Racial and ethnic variation in presentation and outcomes of high-grade soft tissue sarcoma at a Southeastern United States comprehensive cancer center

Abstract

Background:

High-grade soft tissue sarcoma is rare and associated with poor prognosis. This study examines racial and ethnic variation in presentation and outcomes at a Southeastern US cancer center.

Methods:

Among an institutional cohort of patients seen between January 2016–December 2021, racial and ethnic differences were evaluated using chi-squared tests, Kaplan Meier curves, and Cox proportional hazards models.

Results:

There were 295 patients (71 % Non-Hispanic White, 24 % Black, 3 % Hispanic White, 2 % Other). Black representation was greater than national cohorts (24 % vs. 12 %). Histological subtype varied by race/ethnicity (p = 0.007). Adjusting for histology and stage, survival was worse for Black vs. White patients (HR 1.71, 95 % CI 1.07–2.76) and those with metastatic disease (5.47, 3.54–8.44). In non-metastatic patients, survival differences for Black vs. White patients were attenuated by receipt of multi-modal treatment (1.53, 0.82–2.88).

Conclusion:

Observed racial disparities in survival of high-grade sarcoma may be addressed by early, multidisciplinary management.

1. Introduction

Soft tissue sarcomas are a heterogeneous group of mesenchymal tumors consisting of numerous histological subtypes that can arise due to genetic predisposition, environmental and occupational exposures, or secondary to other treatments including external radiation therapy.^1,2 Soft tissue sarcomas can occur anywhere in the body, most often the extremities, trunk, retroperitoneum, and head and neck. These tumors can present as an asymptomatic mass or be accompanied by symptoms due to tumor distortion or infiltration of normal structures.¹ Prognosis is based primarily on tumor histology as well as tumor size and grade, with high-grade tumors being most aggressive with highest risk for recurrence, metastasis, and death. For this reason, high-grade tumors benefit from multi-disciplinary evaluation and multi-modal therapy.^3,4

Given the rarity of these tumors, representing less than 1 % of malignancies,^2,5 there has been limited study of the association between demographics, patient presentation, and prognosis.^6–9 Prior work has demonstrated racial and ethnic variation in the reported incidence of soft tissue sarcoma histological subtypes in adolescents and young adults.⁹ Among an adult sarcoma cohort in the Surveillance, Epidemiology, and End Results (SEER) database, there was a stronger association between race/ethnicity and sarcoma etiology than socioeconomic status, highlighting the importance of genetic variation associated with ancestry.⁶ These studies report on incidence, but not outcomes of soft tissue sarcoma based on race and ethnicity, which are difficult to study using national databases due to lack of treatment details and oncologic outcomes.

The population of our Southeastern US cancer center is unique in having a large Black population, including a high proportion of rural-dwelling Black patients. For this reason, we sought to examine racial and ethnic variation in presentation and outcomes of high-grade soft tissue sarcoma in our cohort of adult patients to understand potential opportunities to address existing disparities.

2. Materials and methods

The study cohort included adult patients with primary, high-grade soft tissue sarcoma who were seen at our comprehensive cancer center between January 2016 and December 2021. The University of Alabama at Birmingham O’Neal Comprehensive Cancer Center is the only NCI-designated comprehensive cancer center in the state and effectively serves the state of Alabama as well as parts of Mississippi, Georgia, and Florida. Patients were identified using a combination of cancer registry and billing data to ensure complete case capture. Data were collected via electronic medical record review including demographic, histologic, stage, treatment, recurrence, and survival information. Patients did not have to receive all aspects of cancer care at UAB to be included in the study but did need to have external treatment information (e.g. details of radiation treatment provided locally) available for review in the UAB medical record. Race and ethnicity were categorized as Non-Hispanic White (White), Non-Hispanic Black (Black), Hispanic White, Hispanic Black, and Other. Race and ethnicity were based on self-report as documented in the electronic medical record (EMR) and were limited to these categories. Treatment included primary site surgery, radiation, and systemic therapy. All data were generated in the course of routine care and were abstracted secondarily for research purposes. The study was approved by the UAB Institutional Review Board.

Zip code level social vulnerability was measured using the Center for Disease Control and Prevention Social Vulnerability Index (SVI).¹⁰ SVI is reported at the census tract level, and SVI scores range from 0.0 to 1.0 with higher scores indicating greater social vulnerability. The SVI ranks each census tract on 16 social factors from American Community Survey data, and these factors are grouped into four related domains. The domains include (1) socioeconomic status (below 150 % poverty, unemployed, housing cost burden, no high school diploma, no health insurance), (2) household characteristics (aged 65 or older, aged 17 or younger, civilian with a disability, single-parent households, English language proficiency), (3) racial and ethnic minority status, and (4) housing type & transportation (multi-unit structures, mobile homes, crowding, no vehicle, group quarters).¹⁰ Domains 1, 2, and 4 were included while domain 3 (race and ethnicity) was excluded from the multivariable models due to co-linearity with our race and ethnicity exposure variables. Because only zip codes were available for our cohort, we obtained zip code level estimates of SVI by calculating mean weighted values for each zip code based on its constituent census tracts.

Racial and ethnic differences were evaluated using chi-squared and Wilcoxon rank sum tests. For survival analyses, only patients of Black and White race were included due to low cohort sizes for other racial and ethnic groups. We created Kaplan Meier curves to compare overall survival and used Cox proportional hazards models to determine specific factors associated with survival. Covariates included in the Cox proportional hazards models were age, metastatic versus non-metastatic at presentation, histology category, and the three SVI domains (1,2, and 4). Separate analyses were performed for all patients and then limited to those with non-metastatic disease at presentation to examine the effect of primary tumor treatment on survival. Results are reported as Hazard Ratios (HR) with 95 % confidence intervals.

3. Results

A total of 295 patients were included. The cohort was 71 % White, 24 % Black, 3 % Hispanic White, and 2 % Other. There were no Hispanic Black patients. The “Other” category included all other racial/ethnic groups and included Asian (N = 2), American Indian or Alaskan Native (N = 3), and unknown (N = 3). Patient characteristics, tumor characteristics, and sarcoma histologies for the overall cohort are delineated in (Table 1). The median age of the cohort was 63 years. Overall, 82 % of the cohort had no prior personal cancer history, and only 1 % of the cohort had any known family history of sarcoma. The most common histologies included leiomyosarcoma (LMS) (22 %), undifferentiated pleomorphic sarcoma (UPS) (27 %), and liposarcoma (LPS) (15 %). There were significant differences in sarcoma histology by race (p = 0.007) as depicted in (Fig. 1). The most common tumor histologies seen in Black patients (N = 70) were LMS (30 %), UPS (21 %), and Other (23 %) which included rhabdomyosarcoma, myxofibrosarcoma, and sarcoma not otherwise specified (NOS). In the White cohort (N = 210), UPS (30 %) was the most common histology. All angiosarcoma cases were in White patients.

Table 1.

Overall cohort patient characteristics, tumor characteristics, and sarcoma histologies.

	Overall (N = 295)
Patient Characteristics
Age at diagnosis (years)	63 (52–72)
Non–metastatic at presentation	243 (82.4 %)
Female	157 (53.2 %)
BMI, median	28.5 (25.0–33.5)
Prior history of unrelated cancer	53 (18.0 %)
Family history of sarcoma	3 (1.0 %)
Soft Tissue Sarcoma Histologies
Angiosarcoma	14 (4.7 %)
Leiomyosarcoma	66 (22.4 %)
Liposarcoma	45 (15.3 %)
Myxofibrosarcoma	20 (6.8 %)
Undifferentiated Pleomorphic Sarcoma	79 (26.8 %)
Spindle Cell Sarcoma	23 (7.8 %)
Other	45 (15.3 %)
Missing	2 (0.7 %)
Tumor Characteristics
Tumor size (cm)	8.9 (5.0–14.7)
Mitoses present	112 (38.0 %)
Necrosis	115 (39.0 %)
Hemorrhage	61 (20.7 %)

^*Data are presented as median (IQR) for continuous measures and n (%) for categorical measures.

^**“Other” histologies include but are not limited to synovial sarcoma, sarcoma NOS, rhabdomyosarcoma, malignant peripheral nerve sheath tumor, and epithelioid sarcoma.

Fig. 1.

Sarcoma histology by race (p = 0.007).

The majority of patients presented with localized disease at 82 % (N = 243), compared to 18 % (N = 52) of patients who presented with metastatic disease. Within the localized subgroup, 7 % received neoadjuvant chemotherapy, 46 % received perioperative radiation, and 91 % underwent surgical resection. In a median follow up time of 493 days (IQR 247–1125), 19 % had local recurrence with a median time to recurrence of 331 days (IQR 161–610) and 28 % developed metastatic disease at a median of 262 (IQR 105–497). Thirty percent died during follow-up with a median time to death of 420 days (IQR 205–718). Of the patients who presented with metastatic disease, 50 % of these patients received chemotherapy and 17 % received palliative radiation. In 256 days (IQR 87–430) follow-up time, 71 % were reported to be deceased by December 2022 with a median time to death of 205 days (IQR 95–347).

Due to retroperitoneal sarcomas historically having a worse overall prognosis, treatment information based on sarcoma location was also calculated (Supplemental Table 1). There was no difference in receipt of systemic therapy (p = 0.94), perioperative radiation (p = 0.18), or surgical resection (p = 0.06) based on tumor location. Because there were few patients of Hispanic ethnicity or Other race, comparative analyses were limited to presentation and outcomes for Black versus White patients (Table 2). Black patients were more likely to be female and present at an earlier age than White patients (p = 0.01). White patients were more likely to have a history of another unrelated cancer (p = 0.02). SVI also differed for Black versus White patients in all 3 of the tested domains (Table 2). Mean SVI was higher in the Black population in the socioeconomic domain (0.77 vs. 0.56, p < 0.001), household characteristic domain (0.71 vs. 0.64, p = 0.02), and housing type and transportation domain (0.47 vs. 0.46, p = 0.03).

Table 2.

Data stratified by Black versus White race.

All Patients	White (N = 219)	Black (N = 70)	p-value
Patient Characteristics
Age at diagnosis (years)	65 (52–73)	58 (51–66)	0.01
Female	104 (47.5 %)	49 (70.0 %)	0.01
BMI, median	28.1 (24.6–32.3)	29.3 (25.8–35.2)	0.06
Prior history of unrelated cancer	46 (21.0 %)	6 (8.6 %)	0.02
Tumor Characteristics
Tumor size (cm)	8.5 (4.5–14.5)	9.5 (6.3–15.4)	0.20
Mitoses	84 (38.3)	25 (35.7)	0.93
Necrosis	78 (35.6 %)	34 (48.6 %)	0.13
Hemorrhage	42 (19.2 %)	17 (24.3 %)	0.14
Stage at Diagnosis			0.82
Non-metastatic	181 (82.6 %)	57 (81.4 %)
Metastatic	38 (17.4 %)	13 (18.6 %)
Treatment, n (%)
Neoadjuvant chemotherapy	14 (6.4 %)	8 (11.4 %)	0.38
Surgery	191 (87.2 %)	57 (81.4 %)	0.42
Radiation	91 (41.6 %)	27 (38.6 %)	0.66
Outcomes
Follow-up time (days)	474 (247–1075)	322 (134–704)	0.01
Death	77 (35.2 %)	30 (42.9 %)	0.25

For Black vs. White patients who presented without metastatic disease, there was no difference in receipt of neoadjuvant chemo (p = 0.08), radiation (p = 0.66), or surgery (p = 0.23) (Table 3).

Table 3.

Data for patients who were non-metastatic at diagnosis stratified by Black versus White race.

Non-metastatic at Diagnosis	White (N = 181)	Black (N = 57)	p-value
Patient Characteristics
Age at diagnosis (years)	65 (53–74)	58 (52–67)	0.01
Female	85 (47.0 %)	37 (64.9 %)	0.06
BMI, median	28.3 (24.9–32.3)	29.0 (25.6–35.2)	0.18
Prior history of unrelated cancer	41 (22.7 %)	5 (8.8 %)	0.02
Treatment
Neoadjuvant chemotherapy	10 (5.5 %)	7 (12.3 %)	0.08
Surgical resection	168 (92.8 %)	50 (87.7 %)	0.23
Perioperative radiation	82 (45.3 %)	26 (45.6 %)	0.66
Outcomes
Local recurrence	41 (22.7 %)	5 (8.8 %)	0.06
Follow-up time (days)	549 (277–1191)	374 (217–896)	0.04
Death	51 (28.2 %)	19 (33.3 %)	0.52

^*Data are presented as median (IQR) for continuous measures and n (%) for categorical measures.

In the adjusted analysis of overall survival, Black patients had worse overall survival (HR 1.71, 95 % CI 1.07–2.76) (Fig. 2). Other factors associated with survival included metastatic disease at diagnosis (HR 5.47, 95 % CI 3.54–8.44) as well as LMS (HR 0.54, 95 % CI 0.31–0.93) and LPS (HR 0.53, 95 % CI 0.29–0.97) histological subtypes (Fig. 2). The association between Black race and worse overall survival persisted in the model limited to patients without metastases at presentation that adjusted for age, histology, and SVI domains (HR 1.78, 95 % CI 1.00–3.17). However, after adjustment for treatment factors including perioperative radiation, neoadjuvant chemotherapy, and surgical resection, the observed difference by race was no longer demonstrated (HR 1.53, 95 % CI 0.82–2.88) (Fig. 3). Further, all three treatment modalities were associated with improved survival – surgery (HR 0.07, 95 % CI 0.03–0.20), radiation (HR 0.53, 95 % CI 0.31–0.90), and systemic therapy (HR 0.27, 95 % CI 0.09–0.83).

Fig. 2.

Likelihood of death by patient and tumor characteristics. *Model also adjusted for SVI domains 1, 2, and 4 (not shown).

Fig. 3.

Likelihood of death in patients presenting initally without metastatic disease. *Model also adjusted for SVI domains 1, 2, and 4 (not shown).

4. Discussion

In this institutional cohort of patients treated at an NCI-designated comprehensive cancer center in the Southeastern US, there were significant differences in high-grade soft tissue sarcoma histology and outcomes by race. Specifically, findings demonstrated worse overall survival for Black versus White patients. Among patients who presented with non-metastatic disease, the observed differences in survival between Black and White patients were attenuated after adjusting for receipt of multimodal therapy including surgery, radiation, and systemic therapy.

Multi-disciplinary care provided at our institution included assessment at a Multidisciplinary Clinic attended by surgical oncologists, medical oncologists, and radiation oncologists as well as case discussion in weekly Multidisciplinary Tumor Board attended by the same provider groups along with pathologists and radiologists who have specific expertise in sarcoma. This study demonstrated that observed racial disparities in sarcoma outcomes were no longer observed in the patients who received multimodal treatment, highlighting the potential value of this institutional strategy in reducing observed disparities.

While several studies have similarly demonstrated variation in soft tissue sarcoma histology by race,^7,9 there continues to be conflicting evidence regarding the role race plays in the outcomes of these aggressive neoplasms.^11–14 Previous studies have utilized a large national cancer database or combination of several smaller databases, and have similarly found differences in survival for Black versus White patients.^12–14 Still others found no racial differences in sarcoma survival (10). These studies have more homogenous populations and are limited in their ability to examine the effect of treatment on observed disparities. Our cohort is distinct in that our region had a larger Black population than seen nationally (12.1 % vs. 24 %), increasing power to detect differences by race.

The ability to include treatment details in survival analyses enables a more nuanced understanding of contributors to observed survival differences. Our adjusted model of survival among patients initially presenting with non-metastatic disease included receipt of multi-modal therapy (surgery, radiation and systemic treatment). In the adjusted model, we found that surgery, radiation, and systemic therapy reduced the likelihood of death. This finding emphasizes the importance of multimodal treatment in mitigating racial disparities in outcomes for high-grade sarcoma.¹⁵ Both earlier treatment and a multidisciplinary approach have proven beneficial in the context of other cancer types and remain an area for improvement in sarcoma management.^16–18 Future work should focus on early cancer detection (i.e. presentation when non-metastatic) and referral to multidisciplinary sarcoma teams.

The thirty percent death rate seen in our initially non-metastatic cohort, though high, is not largely different than what we would expect to find. Up to half of originally localized soft tissue sarcomas will go on to develop metastatic disease with initial tumor grade being the most important prognostic factor.¹⁹ As this cohort consists entirely of high-grade sarcomas, we would expect survival to be low even in those initially presenting without metastatic disease. Overall, a 10-year overall survival of 50 % has been in demonstrated in localized resectable STS, and once metastasis is detected, median survival is approximately twelve months.^19,20

A limitation of this study is inability to capture complete treatment or outcomes data for patients who received a portion of their care outside our institution. Our group recently demonstrated in another study done by Montgomery et al. that after consultation with our team, surgery was performed outside of our institution in less than 2 % of cases, systemic therapy in less than 3 % of cases, and radiation in 14.3 % of cases.²¹ Another persistent barrier identified in this work was that 52 % of patients had delays of care which was defined as it taking more than 21 days to receive initial treatment, with 17 % of treatment delays attributable to patient-related factors (e.g. transportation and insurance) and the majority to system-level factors related to obtaining pertinent studies and coordination of care.²¹ Finally, we did not document the proportion of patients who received all recommended elements of a given treatment plan, and appreciate that failure to receive intended therapy may be due to both tumor-specific and patient-specific factors.

Additional study limitations include its retrospective nature. While we were able to include area-level social determinants of health, we were not able to obtain data on individual-level social determinants of health such as socioeconomic status, education, and travel distance which would provide greater nuance in the adjusted analyses. Future multi-institutional prospective studies with robust collection of individual-level social determinants of health could provide better understanding of the root causes of observed racial disparities in sarcoma outcomes.

5. Conclusion

Observed racial disparities in survival of high-grade sarcoma may be addressed by further utilization of a multi-modal approach. Future work will focus on addressing these inequities through assessment of individual barriers and early referral to multidisciplinary sarcoma teams.