Head and Neck Cutaneous Soft-Tissue Sarcoma Demonstrate Sex and Racial/Ethnic Disparities in Incidence and Socioeconomic Disparities in Survival

Background: Cutaneous soft-tissue sarcoma (CSTS) of the head and neck are rare and are known to have aggressive clinical course. The current study utilizes a population-based registry in the U.S. to characterize these malignancies and explore disparities. Methods: National Cancer Institute’s (NCI) Surveillance, Epidemiology and End Result (SEER) database from 2000 to 2018 was queried to report incidence and survival data in 4253 cases in the U.S. Results: Males were 5.37 times more likely and Non-Hispanic-White people (NHW) were 4.62 times more likely than females and Non-Hispanic-Black people (NHB) to develop CSTS of the head and neck. The overall incidence was 0.27 per 100,000 persons in 2018, with a significant increase since 2000. Advanced age and stage, histologic group other than ‘fibromatous sarcoma’ and lower SES groups were independent factors for worse overall survival. Conclusions: CSTS of the head and neck demonstrate sex and racial/ethnic disparities in incidence and socioeconomic disparities in overall survival. Level of evidence: II.


Introduction
Cutaneous soft tissue sarcomas (CSTS) of the head and neck are rare [1] and represent a wide array of histopathological diagnoses [2]. CSTS of the head and neck often have an aggressive clinical course relative to other cutaneous malignancies [2]. They may be associated with significant cutaneous and systemic manifestations, and may have dramatic impacts on quality of life (QoL) [2]. Previous studies in the literature have reported the experience of single centers with resultant limited sample sizes [3][4][5] and, given the rarity of the disease, have been limited in their statistical power to analyze prognostic factors. In contrast, existing population-based studies have focused on cutaneous sarcoma located anywhere in the body [6] or aggregated soft tissue and bone sarcomas of the head and neck for combined analysis [7,8]. An Australian study addressing head and neck sarcoma has provided a sub-analysis of CSTS [9]. Thus, a dedicated population-based analysis of cutaneous soft tissue sarcoma (CSTS) of the head and neck in the US is lacking in the literature.
The National Institute of Minority Health and Health Disparities defines health disparities research as that which addresses health differences in socially disadvantaged populations related to specific outcomes [10]. In the context of cancer, disparities in the incidence, prevalence, rate of screening, stage at initial presentation, morbidity, survival, and financial burden of disease have been reported for multiple primary malignancies [11]. Disparity research exploring CSTS of the head and neck is lacking in the literature. We aimed to explore and report the disparities in incidence for CSTS of the head and neck.
National Comprehensive Care Network (NCCN) guidelines recommend proper biopsy technique, followed by surgical resection with adequate margins and complete histological analysis for the treatment of CSTS of the head and neck [1]. Disparate access to cancer treatment in general [12,13] and surgical oncology in particular have been previously reported in the literature [14,15]. Patients with CSTS of the head and neck typically present with a skin mass and may present to a physician without an oncologic training. As a consequence, these patients may undergo an unplanned biopsy/surgical excision without attention to oncologic principles [1,2]. Considering surgical resection is the cornerstone for cancer therapy for CSTS of the head and neck, our goal was to explore the disparities in overall survival. We have utilized the National Cancer Institute's (NCI), population-based Surveillance, Epidemiology and End Result (SEER) database for our analysis.

Methods
The cohort of cases for the current study was isolated using the NCI's SEER program [11]. Presently, SEER collects the data from 22 registries covering approximately 48% of the US population [11]. We utilized the ICD-O-3 codes for malignant histologic behavior and primary location to isolate a total of 4253 cases. Histologic recode broad groups included 'soft tissue sarcoma 8800-8809, fibromatous sarcoma 8810-8839, myxomatous sarcoma 8840-8849, lipomatous sarcoma 8850-8889, myomatous neoplasms 8890-8929, complex mixed and stromal sarcoma 8930-8999 and blood vessel sarcoma 9120-9169'. The number of patients belonging to each broad group are summarized in the Supplemental Table. In addition, frequency table detailing the number of patients corresponding to individual ICD-O-3 code are also presented in the Supplementary Material. Given the frequency distribution and for the statistical analysis, we combined the patients with myxomatous, lipomatous and complex mixed and stromal sarcoma categories in a single category. Patients with diagnosis of Kaposi sarcoma 9140.3 were excluded from this analysis, as mainstay of treatment for Kaposi sarcoma is medical therapy and radiation. Primary location included sites 'C44.0 through C44.4 corresponding to skin of lip, NOS, eye lid NOS, external ear, skin other/unspecified parts of face and skin of scalp and neck, respectively'. The information was extracted from the SEER dataset (18 registries from 2000-2018). Information regarding patient demographics, grade, stage, size, year of diagnosis, surgical and radiation treatment, and overall survival time until death or loss to follow-up was identified. Information regarding socioeconomic status (SES) and insurance status was extracted using the custom SEER census tract level and rurality database from 2000 to 2016 [16]. Patients with no insurance were grouped together with patients on Medicaid. This was done as patients presenting with no insurance to a healthcare facility are enrolled in Medicaid [17]. Small-area SES was analyzed as a composite index calculated by SEER using the method described by Yost et al. [18] Census tract-level SES indicator variables of median household income, median house value, median rent, percentage of the population below 150% of the poverty line, an education index, percentage of the population with working class occupations and percentage of population older than 16 years in the workforce without a job were utilized [18]. The data are presented as quintiles, group 1 representing the lowest SES and group 5 representing the highest SES. Patients with missing data were excluded from each respective univariable and multivariable analysis. Census urban-area-based categorization was used to stratify the cohort in 'Urban' and 'Rural' groups. 'All rural' and 'mostly rural' were grouped together under 'Rural'. Similarly, 'all urban' and 'mostly urban' were grouped together under 'Urban'.
Patient age was converted to a categorical variable (0-14, 15-39, 40-64, ≥65). We chose this stratification to align with adolescent and young adult population demographics being defined at 15-39 [19,20]. Staging categories of local, regional and distant disease were used according to SEER staging system [21]. Tumor size was also converted into a categorical variable (≤5 cm, >5 cm). Size cut off of 5 cm was used as per AJCC 8th ed recommendation for axial soft tissue sarcoma [22]. Surgical procedures were categorized into 'skin biopsy including Mohs surgery' and 'wide excision including amputation'. The primary outcome in the current investigation is 'overall survival'. SEER* Stat software (version 8.3.8, NCI, Bethesda, MD, USA) was used to analyze incidence rates which were age adjusted and normalized using the 2000 US Standard population using the dataset '18 registries 2000-2018'. Statistical analysis was performed using SPSS Statistical package version 27.0 (SPSS Inc., Chicago, IL, USA). Log-rank test was utilized for categorical values to evaluate the effects of demographic, clinical, pathological, treatment and socioeconomic variables. A multivariable analysis was performed to determine independent predictors of outcome using the Cox proportional hazards model.

Incidence: Sex and Racial/Ethnic Disparities
The incidence of CSTS of the head and neck was 0.27 per 100,000 persons in 2018 and has increased significantly since 2000 ( Figure 1a). The annual percentage change was 1.94 with p < 0.05 (Figure 1a). The incidence of CSTS of the head and neck for male patients was 0.51 per 100,000 persons ( Figure 1b) as compared to 0.092 for female patients, in 2018 ( Figure 1b). Over the study period, males were 5.37 times more likely as compared to females to be diagnosed with CSTS of the head and neck. Non-Hispanic White people had an incidence of 0.325 per 100,000 persons in 2018 (Figure 1c), highest among all racial/ethnic groups. As compared to NHB (incidence 0.078 per 100,000 persons Figure 1c), NHW were 4.62 times more likely to develop CSTS of the head and neck.

Incidence: Sex and Racial/Ethnic Disparities
The incidence of CSTS of the head and neck was 0.27 per 100,000 persons in 2018 and has increased significantly since 2000 ( Figure 1a). The annual percentage change was 1.94 with p < 0.05 (Figure 1a). The incidence of CSTS of the head and neck for male patients was 0.51 per 100,000 persons ( Figure 1b) as compared to 0.092 for female patients, in 2018 ( Figure 1b). Over the study period, males were 5.37 times more likely as compared to females to be diagnosed with CSTS of the head and neck. Non-Hispanic White people had an incidence of 0.325 per 100,000 persons in 2018 (Figure 1c), highest among all racial/ethnic groups. As compared to NHB (incidence 0.078 per 100,000 persons Figure 1c), NHW were 4.62 times more likely to develop CSTS of the head and neck.

Survival and Univariable Analysis
The five-and ten-year overall survival rate for the entire cohort were 0.44 and 0.27, respectively (Table 2). Univariable and multivariable survival analyses are shown in Tables 2 and 3. On univariable analysis 'younger' age (p < 0.001), 'female' sex (p < 0.001), 'NHB' racial/ethnic group (p < 0.001), 'well differentiated' grade (p < 0.001), 'localized' stage (p < 0.001), size of primary tumor less than 5 cm (p = 0.04), histologic broad group of myxosarcoma, liposarcoma and complex mixed and stromal sarcoma (p < 0.001), surgical excision (p < 0.001), surgical procedure 'wide excision including amputation surgery' (p = 0.025), lack of radiotherapy and chemotherapy (p < 0.001) and urban origin of patients (p = 0.038) were significantly associated with improved survival.    Chi-square test of independence was performed to examine the relation between 'SES' and 'surgery'/'surgical procedure'. In the current analysis, the relation was not statistically significant (data not shown).
Due to the high degree of collinearity between 'surgical excision' and 'surgical procedure', we ran multivariable analysis with each of the variables separately. The determination of independent prognostic factors using Cox P-H model yielded similar results with either of the variables. We have presented below the multivariable analysis using 'surgical procedure' variable (Table 3).

Multivariable Analysis: Independent Predictors of Overall Survival
On multivariable analysis (Table 3) age group '40-64 years', 'localized' and 'regional' stage, and 'fibromatous sarcoma' histologic group were independent predictors of improved overall survival. SES groups 1, 3 and 4 were independent predictors of worse overall survival. A representative survival curve is shown in Figure 2. Chi-square test of independence was performed to examine the relation between 'SES' and 'surgery'/'surgical procedure'. In the current analysis, the relation was not statistically significant (data not shown).
Due to the high degree of collinearity between 'surgical excision' and 'surgical procedure', we ran multivariable analysis with each of the variables separately. The determination of independent prognostic factors using Cox P-H model yielded similar results with either of the variables. We have presented below the multivariable analysis using 'surgical procedure' variable (Table 3).

Multivariable Analysis: Independent Predictors of Overall Survival
On multivariable analysis (Table 3) age group '40-64 years', 'localized' and 'regional' stage, and 'fibromatous sarcoma' histologic group were independent predictors of improved overall survival. SES groups 1, 3 and 4 were independent predictors of worse overall survival. A representative survival curve is shown in Figure 2.

Previous Literature
To the best of our knowledge, this is the first population-based study for patients with CSTS of the head and neck, delineating the survival, prognostic factors and disparities in incidence and overall survival. Previous reports consist of single-center-based data focusing on soft tissue sarcoma of the head and neck [3][4][5]. Patients with CSTS of the head and neck were included as a part of the cohort, however, a focused analysis of survival or prognostic factors was not reported. This was in part due to rarity of the disease; few patients with the diagnosis of CSTS of the head and neck precluded any meaningful analysis. Single-center data are also susceptible to selection bias. Rouhani et al. have utilized SEER database to report on incidence and survival of CSTS in the U.S. [6]. Their analysis focuses on the most common histopathological subtypes; however, an analysis focused on anatomical regions is lacking. Others have used utilized population-based data to report on 'sarcomas of the head and neck' and 'leiomyosarcoma of the head and neck' [7,8]. These reports lacked the focus on CSTS. The current report is unique as it focuses on the incidence, survival and associated disparities among patients with CSTS of the head and neck.

Sex and Racial/Ethnic Disparities in Incidence
This is the first report of incidence of CSTS of the head and neck and highlights the sex and racial/ethnic disparities in incidence. Males and NHW are~five times more likely to present with CSTS of the head and neck (Figure 1). Rouhani et al. reported CSTS incidence rate ratio for male to female of 4.7 for malignant fibrous histiocytoma (MFH), 3.7 for leiomyosarcoma (LS), 2.0 for Angiosarcoma (AS), and 0.9 for dermatofibrosarcoma proturans (DFSP) [6]. Our results indicate a slightly higher preponderance of CSTS of the head and neck among males with an incidence rate ratio of 5.37. White people have been found to be more likely to develop MFH, LS and AS, while Black people are more likely to develop Kaposi sarcoma and DFSP [6]. In our analysis, NHW are 4.62 times more likely to develop CSTS of the head and neck as compared to NHB. Of note, we have excluded the cases with diagnosis of Kaposi sarcoma in the current analysis.

Prognostic Factors
Other than 'year of diagnosis' and 'insurance status', all other factors achieved statistical significance on univariable analysis (Table 2). However, using the Cox P-H model for multivariable analysis, only age group '40-64 years', stage other than 'distant' and histologic broad group 'fibromatous sarcoma' were independent protective factors for improved overall survival. Since a majority of patients with CSTS of the head and neck were older than 40 years of age (92.3%, Table 1); age group '40-64 years' represents improved overall survival with younger age at diagnosis. Age group '0-14 years' was censored in the multivariable analysis and the AYA age group '15-39 years' had only 10 patients (Table 3). This finding is consistent with improved outcomes for younger age previously reported by Peng et al. for head and neck sarcomas [7]. Stage other than 'distant' was an independent protective factor for overall survival, as expected, highlighting the importance of early diagnosis and local control to prevent systemic disease.
The  [23]. It is now known as undifferentiated pleomorphic sarcoma (UPS) [23]. However, since the terminology 'MFH' has been used in the SEER database, we have reported it as such for the sake of consistency and to avoid any confusion.

SES Disparities in Outcomes
Previous reports have highlighted the importance of surgical resection and local disease control in the treatment of sarcoma [3][4][5]7,8]. These studies, however, did not explore the impact of SES. In the current study, 'surgical resection' and surgical procedure of 'wide resection including amputation' were statistically significant on univariate analysis (Table 2). However, on multivariable analysis, neither surgical resection, nor surgical procedure, achieved statistical significance (Table 3). On the other hand, our analysis highlighted the prognostic significance of higher SES in patients with CSTS of the head and neck (Table 3). This is a unique finding that has not been previously reported in the literature. This finding highlights a potential issue of access to health care in general and surgical oncology in particular among patients with CSTS of the head and neck. Disparities in access to health care [12,13] and surgery specifically, have been previously implicated for patients with cancer [14,15]. The current analysis, however, did not reveal any correlation between SES and surgery/surgical procedure. Multivariable Cox P-H model reveals loss of significance for treatment modalities (surgery, radiotherapy and chemotherapy) when coanalyzed with SES status, raising the possibility of a correlation between SES and treatment modalities collectively.
Limitations of the current study include lack of any information on specific chemotherapy or any other medical therapy in the SEER database. Similarly, no information regarding any medical history, radiological studies or serological work up is provided in the database limiting our analysis. Epidemiological studies comparing SEER areas to non-SEER areas in the U.S. conclude that their age and sex distributions are comparable except that SEER areas tended to be more affluent and more urban than non-SEER areas. Staging can be a potential pitfall in all studies based on the database as lack of any radiological record makes it impossible to verify the stage at diagnosis.
Despite these limitations, the current study constitutes a significant step towards identification of independent factors associated with improved survival and highlights sex and racial/ethnic disparities in incidence and SES disparities in overall survival. The latter finding further highlights a potential issue with access to health care.