five-year overall (OS) survival.Notably, this trend persisted when nSES was stratified by race and ethnicity (Fig. 1) in the OCC cohort for NH White (54.8% vs. 45.4% for high vs. low nSES), NH Black (36.1% vs. 28.3%),and Pacific Islander patients (56.5% vs. 20.0%).In Hispanic patients with OCC, there was not a statistically significant difference in Sex Female Sex Female five-year OS (47.1% vs. 41.6%,P ¼ 0.17).There was a statistically significant difference in five-year OS among Asian patients with OCC, but belonging to the middle nSES census tract was associated with poor five-year OS.The association between low nSES and poor five-year OS is also not observed in American Indian/Alaska Native patients with OCC.In the laryngeal cancer cohort stratified by race and ethnicity (Fig. 2), living in low nSES census tracts was associated with poor fiveyear OS for NH White (55.4% vs. 44.7%),NH Black (47.0% vs. 39.0%),Hispanic (52.9% vs. 46.1%),and Asian patients (59.8% vs. 46.8%).Pacific Islander and American Indian/Alaska Native patients with laryngeal cancer living in middle nSES census tracts was associated with poor five-year OS, although this did not reach statistical significance in either group (P ¼ 0.34, P ¼ 0.13).

OCC multivariable analyses-NH Black patients have worse fiveyear OS despite high nSES
For the multivariable analyses of both cohorts, we used the Cox proportional hazards regression model adjusted for age in 10-year groups, sex, stage, and subsite.Race, ethnicity, and nSES were significant determinants of 5-year OS. Figure 3 shows that NH Black patients with OCC had worse five-year OS than NH White patients [HR, 1.31; 95% confidence interval (CI), 1.20-1.43].Higher nSES was associated with better five-year OS overall in patients with OCC.Upon cross-classifying by race (Fig. 4), NH Black patients fared worse than NH White patients within every level of nSES.The association of NH Black race and ethnicity compared with NH White is also greater in the low nSES group than the high nSES group, although this was not statistically significant (HR 1.69/HR 1.35 ¼ 1.25 in the low nSES group and HR 1.14/HR 1 ¼ 1.14 in the high nSES group; Supplementary Table S3).Asian patients of high nSES had the best five-year OS (HR, 0.86; 95% CI, 0.69-1.06).The association of Asian race and ethnicity compared with NH White in the low nSES group is both significantly higher and in the opposite direction than in the high nSES group (HR 1.62/HR 1.35 ¼ 1.20 in the low nSES group and HR 0.86/HR 1 ¼ 0.86 in the high nSES group).

Laryngeal cancer multivariable analyses-Asian patients have better five-year OS despite low nSES
Asian race was associated with better five-year OS (HR, 0.82; 95% CI, 0.74-0.92)while NH Black race was associated with poorer fiveyear OS (HR, 1.12; 95% CI, 1.07-1.17)when each is compared with NH White patients with laryngeal cancer (Fig. 3).Although higher nSES was associated with better five-year OS, NH Black patients had significantly higher risk of death compared with NH White patients within the high nSES tertile (HR, 1.26; 95% CI, 1.07-1.49;Fig. 4).In contrast, Asian patients with laryngeal cancer demonstrated significantly better outcomes than NH White, NH Black, and Hispanic patients within the high nSES group (HR, 0.81; 95% CI, 0.68-0.97).The association of Asian race and ethnicity compared with NH White is similar in both the low and high nSES groups (HR 1.16/HR 1.33 ¼ 0.87 in the low nSES group and HR 0.81/HR 1 ¼ 0.81 in the high nSES group).In both the intermediate and low nSES groups, Asian patients with laryngeal cancer showed no significant difference in five-year OS compared with NH White patients of high nSES (intermediate nSES HR, 0.91; 95% CI, 0.76-1.10;low nSES HR, 1.16; 95% CI, 0.93-1.45;Supplementary Table S3).This is in contrast to NH Black patients, who had poorer five-year OS compared with NH White patients in every nSES tertile.The association of NH Black race and ethnicity compared with NH White is also greater in the high nSES group than the low nSES group, although this was not statistically significant (HR 1.26/HR 1 ¼ 1.26 in the high nSES group and HR 1.45/HR 1.33 ¼ 1.09 in the low nSES group).Overall, our analyses revealed that race and SES are independent on the multiplicative scale, as the multiplicative interactions were not significant (Supplementary Table S3).

Disaggregation of Asian and Pacific Islander patients with OCC and laryngeal cancer-variable association between nSES and five-year OS
In the univariable analysis of nSES and race and ethnicity, there were unexpected patterns of survival observed in Asian and Pacific Islander patients with OCC and laryngeal cancer that contrasted with patterns of survival in other groups.When Asian and Pacific Islander patients are disaggregated into East Asian, Southeast Asian, Indian/Pakistani Asian, and Pacific Islander groups, there is significant variability in survival by nSES.In OCC (Supplementary Fig. S3), Pacific Islander patients in low nSES census tracts have significantly poorer five-year OS (P ¼ 0.0043), an association that is not seen in other subgroups.Of note, Indian/Pakistani Asian patients from intermediate nSES demonstrated poorer five-year OS survival than low and high nSES patients (P ¼ 0.015).There was no significant difference in survival among nSES for Southeast Asians or East Asians (P ¼ 0.76 and P ¼ 0.078, respectively).In laryngeal cancer (Supplementary Fig. S4), living in a low nSES census tract is only associated with poorer five-year OS among Indian/Pakistani Asian patients (P ¼ 0.0068).

Discussion
Our study investigates intersectional health disparities of nSES and race and ethnicity for OCC and laryngeal cancer patients of NH Black, NH White, Hispanic, and Asian racial identities with a disaggregated subset analysis of East Asian, Southeast Asian, Pacific Islander, and Indian/Pakistani patients within the Asian group.Although prior studies on head and neck cancer (HNC) survival among racial and ethnic minority groups have found relatively poor survival among NH Black patients, our study expands the scope of comparison by including NH White patients and disaggregating Asian and Pacific Islander survival data (12).Our exploration of the joint effects of race, ethnicity, and nSES unmasked survival patterns that would not have been apparent otherwise.These patterns highlight the inextricability of race, ethnicity, and nSES in analyses of patient outcomes.The discovery that the impact of nSES on survival varies significantly between racial and ethnic groups contributes substantively to demonstrating the importance of recognizing intersectionality when studying survival.
The overall survival differed dramatically between different racial and ethnic groups of similar nSESs, indicating that patients face intersectional barriers that extend beyond race and ethnicity or nSES alone.This is consistent with a study on HNC survival using data from 1998 to 2002 revealing that among NH Black, NH White, and Hispanic oral cavity and larynx cancer patients, NH Black patients had worse survival that was not solely explained by demographics (7).Furthermore, although lower nSES was generally associated with poorer survival for patients with laryngeal cancer compared with the reference group of high nSES NH White patients, Asian patients deviate from this trend.These observations suggest that the impact of nSES on survival is not uniform between populations.A recent analysis of the interaction between gender and race in HNC survival similarly demonstrated interaction between multiple factors (13).Therefore, health disparities in patients with OCC and laryngeal cancer reflect broader intersectional inequities in the cancer control continuum from detection and diagnosis to medical treatment and survivorship.

Detection and diagnosis: Socioeconomic and racial and ethnic factors contribute to risk and delay diagnosis
Patients from low nSES face social and structural barriers to affording and accessing health care.The lack of health and dental insurance or difficulty meeting copays and deductibles often discourages patients from seeking medical attention until their conditions have progressed severely (14).The lack of access to transportation may also deter patients from routine health checkups, preventative dental care, and early screenings.Consequently, diseases like OCC and laryngeal cancer may not be diagnosed until tumors lead to complications with speech and swallowing, which increase the likelihood of diagnosis at a later stage (15).
Racial and ethnic identity adds another layer of complexity to nSES in detection and diagnosis.We found that despite the association between higher nSES and better survival, NH Black patients experience worse survival than other racial groups within the high nSES tertile.While the mechanisms for this disparity are likely multifactorial, one may be delayed diagnosis among NH Black patients, who may experience greater medical mistrust and perceived bias in medical settings (15)(16)(17).Historical traumas inflicted upon Black communities and lived experiences of discrimination may also contribute to mistrust of medical institutions (18,19).Unlike White patients for whom higher nSES leads to protective privilege and lower discrimination, Black patients of high nSES continue to report more discrimination than their peers (8).These factors likely shape Black patients' decisions to access medical care before their cancer diagnosis, as demonstrated by Black patients presenting with OCC at later stages than NH White patients (20).These patterns of NH Black survival support findings from a prior study by Taylor and colleagues on HNC survival among racial and ethnic minority groups, which showed NH Black patients experienced worse survival than other minority groups when SES was accounted for (12).Although Taylor and colleagues also examined HNC survival among Asian and Pacific Islander patients, a growing body of literature has demonstrated differences in patient factors and survival outcomes between Asian and Pacific Islander patients.The aggregation of Asian and Hawaiian or Pacific Islander patients masks differences in SES, health care access, comorbidities, and survival outcomes between the populations (21).A publication by Taparra and colleagues delineates a historically rooted argument for the disaggregation of this data, which includes cultural and socioeconomic differences between immigrants of those identities (22).
By aggregating data from these very distinct populations, researchers mask remarkable differences in prevalence, prognosis, and survival.Therefore, we separated Asian from Pacific Islander survival data and discovered that Asian patients in the low nSES tertile of the laryngeal cancer cohort displayed survival comparable with NH White patients in the high nSES tertile.This finding is consistent with results from a prior study by Moon and colleagues, demonstrating that Asian patients had better survival than NH White and Pacific Islander patients for head and neck cancer (23).They posited that the difference in survival may result from a higher degree of comorbidity burden in the Pacific Islander population.In our case, it is likely that a reduced comorbidity burden among the disaggregated Asian patients in our laryngeal cancer and OCC cohorts may be a contributing factor to their better five-year OS.
Treatment: differences in medical management and access drive outcome disparities OCC and laryngeal cancer require longitudinal treatment and follow-up regardless of treatment modality.Treatment adherence is essential for long-term survival.Patients from lower nSES face many obstacles related to treatment access and adherence.Patients from economically disadvantaged backgrounds can take less time off work or child care for medical visits (24).These patients may also lack transportation to reach the clinic or hospital, challenging treatment adherence.They are also less likely to obtain mental health services and counseling throughout treatment, associated with lower treatment adherence rates and poorer psychological health (25).Patients with cancer who belong to low nSES have higher suicide risks than other patients with cancer (26).In addition, patients with head and neck cancer experience unique psychosocial challenges that contribute toward a suicide rate double those of other cancers (27).Even among patients with cancer, the impact of low nSES on OCC and laryngeal cancer patient outcomes is particularly devastating.
Management for OCC and laryngeal cancer can be surgical or nonsurgical, either with or without radiotherapy (RT) and chemotherapy.Surgical therapy is associated with better outcomes than RT for OCC.However, NH Black patients receive surgery less often and undergo RT at higher rates than other groups, despite its association with worse survival outcomes (28).While this pattern may not be definitively attributable to patient preferences impacted by social barriers or medical mistrust in the NH Black community, literature on physician-patient communication suggests that NH Black patients may receive inadequate patient education and shared decision-making when discussing surgical management.Patient-reported shared decision-making decreases with their nSES (29).In addition, shared decision making may be further complicated for Black patients who experience physician mistrust (30).These factors likely contribute toward a preference for nonsurgical management of OCC in Black patients, which is associated with worse overall survival.

Survivorship: cultural and community factors may affect longitudinal care and survival
Even after treatment, patients with cancer from low nSES are less likely to participate in follow-up care discussions with physicians on topics including long-term treatment effects, risk of recurrence, and surveillance for recurrence (31).Decreased adherence to functional therapies during and after treatment, such as speech and swallow therapy, leads to poorer rehabilitation and quality of life posttreatment.Lower education levels, conflicting work and family responsibilities, lack of time, and lack of support have been associated with poorer adherence to dysphagia therapy in patients with head and neck cancer (31).These barriers to health equity highlight the unfortunate reality that socioeconomic disparity permeates through every phase of the continuum of care.
Hispanic patients fare better than NH Black patients despite census data showing that they are similar in socioeconomic status on a population level.Sociologic research suggests that the difference in outcomes of Hispanic and NH Black patients of similar nSES, or the "Hispanic Paradox," may result from increased resiliency in Hispanic patients who are immigrants (32).However, the precise mechanisms for this phenomenon are currently unknown and warrant investigation.In addition, further research is needed to analyze the role of factors such as comorbidity and smoking in survivorship differences between racial and ethnic and nSES groups.Higher rates of comorbidity and smoking are independently associated with poorer cancer outcomes, but their impacts on survivorship are likely compounded for patients facing intersectional disadvantages in the health care system (3,33).
The multifaceted influence of culture and community on survival is also evident in the trends observed in the subgroup analysis of Asian patients.Although higher nSES is broadly associated with better survival, KM analysis of Indian/Pakistani OCC patients revealed that the intermediate nSES group demonstrated the poorest survival.Interestingly, this pattern is not reflected for Indian/Pakistani patients with laryngeal cancer.One potential explanation for this pattern involves the regional practice of smoking and chewing tobacco preparations made of crushed betel nut, such as gutka or paan in India and Pakistan (34).Consumption of crushed betel nut as a chewing preparation is linked to increased rates of oral cancers (35,36).It is possible that the nSES tertiles may represent different waves of South Asian diaspora migration to the United States.The low nSES tertile may represent Indian/Pakistani immigrants who arrived decades ago and have undergone greater degrees of assimilation and are less adherent to traditional practices and have lower rates of betel chewing.Intermediate nSES patients may represent more recent waves of highly educated immigrants who arrived later (37).The intermediate nSES group that immigrated more recently may have the highest rates of smoking and betel use and therefore the poorest OCC survival.However, our understanding of the reasons for these survival patterns is limited by the lack of information on nativity, degree of acculturation, and length of habitation in the United States, among other factors.

Limitations and future directions
Our findings demonstrated that the impact of nSES on survival is variable between different racial and ethnic groups in the United States.However, more research is required to identify the mechanisms contributing to these differences on a population level.Asian patients within the laryngeal cancer cohort demonstrate better survival outcomes than other racial groups in the high nSES tertile and exhibit survival outcomes in the intermediate and low nSES tertiles comparable with NH White patients of high nSES.Extant literature suggests that factors in the social environment, such as ethnic density, may influence cancer survival outcomes.Living in ethnic enclaves has been associated with lower cancer mortality in Latinx and Asian communities (38,39).Conversely, living in ethnically dense Black communities is associated with increased cancer mortality (40).These contrasting narratives of ethnic density are consistent with our study's findings that protective factors for survival in one identity group may play a detrimental role in outcomes for another.
One limitation of this study is the reliance on race to group patients of various ancestries and ethnic backgrounds on a population level.The use of race in biomedical research is constrained because race is a social construct that does not reflect a biological distinction between humans (6,41).Although there is increasing recognition of the need to disaggregate Asian survival data in epidemiologic research, there is progress to be made for Hispanic patients.The Hispanic ethnicity is inclusive of over twenty different countries across North America, Central America, and South America.We are also limited in our understanding of nativity, length of habitation, and degree of acculturation among Hispanic communities.We recommend that efforts be made to promote disaggregation of the Hispanic ethnicity in future epidemiologic research.
This study is also constrained by its retrospective nature.Data collection and analysis are limited to the categorical variables provided by the SEER database.Furthermore, the SEER database does not provide individual subject-level socioeconomic status data.A limitation of census tract analysis is the simplification of nSES heterogeneity within census tracts.Finally, there are no geographic identifiers included in this database due to confidentiality concerns.Thus, we are unable to account for geographic clustering of this study.
In reality, the driver of disparate health outcomes is not race, but racism (43,44).The impact of systemic racism on health cannot be ignored, and it is necessary to recognize its role in outcome disparities instead of attributing differences to race and ethnicity alone.The geography of residence is complicated by the long history of segregation, redlining, and structural racism in the United States (35).This history continues to shape how the marginalization of Black communities shapes their access to medical care, healthy food options, and social services even to the present day (45,46).Recognizing the role of systemic racism as a mediator of health outcomes is prerequisite to understanding that the driver of worse survival for NH Black patients is not their race-it is the culmination of centuries of oppression NH Black communities have been subjected to throughout American history.
The current study shows that race, ethnicity, and nSES are significant independent predictors of survival.Further research is required to identify the causes of worse survival for NH Black patients from high nSES and to better characterize risk and survival for Hispanic patients of different cultures and nationalities.

Figure 1 .
Figure 1.Five-year overall survival of patients with oral cavity cancer in racial and ethnic groups stratified by nSES.Six Kaplan-Meier curves depicting the 5-year overall survival of oral cavity cancer patients in different racial and ethnic groups by nSES group.Top right, Pacific Islander; top middle, Hispanic; top left, NH White; bottom right, American Indian/Alaska Native; bottom middle, Asian; bottom left, NH Black.

Figure 2 .
Figure 2. Five-year overall survival of patients with larynx cancer in racial and ethnic groups stratified by nSES.Six Kaplan-Meier curves depicting the 5-year overall survival of patients with larynx cancer in different racial and ethnic groups by nSES group.Top right, Pacific Islander; top middle, Hispanic; top left, NH White; bottom right, American Indian/Alaska Native; bottom middle, Asian; bottom left, NH Black.

Figure 3 .
Figure 3. Multivariable analyses of survival in oral cavity and larynx cancer.Two forest plots depicting the hazard ratio and 95% CIs of patients with oral cavity (top) and larynx cancer (bottom) in a multivariable analysis by race and ethnicity and nSES.Group 1 indicates low nSES, Group 2 indicates intermediate nSES, and Group 3 indicates high nSES.

Figure 4 .
Figure 4. Survival in patients with oral cavity and larynx cancer by race and ethnicity and nSES.Figure 4 is two forest plots depicting the hazard ratio and 95% CIs of patients with oral cavity and larynx cancer in a multivariable analysis by nSES and cross-classified by race and ethnicity.Group 1 indicates low nSES, Group 2 indicates intermediate nSES, and Group 3 indicates high nSES.

Table 1 .
Oral cavity cancer patient descriptive statistics and characteristics by race and ethnicity and nSES in the United States, 2013-2018.

Table 2 .
Laryngeal cancer patient descriptive statistics and characteristics by race and ethnicity and nSES in the United States, 2013-2018.