Treatment at Low Volume Hospitals is Associated with Reduced Short- and Long-Term Outcomes for Patients with Retroperitoneal Sarcoma

Abstract

Background

Retroperitoneal sarcomas (RPS) are rare tumors for which complete surgical resection remains the mainstay of treatment. We sought to determine the impact of hospital case volume on RPS outcomes.

Methods

We identified 6950 patients with primary RPS who underwent surgical resection from the National Cancer Data Base (1998–2011). Treating hospitals were classified by annual case volume; low volume hospitals (LVHs) and high volume hospitals (HVHs) were defined as ≤10 and >10 cases/year, respectively. Overall survival (OS) was compared using Kaplan-Meier curves. Cox proportional hazard models were created to compare risks.

Results

Of 1131 reporting hospitals, most (n=1127, 99.6%) were LVHs treating the majority of patients (n=6270; 90.2%). Patients treated at LVHs were more likely to have lower grade and smaller tumors, receive radiation therapy, and undergo incomplete gross (R2) resection. Patients treated at HVHs had lower 30-day readmission rates (1.8% vs 3.4%, p<0.001), 30-day (1.9% vs 3.1%, p=0.004) and 90-day mortality (3.2% vs 5.7%, p=0.007); longer median OS (76.2 vs 64.2 months, p<0.001) and higher 5-year OS (58% vs 52%, p<0.001). After controlling for age, gender, insurance status, tumor size, tumor grade, resection margin status, and radiation administration, treatment at a HVH was independently associated with a reduced risk of death (HR 0.77, 95% CI 0.65–0.91, p=0.003).

Conclusion

Primary RPS are rare tumors for which few surgeons and institutions have significant experience and expertise in their multidisciplinary management and surgical resection. Although additional studies are needed, patient outcomes may be impacted by treating facility case volume and experience.

Introduction

Soft tissue sarcomas (STS) are rare and heterogeneous tumors for which the mainstay of treatment for primary localized disease is complete surgical resection. Of the 12,390 new cases of STS in the United States in 2017,¹ ~15% present in the retroperitoneum.² Retroperitoneal sarcomas (RPS) pose significant technical challenges for the surgeon, given their typically large size and abutment or involvement of adjacent critical structures. Resection of these tumors commonly necessitates multi-organ resection and vascular reconstruction and achieving a macroscopically complete surgical (R0/R1) resection is often difficult.

European and North American expert sarcoma surgeons in conjunction with the European Organization for Research and Treatment of Cancer Soft Tissue and Bone Sarcoma Group have published consensus statements to describe a reproducible and standardized approach to these tumors³ in which they and others recommend that patients with RPS be managed by experienced multidisciplinary teams in specialized sarcoma referral centers.^3–5 Whether treatment at specialized referral centers and high volume hospitals (HVHs) compared to low volume hospitals (LVHs) is associated with better cancer outcomes has been the subject of many studies across disease sites^6–11 and in patients undergoing complex oncologic resections including head/neck surgery,¹² radical cystectomy,¹³ lung resection,¹⁴ esophagectomy,^15,16 colorectal surgery,¹⁷ and hepato-pancreatico-biliary (HPB) surgery.^18–20

Within the sarcoma literature, a small number of retrospective studies have reported an association between sarcoma treatment at either specialty referral centers or HVHs and better patient outcomes. However, these studies are limited as they included 1) STS of multiple disease sites (not limited to the retroperitoneum),²¹ 2) patients who did not undergo RPS resection,²² and 3) often defined hospitals performing as few as 5 RPS resections/year as HVHs.^21,22 The aim of this study was to determine the impact of hospital case volume on outcomes among patients who underwent surgical resection of primary RPS using a large national cancer database.

Materials and Methods

Data source

The National Cancer Database (NCDB) is a hospital-based cancer registry sponsored by the American College of Surgeons and the American Cancer Society that prospectively captures approximately 70% of all new cancer cases in the United States. The data includes clinicopathologic, treatment, and outcome variables and are de-identified. This study was thus considered exempt by The University of Texas MD Anderson Cancer Center Institutional Review Board.

Analytic cohort

Patients diagnosed with primary RPS (1998–2011) were identified from the NCDB Participant Use File (n=12,423) using the International Classification of Diseases for Oncology third edition (ICD-O-3) code C480 (Figure 1). All histologic subtypes included in the data were individually vetted to exclude non-sarcomatous or mixed histologies (Supplemental Table 1).²³ Additional subgroups excluded from the analytical cohort were patients that did not undergo surgical resection, pediatric patients, central nervous system and bone sarcomas, patients not treated at the reporting hospital, and patients with incomplete information. The final cohort (n=6950) was grouped by hospital volume into LVHs (≤10 cases/year) or HVHs (>10 cases/year) (Table 1). We chose not to use the median number of RPS resections performed per year by reporting hospitals as a cutoff to define LVHs versus HVHs because the median number of cases/year was only 1.1. Furthermore, hospitals performing 2 cases/year were among the 90^th percentile and 5.4 cases/year among the 99^th percentile. Thus, we defined 10 RPS resections/year as a cutoff for LVHs versus HVHs as we felt this volume would endow a hospital with sufficient experience and expertise in the treatment of these rare tumors. Using receiver operating characteristic (ROC) curves, we were unable to identify a better cutoff to define LVHs versus HVHs (data not shown).

Figure 1.

Flow chart demonstrating how the final cohort of 6950 patients who underwent surgical resection of primary retroperitoneal sarcoma was obtained.

*All percentages calculated from the total Participant Use File

Table 1.

Demographics and treatment information of patients with primary retroperitoneal sarcoma who underwent surgical resection, stratified by treating hospital volume.

	All (n=6950)		≤10/year (n=6270, 90.2%)		>10/year (n=680, 9.8%)		p value
	N	Col%	N	Col %	N	Col %
Age (median, range)	62 (18–90)		62 (18–90)		60 (19–90)		<0.001
Sex							<0.001
Male	3215	46.3	2853	45.5	362	53.2
Female	3735	53.7	3417	54.5	318	46.8
Race							<0.001
Caucasian	5479	78.8	4901	78.2	578	85.0
African American	667	9.6	631	10.1	36	5.3
Asian/Pacific Islander	304	4.4	280	4.5	24	3.5
Hispanic	445	6.4	410	6.5	35	5.2
Other	55	0.8	48	0.8	7	1.0
Median household income in zip code of residence							<0.001
≥$63,000	2259	32.5	1978	31.6	281	41.3
$48,000–$62,999	1844	26.5	1671	26.7	173	25.4
$38,000–$47,999	1553	22.4	1415	22.6	138	20.3
<$38,000	1095	15.8	1033	16.5	62	9.1
Unknown	199	2.9	173	2.8	26	3.8
Educational attainment in zip code of residence, % graduated high school							<0.001
≥93%	1765	25.4	1535	24.5	230	33.8
87.0–92.9%	2228	32.1	2018	32.2	210	30.9
79–86.9%	1659	23.9	1532	24.4	127	18.7
<79%	1102	15.9	1015	16.2	87	12.8
Unknown	196	2.8	170	2.7	26	3.8
Insurance							<0.001
Private Insurance	3344	48.1	3072	49.0	272	40.0
Medicare	2649	38.1	2478	39.5	171	25.2
Medicaid	328	4.7	309	4.9	19	2.8
Other Government	72	1.0	69	1.1	3	0.4
Not Insured	208	3.0	200	3.2	8	1.2
Unknown	349	5.0	142	2.3	207	30.4
Charlson-Deyo score							0.020
0	3764	54.2	3361	53.6	403	59.26
1	801	11.5	728	11.61	73	10.74
2	211	3.0	200	3.19	11	1.62
Unknown	2174	31.3	1981	31.59	193	28.38
Facility type							<0.001
Academic/Research Program	3806	54.8	3126	49.9	680	100
Comprehensive Community Cancer Program	2702	38.9	2702	43.1	.	.
Community Cancer Program	434	6.2	434	6.9	.	.
Other specified types of cancer programs	8	0.1	8	0.1	.	.
Grade							<0.001
Low	3076	44.3	2815	44.9	261	38.4
Intermediate/High	3874	55.7	3455	55.1	419	62.6
Tumor size							0.048
≤5 cm	594	8.6	550	8.8	44	6.5
5–10 cm	1239	17.8	1119	17.9	120	17.7
>10 cm	4512	64.9	4044	64.5	468	68.8
Unknown	605	8.7	557	8.9	48	7.1
Margin							<0.001
R0/R1	4492	64.6	4087	65.2	405	59.6
R2	291	4.2	280	4.5	11	1.6
Unknown	2167	31.2	1903	30.4	264	38.8
Received chemotherapy							<0.001
No	5763	82.9	5192	82.8	571	84.0
Yes	959	13.8	854	13.6	105	15.4
Unknown	228	3.3	224	3.6	4	0.6
Radiation therapy							<0.001
No	4961	71.4	4398	70.1	563	82.8
Yes	1868	26.9	1751	27.9	117	17.2
Unknown	121	1.7	121	1.9	.	.
Readmission within 30 days of surgical discharge							<0.001
No readmission	4387	63.1	3913	62.4	474	69.7
Readmission	227	3.3	215	3.4	12	1.8
Unknown	2336	33.6	2142	34.2	194	28.5
30 day mortality							0.004
Alive or death >30 days of surgery	6387	91.9	5818	92.8	569	83.7
Death ≤30 days of surgery	359	2.9	191	3.1	13	1.9
Alive with <30 days follow-up or date of last contact unknown	359	5.2	261	4.2	98	14.4
90 day mortality							0.007
Alive or death >90 days of surgery	6355	91.4	5728	91.4	627	92.2
Death ≤90 days of surgery	382	5.5	360	5.7	22	3.2
Alive with <90 days follow-up or date of last contact unknown	213	3.1	182	2.9	31	4.6

R0/R1: complete gross resection

R2: incomplete gross resection

Study variables

Study variables included patient age, gender, race, income, education, insurance status, Charlson-Deyo comorbidity index, treating facility type (academic, comprehensive, or community), hospital volume (# cases/year reporting to the NCDB), tumor grade, tumor size, and therapies received including radiation and/or chemotherapy. In the NCDB, income is estimated by the median household income in the zip code of each patient’s area of residence based on the 2012 American community Survey data. Educational attainment is estimated by the number of adults in the patient’s zip code who graduated from high school based on the 2012 American Community Survey data. Outcomes of interest included rates of complete (R0/R1) RPS resection, 30-day readmission, 30-day mortality, 90-day mortality, and overall survival (OS).

Statistical analysis

Survival data was available for the years 1998–2011. The Kaplan-Meier method and log-rank test were used to calculate and compare the unadjusted OS curves by hospital volume. Univariate and multivariate Cox regression models were used to identify covariates associated with OS. Age, gender, race, insurance, tumor size, surgical margin, tumor grade, radiation treatment, and hospital volume were considered in the univariate Cox regression models; a stepwise method was used to select the final multivariate models. All variables (except chemotherapy administration) included in the model met the proportional hazards function assumption with exclusion of unknown cases from the analysis except for chemotherapy administration, which consistently violated the proportional hazards assumption. Therefore, two separate models for OS were performed: (1) All patients, stratified by chemotherapy status and (2) patients that did not receive chemotherapy. The adjusted survival curves were estimated from the final Cox regression models after adjusting for significant covariates. We performed a logistic regression analysis to determine factors predictive of treatment at a LVH (Supplemental Table 2). We next performed sensitivity analyses to evaluate the reliability of our model. Propensity score (PS) methods (PS stratification, normalized weighting, and covariate adjustment) were applied to the Cox regression model (Supplemental Table 3) with hospital volume as the dependent variable and age at diagnosis, sex, race, insurance, and education as independent variables. The resulting HRs were similar to that of the original model and thus this was retained as our final model. P<0.05 was considered statistically significant. SAS version 9.4 (SAS Institute, Cary, NC) was used to conduct all analyses.

Results

The majority of primary RPS are treated at LVHs

Between 1998–2011, 1131 reporting hospitals performed surgical resection for primary RPS (Table 1). The vast majority of reporting hospitals (n=1127, 99.6%) were LVHs that performed ≤10 primary RPS resections/year. Of the total 6950 patients with primary RPS who underwent surgical resection during the study period, the majority were treated at LVHs (n=6270, 90.2%). A minority of patients (n=680, 9.8%) were treated at the 4 HVHs, defined as facilities which averaged >10 cases/year. Although the number of surgical cases performed for RPS steadily increased from 1998 to 2011, the percentage of resections performed at HVHs remained relatively constant (range 7.6–12.2%) over the study period (Figure 2, Supplemental Table 4).

Figure 2.

Total number of primary retroperitoneal sarcoma surgical resections versus percent of cases resected at high volume hospitals by year.

Patients treated at LVHs versus HVHs differ in their baseline characteristics and treatment modalities

Patient demographics and clinical characteristics and by treating hospital volume are detailed in Table 1. Comparison of patients by treating hospital volume revealed differences between subgroups treated at LVHs versus HVHs.

There were notable socioeconomic differences between RPS patients treated at LVHs versus HVHs. Patients treated at LVHs were more likely to be ethnic minorities rather than white Caucasian and of lower educational background and income (all p<0.001). Additionally, LVHs treated a greater proportion of uninsured (3.2% versus 1.2%) or Medicaid/Medicare (45.4% versus 28%) patients (p<0.001).

HVHs were also distinct from LVHs with respect to patient clinical and tumor characteristics and the multidisciplinary treatment modalities used. HVHs treated patients with RPS that were larger (p=0.048) and of high tumor grade (p<0.001), features associated with poorer clinical outcomes.^4,24,25 Those treated at HVHs waited longer between diagnosis and start of treatment (median 28 versus 5 days, p<0.001) and between diagnosis and definitive surgical resection (median 36 versus 10 days, p<0.001). HVHs were less likely to use radiation therapy (17.2% vs 27.9%, p<0.001), although use of systemic therapy was similar between LVHs and HVHs.

Primary RPS treatment at HVHs is associated with improved outcomes

We next compared outcomes of patients who underwent resection of primary RPS at LVHs versus HVHs. One of the most important prognostic factors for patients with RPS is the ability to resect all macroscopic disease (R0/R1 resection). Surgery at HVHs was associated with fewer incomplete (R2) gross resections (1.6% versus 4.5%, p<0.001).

Compared to patients treated at LVHs, those treated at HVHs had lower 30-day readmission rates (1.8% versus 3.4%, p<0.001), 30-day mortality (1.9% versus 3.1%, p=0.004), and 90-day mortality (3.2% versus 5.7%, p=0.007) (Table 1). RPS treatment at HVHs was also associated with longer median OS (76.2 versus 64.2 months) and 5-year OS (57.7% versus 52.0%) (Figure 3). Additional analyses suggest a dose-effect associated with increasing hospital case volume and better patient outcomes and found progressive improvement in patient outcome with increasing hospital case volume (0–5 per year, 6–10 cases/year, and >10 cases per year; Supplemental Figure 1).

Figure 3.

Overall survival of patients with primary retroperitoneal sarcoma undergoing surgical resection by hospital volume (1998–2011). Survival curves for overall survival among (A) all patients, unadjusted; (B) excluding patients who received chemotherapy, unadjusted; (C–D) adjusted for older age, gender, insurance, larger tumor size, radiation therapy and (C) low & (D) intermediate/high grade RPS.

To examine whether hospital volume was independently associated with OS, univariate and multivariate Cox regression models were used to identify covariates associated with OS. Chemotherapy as a covariate violated the proportional hazards assumption in the model. Patients who received chemotherapy (n=959, 13.8%) differed significantly with respect to multiple factors compared to those that did not receive chemotherapy (n=5763, 90.1%) as part of their multidisciplinary treatment plan (Supplemental Table 5). Thus, analysis was limited to patients who did not receive chemotherapy as part of their RPS treatment (Supplementary Table 6).

On multivariate analysis, treatment at a HVH was associated with a reduced risk of death compared to treatment at a LVH (HR 0.77, 95% CI 0.65–0.91, p=0.003; Table 2). Older age, male gender, larger tumor size, higher tumor grade, and incomplete gross (R2) resection were also associated with worse OS. We found similar results for univariate and multivariate analyses to identify factors predictive of OS among all patients, stratified by chemotherapy status (Supplemental Table 7). We also found similar results when separate analyses were performed limited to patients with Charlson-Deyo score available in the NCDB (2003–2011, n=3524) (Supplemental Table 8).

Table 2.

Cox model for overall survival among patients with primary retroperitoneal sarcoma surgically resected, excluding patients who received chemotherapy (n=4969).

Description	Univariate model				Multivariate model
	HR	95% CI		p value	HR	95% CI		p value
Age (ref. ≤60)				<0.001				<0.001
>60	1.96	1.80	2.14		1.64	1.47	1.84
Gender (ref. Male)				<0.001				<0.001
Female	0.81	0.75	0.88		0.84	0.77	0.91
Race (ref. White)				0.003
African American	0.91	0.79	1.05
Asian/Pacific Islander	0.71	0.56	0.88
Hispanic	0.82	0.68	0.99
Insurance (ref. Private)				<0.001				<0.001
Medicare	1.85	1.70	2.01		1.31	1.17	1.46
Medicaid	1.27	1.03	1.57		1.33	1.08	1.64
Other Government	1.07	0.67	1.70		1.11	0.70	1.77
Not Insured	1.40	1.09	1.81		1.37	1.06	1.77
Tumor size (ref. ≤5 cm)				<0.001				<0.001
5–10 cm	1.07	0.90	1.26		1.03	0.87	1.22
>10 cm	1.29	1.11	1.49		1.28	1.11	1.48
Margin (ref. R0/R1)				<0.001				<0.001
R2	2.34	1.94	2.83		2.45	2.02	2.96
Unknown	1.41	1.29	1.54		1.46	1.34	1.60
Hospital volume (ref. >10/year)				<0.001				0.003
≤10/year	1.27	1.07	1.51		1.30	1.10	1.55
Received radiation therapy (ref. No)				0.006				<0.001
Yes	0.89	0.81	0.97		0.80	0.73	0.88
Grade (ref. Low)				<0.001				<0.001
Intermediate/High	2.05	1.88	2.23		2.24	2.05	2.45

R0/R1: complete gross resection

R2: incomplete gross resection

Discussion

In this study, we examined the effect of hospital case volume on perioperative and survival outcomes in patients undergoing surgical resection for primary RPS using the NCDB. The results of this study demonstrate that treatment of primary RPS at HVHs is associated with reduced perioperative death and improved OS. Unfortunately, however, the vast majority of primary RPS resections during the study period were performed at LVHs that have limited experience in the multidisciplinary care of these rare tumors. It is important to note that although 10 RPS resections/year (<1 case per month) was chosen as the cutoff for LVHs versus HVHs in this study, expert sarcoma referral centers likely have higher volumes than this cutoff.

To our knowledge, this is the first study to investigate the impact of hospital volume on outcomes among patients undergoing surgical resection for RPS. Two prior studies examined hospital volume and outcomes in sarcomas patients, however these were limited for several reasons.^21,22 Gutierrez et al reported outcomes among 4205 patients with STS undergoing surgical resection, finding lower 30- and 90-day mortality in HVHs.²¹ However, the majority of patients included had STS of the extremity, trunk, and head/neck locations and <42% of patients included had RPS. Also using the NCDB, Maurice et al reported outcomes among 3141 RPS patients, however, 30% of patients in this study did not undergo surgical resection and HVH was defined as a hospital performing at least 5 resections/year.²² RPS treatment at HVHs in this study was associated with higher rates of surgical resection and R0/R1 resection, however the authors did not find a survival difference between patients treated at HVHs and LVHs.

Although the association between hospital volume and patient outcomes has only recently begun to be examined for sarcomas, it has been extensively examined across other patient populations.^{6–8,11,20,26} Compared to treatment at LVHs, treatment at HVHs has been associated with improved patient outcomes following various complex oncologic resections including head/neck surgery,¹² esophagectomy,^9,15,16 colorectal surgery,^9,17,27 HPB,^9,19,20,28 cystectomy,¹³ nephrectomy,¹⁰ and major lung resection.¹⁴

The relationship between hospital volume and patient outcomes may reflect differences in multiple aspects of the multidisciplinary care of the oncology patient undergoing complex surgical resection, such as the patient with primary resectable RPS. These may include not only differences in surgeon volume and experience at HVHs which may result in differences in surgical approach to RPS resection, completeness of surgical resection, and quality of surgery but also extend to differences in how multimodality therapy is used, how patients are surveyed, and how disease recurrence is managed.

The impact of surgeon volume on patient outcomes has been demonstrated in other patient populations such as following adrenalectomy,²⁹ thyroidectomy,³⁰ and HPB surgery.^18,31 In this study, despite limitations of data available in the NCDB regarding perioperative complications, there are immediate differences in quality of surgery with higher rates of incomplete (R2) RPS resection occurring at LVHs compared to HVHs. Additionally, there is significant differences in utilization of radiation therapy for RPS between HVHs and LVHs. Finally, surgical resection at HVHs compared to LVHs is associated with significantly lower 30-day readmission.

There are also data across disease sites to suggest that HVHs may be better able to “rescue” patients following postoperative complications. Failure to rescue (FTR) is defined as operative mortality after a major perioperative complication and has been demonstrated to differ between HVHs and LVHs in other patient populations, including patients undergoing high-risk cancer operations such as gastrectomy, esophagectomy, pancreatectomy, and liver resection.^18,32–34 In this study, although FTR cannot be determined using data available in the NCDB, it might be extrapolated that LVHs may have had higher FTR rates based on the findings of higher 30- and 90-day mortality rates.

Lastly, studies have suggested that there may be differences in rates of adherence to sarcoma treatment guidelines between HVHs and LVHs. Incomplete adherence to sarcoma treatment guidelines has consistently been shown to be associated with worse outcomes in studies of patient populations in the United States³⁵ and Europe.^36,37 Furthermore, these studies report that treatment within specialized sarcoma centers, HVHs, and within a cancer network is more likely to adhere to treatment guidelines.^38–40

Despite a significant body of literature demonstrating the association between care at HVHs and improved outcomes, a significant proportion of patients seek care at LVHs. Although some patients may prefer to undergo a complex surgical procedure at a local LVH,⁴¹ there is concern that socioeconomic barriers may prevent many patients from accessing care at HVHs and specialized referral centers.^20,42,43 In this study, demographic data available in the NCDB suggests that patients who underwent surgery at LVHs may be of lower socioeconomic status, raising concerns that socio-economic circumstances may be barriers to patients’ ability to access care at HVHs for RPS.

There are several limitations of this study, including those inherent to retrospective studies using a large national database. Data available through the NCDB are submitted by participating hospitals across the US and thus not only reflect potential differences in RPS management across institutions but may also reflect inconsistencies in pathologic diagnosis of RPS histologies and data entry by research personnel. Furthermore, for each variable examined, a minority of cases within the analytic cohort had missing or insufficient data for analysis. Additionally, the NCDB does not include data pertaining to disease recurrence/progression, details of systemic chemotherapy administered or radiation therapy regimens. Despite these limitations, however, the NCDB remains a valuable resource to study patient- and hospital-related factors that may impact patient care and oncologic outcomes particularly in rare malignancies such as primary RPS.

Conclusion

Multidisciplinary care of the patient with RPS is complex. Although additional studies are needed, the results of the current study suggest that outcomes among patients with primary RPS undergoing surgical resection may be impacted by treating facility case volume and experience. Furthermore, patients who underwent surgery at LVHs in this study were of lower socioeconomic status, raising concerns that financial and social circumstances may be clinically relevant barriers to patients’ ability to seek and access care at HVHs for RPS.

Supplementary Material

Supp TableS1-8

Supplemental Table 1. International Classification of Diseases for Oncology (ICD-O-3) topographical and morphological codes included.

Supplemental Table 2. Factors associated with treatment at a low volume hospital (≤ 10 cases/year)

Supplemental Table 3. Sensitivity analysis of hospital volume - results of various propensity score methods

Supplemental Table 4. Total number of primary retroperitoneal sarcoma surgical resections versus percent of cases resected at high volume hospitals by year.

Supplemental Table 5. Demographics and treatment information of patients with primary retroperitoneal sarcoma who underwent surgical resection, stratified by chemotherapy status.

Supplemental Table 6. Demographics and treatment information of patients with primary retroperitoneal sarcoma who underwent surgical resection (without chemotherapy), stratified by treating hospital volume.

Supplemental Table 7. Cox model for overall survival among patients with primary retroperitoneal sarcoma surgically resected, stratified by chemotherapy status.

Supplemental Table 8. Cox model for overall survival among patients with primary retroperitoneal sarcoma surgically resected with Charlson-Deyo score available (2003–2011 only), excluding patients who received chemotherapy (n=3524).

Supp figS1

Supplemental Figure 1. Overall survival of patients with primary retroperitoneal sarcoma undergoing surgical resection by hospital volume (1998–2011). Survival curves for overall survival among all patients adjusted for older age, gender, insurance, larger tumor size, radiation therapy and (A) low and (B) intermediate/high grade RPS.

Synopsis.

Few institutions have significant experience in the multidisciplinary management of retroperitoneal sarcomas (RPS), rare tumors that often require complex multi-visceral resection. Treatment of primary RPS at high volume hospitals is associated with improved patient outcomes.