Pathologic Nodal Staging for Clinically Node Negative Soft Tissue Sarcoma of the Extremities

Abstract

Background and Objectives:

Synovial, clear cell, angiosarcoma, rhabdomyosarcoma, and epithelioid (SCARE) soft tissue sarcoma are at risk for nodal involvement, although the nodal positivity rates and impact on prognostication in clinically node negative patients are not well described.

Methods:

Patients with extremity SCARE sarcoma without clinical nodal involvement undergoing surgical resection in the National Cancer Database (2004–2017) were included. Logistic regression was used to evaluate the likelihood of nodal surgery and nodal positivity. Kaplan-Meier method and Cox regression were used to assess associations of nodal status to overall survival.

Results:

We included 4,158 patients, and 669 patients (16%) underwent regional lymph node surgery (RLNS). On multivariable logistic analysis, patients with epithelioid (OR 3.77, p< 0.001) and clear cell (OR 6.38, p<0.001) were most likely to undergo RLNS. 45 patients (7%) had positive nodes. Clear cell sarcoma (14%) and angiosarcoma (13%) had the highest rates of nodal positivity. Patients with positive nodes had reduced 5-year overall survival, and the stratification was largest in clear cell and angiosarcoma.

Conclusions:

Discordance exists between selection for pathologic nodal evaluation and factors associated with nodal positivity. Clinically node negative patients with clear cell and angiosarcoma should be considered for pathologic nodal evaluation.

INTRODUCTION

Soft tissue sarcoma (STS) is an uncommon malignancy with an estimated 13,130 new cases in 2020 in the United States.¹ Classically, these tumors are thought to metastasize by a hematogenous route, most commonly to the lungs,² and not to have lymphatic spread in the same way as carcinoma. For a minority of patients, regional nodal spread has been documented, with estimates generally ranging from 3–10% for all sarcoma subtypes,^3–9 but some have noted incidence as low as 0.9%.¹⁰ However, a subset of STS histologies, known as the SCARE histologies (Synovial sarcoma, Clear cell sarcoma, Angiosarcoma, Rhabdomyosarcoma and Epithelioid sarcoma) demonstrate a higher propensity for regional lymph node spread.^3,4,6,9,11 This possibility has important clinical implications as regional lymphatic spread in STS in the absence of distant metastases has been identified as a factor predicting worse overall survival.^4,6,10,12,13. A prior study examining factors associated with nodal positivity in this database included clinically node positive patients, which may be distinct from microscopically positive disease in clinically node negative patients, and combined the CARE histologies into one group, which did not allow for a discrete examination of the diverse biology under the soft tissue sarcoma umbrella.¹⁴ Therefore, regional lymph node evaluation to identify clinically occult disease may be important for prognosis and subsequent management for certain subtypes at high risk of regional lymphatic spread.

In contrast to melanoma and breast cancer, the role of pathologic nodal staging in adults with STS is not well defined. Prior studies have evaluated the prognostic role of nodal status in high risk STS histologies which have indicated that a positive lymph node predicts worse survival, but these are primarily limited to smaller institutional studies.^11,15–17 There is a paucity of studies with sufficient patients and statistical power to describe the prevalence and impact of positive regional lymph nodes among sarcoma histologies with differing biologies.^6,10,18 Additionally, the way in which patients are selected for nodal evaluation in clinically node negative disease, how that corresponds with probability of positive nodes, and how that is associated with histology specific outcome is not known. In this manuscript, we evaluate factors associated with receipt of regional lymph node surgery (RLNS), factors associated with pathologically positive lymph nodes, and the effect of lymph node status on histology specific survival in order to identify patients that should be considered for pathologic nodal evaluation.

MATERIALS AND METHODS

Data Source

We used the National Cancer Database (NCDB). The NCDB is a joint project of the American Cancer Society and the Commission on Cancer of the American College of Surgeons. The NCDB was established in 1989, and is a nationwide, facility-based comprehensive clinical surveillance oncology dataset pulled from more than 1500 Commission on Cancer accredited facilities, capturing over 70% of all newly diagnosed cancer cases each year.¹⁹

The database of all incident cases of soft tissue neoplasms from 2004 – 2017 was queried for patients diagnosed with non-metastatic extremity sarcoma with SCARE histology without clinically node positive disease who underwent surgery. Selections were made using International Classification of Disease for Oncology, Third Edition [ICD-O-3] primary site codes C491 and C492 and morphologic codes 8804/3, 8900–2/3, 9040–4/3, 9120/3. Patients were excluded if they had missing data on vital status, did not undergo surgery or regional lymph node surgery, had incomplete/missing data on surgery, or clinically positive nodal disease. After inclusion criteria were applied, a total of 4,158 cases were used for analysis. A CONSORT diagram of the inclusion/exclusion criteria is available (Supplemental Figure 1). The study was reviewed and deemed exempt by the University of North Carolina Institutional Review Board.

The primary exposure of interest was histology. Our focus was on the SCARE histologies – Synovial, Clear cell, Angiosarcoma, Rhabdomyosarcoma, Epithelioid. Other covariates of interest included age at diagnosis, race, ethnicity, site of primary lesion, sarcoma T stage, insurance status, tumor size, sarcoma grade, type of surgery, zip-code associated socioeconomic status, stage of disease, receipt of chemotherapy, receipt of radiation therapy, and overall survival.

Age was divided into 4 categories: <18, 18 – 39, 40–64, >=65 years. Sarcoma T category was created based on tumor size as follows: <=5cm, >5cm to <=10cm, >10cm to <=15cm, >15cm, and a category for those with unknown tumor size. NCDB also includes information on zip code levels of educational attainment (% of adults with less than a high school education) and zip code median household income from the 2012 American Community Survey (collected 2008–2012). Both variables are initially reported as quartiles (low, medium, high, and highest), and for analysis we combined into a single low, medium, and high SES variable. SES was considered “Low” if both educational attainment and median income were in the lowest quartiles, “High” if both educational attainment and median income were in the highest quartiles, and “Medium” if the both of the above criteria were not met. Charlson-Deyo score was used to create a binary variable “No comorbidities” where the answer is “yes” for Charlson score of zero and “no” for Charlson score ≥1. Facility type and location were not included in the analysis because the NCDB suppresses facility information on patients under the age of 40 and they were included in this analysis.

The outcomes of interest were receipt of regional lymph node surgery, likelihood of positive node after regional lymph node surgery, and overall survival as defined by the number of months from diagnosis to last contact (censored) or death. Variables specific to sentinel node procedures are only available for cases diagnosed after 2012, a focus which would have severely limited our sample size in these rare histologies. We used the available general regional lymph node surgery variable to include more cases in the analysis.²⁰

Descriptive statistics were performed on all variables included for the study. Frequency distributions were provided on all categorical variables. Mean and standard deviations were calculated for the continuous variables.

Multivariable logistic regression, adjusted for age, sex, race, comorbidities, sarcoma grade, insurance status, zip-code level SES, tumor size category was used to assess associations between histology and (i) receipt of lymph node surgery in the entire cohort and (ii) positive lymph node in those who underwent lymph node surgery. Variables were selected a priori based on clinical relevance and interpretability. Using the model results, we also calculated adjusted probabilities for receipt of lymph node surgery and positive lymph nodes by histologic subtype.

Kaplan-Meier curves were used to calculate 5-year overall survival and illustrate overall survival. Survival was estimated from the time of diagnosis until death or last follow-up. Multivariable Cox proportional hazards regression was used to examine the association between histology and overall survival, after adjusting for patient and tumor factors. Factors which did not meet the proportional hazards assumption were stratified in the analysis with all other covariates in the model. As overall survival data was only available through 2016, those diagnosed in 2017 were not included in this analysis.

Statistical analyses were carried out using STATA version 16.1 (STATA, College Station, TX, USA). According to NCDB guidelines, analyses which would result in reporting of cell counts <10 were suppressed and omitted from the report. Statistical significance was defined as p<0.05.

RESULTS

Patient Characteristics

We identified 4,158 patients with a SCARE histology soft tissue sarcoma of the extremities without clinical evidence of nodal involvement who underwent resection for non-metastatic disease in the NCDB from 2004–2017, Table 1. In those patients there were 1,308 recorded deaths. The median age was 44 years and 3,378 (81%) patients were white. The majority of patients (86%) had no comorbidities recorded. Median overall follow up was 53.3 months (IQR 27.1–92.4). Tumor histology breakdown was as follows: 2,400 (58%) synovial, 306 (7%) clear cell, 529 (13%) angiosarcoma, 387 (9%) rhabdomyosarcoma, and 536 (13%) epithelioid. The majority of tumors were in the lower extremity, 2,790 (67%), compared to the upper extremity. Forty-one percent of tumors were less than 5 cm, 26% 5–10 cm, 9% 10–15 cm, and 6% greater than 15 cm. Six hundred and forty (15%) patients were surgically managed with amputation, 2,118 (51%) received external beam radiation, and the 1,313 (32%) patients received chemotherapy.

Table 1:

Baseline characteristics of sarcoma patients in the National Cancer Database 2004–2017 overall and by regional lymph node surgery

Cohort Characteristics	No RLNS N=3,489	RLNS N=669	Total N=4,158	P-Value
Age at diagnosis (years), median (IQR)	44.0 (28.0–61.0)	34.0 (22.0–52.0)	43.0 (27.0–59.0)	<0.001
Sex, n(%)				0.014
Female	1,808 (51.8%)	312 (46.6%)	2,120 (51.0%)
Male	1,681 (48.2%)	357 (53.4%)	2,038 (49.0%)
Race, n(%)				0.85
White	2,832 (81.2%)	546 (81.6%)	3,378 (81.2%)
Black	418 (12.0%)	82 (12.3%)	500 (12.0%)
American Indian/Indigenous	18 (0.5%)	<10	21 (0.5%)
Asian/Pacific Islander	116 (3.3%)	22 (3.3%)	138 (3.3%)
Other	49 (1.4%)	<10	54 (1.3%)
Unknown	56 (1.6%)	11 (1.6%)	67 (1.6%)
Hispanic Ethnicity, n(%)	407 (12.2%)	79 (12.4%)	486 (12.2%)	0.89
Primary Payor, n(%)				<0.001
Not Insured	155 (4.4%)	35 (5.2%)	190 (4.6%)
Private Insurance	2,100 (60.2%)	446 (66.7%)	2,546 (61.2%)
Medicaid	398 (11.4%)	86 (12.9%)	484 (11.6%)
Medicare	702 (20.1%)	71 (10.6%)	773 (18.6%)
Other Government	70 (2.0%)	13 (1.9%)	83 (2.0%)
Insurance Status Unknown	64 (1.8%)	18 (2.7%)	82 (2.0%)
Great circle distance (miles), median (IQR)	17.8 (7.1–52.0)	17.6 (7.4–48.1)	17.8 (7.2–51.1)	0.97
Charlson-Deyo Score, n(%)				0.31
≥1	501 (14.4%)	86 (12.9%)	587 (14.1%)
0	2,988 (85.6%)	583 (87.1%)	3,571 (85.9%)
Socioeconomic status, n(%)				0.11
Low SES	312 (9.7%)	51 (8.3%)	363 (9.5%)
Medium SES	2,319 (72.0%)	429 (70.0%)	2,748 (71.7%)
High SES	591 (18.3%)	133 (21.7%)	724 (18.9%)
Tumor Histology, n(%)				<0.001
Synovial	2,126 (60.9%)	274 (41.0%)	2,400 (57.7%)
Clear cell	176 (5.0%)	130 (19.4%)	306 (7.4%)
Angiosarcoma	479 (13.7%)	50 (7.5%)	529 (12.7%)
Rhabdomyosarcoma	339 (9.7%)	48 (7.2%)	387 (9.3%)
Epithelioid	369 (10.6%)	167 (25.0%)	536 (12.9%)
Primary Site, n(%)				<0.001
Upper limb and shoulder	1,108 (31.8%)	260 (38.9%)	1,368 (32.9%)
Lower limb and hip	2,381 (68.2%)	409 (61.1%)	2,790 (67.1%)
Sarcoma T category, n(%)				0.013
<5cm	1,412 (40.5%)	307 (45.9%)	1,719 (41.3%)
>5cm, <=10cm	959 (27.5%)	152 (22.7%)	1,111 (26.7%)
>10cm, <=15cm	316 (9.1%)	46 (6.9%)	362 (8.7%)
>15cm	209 (6.0%)	40 (6.0%)	249 (6.0%)
Unknown	593 (17.0%)	124 (18.5%)	717 (17.2%)
Treatment started, days from diagnosis, median (IQR)	10 (0–31)	11 (0–30)	10 (0–31)	0.68
Definitive surgical procedure, days from diagnosis, median (IQR)	41 (8–99))	48 (22–97)	42 (12–99)	0.003
Surgical Procedure, n(%)				<0.001
Local excision/resection	1,190 (34.1%)	164 (24.5%)	1,354 (32.6%)
Radical excision/resection - limb salvage	1,808 (51.8%)	318 (47.5%)	2,126 (51.1%)
Limb amputation	471 (13.5%)	169 (25.3%)	640 (15.4%)
Major intraarticular amputation	20 (0.6%)	18 (2.7%)	38 (0.9%)
External Beam Radiation, n(%)	1,821 (53.9%)	297 (45.6%)	2,118 (52.6%)	<0.001
Chemotherapy, n(%)				0.22
No	2,285 (65.5%)	458 (68.5%)	2,743 (66.0%)
Yes	1,114 (31.9%)	199 (29.7%)	1,313 (31.6%)
Missing	90 (2.6%)	12 (1.8%)	102 (2.5%)
90-day mortality, n(%)	65 (1.9%)	<10	71 (1.7%)	0.073
Follow up (months), median (IQR)	53.9 (27.2–93.0)	51.2 (26.7–87.7)	53.3 (27.1–92.4)	0.36

RLNS: Regional lymph node surgery; CP: Cancer program; XRT: external beam radiation therapy

Receipt of Regional Lymph Node Surgery

In the study cohort, 669 patients received regional lymph node surgery (RLNS). To assess how patients were selected for RLNS we compared those 669 patients to the 3,489 patients that did not undergo regional lymph node surgery (Table 1). On multivariate logistic regression, patients with epithelioid (OR 3.9, CI 3.0–5.0, p<0.001) and clear cell (OR 6.4, CI 4.7–8.5, p<0.001) were more likely to undergo RLNS compared to reference synovial sarcoma (Figure 1A). Young patients (<18 years) were significantly more likely to undergo RLNS in an age dependent fashion, with patients over 65 years least likely to have RLNS (OR 0.21, CI 0.2–0.4, p<0.001). Zip code-linked socioeconomic and insurance status were not associated with RLNS. Tumor size was not significantly associated with receipt of RLNS, but there was a trend towards increased RLNS with larger tumors. Patients with poorly differentiated/undifferentiated tumors had similar likelihood of RLNS than patients with non-poorly differentiated tumors. Adjusted probability of receiving RLNS by histology was highest for clear cell (41%) and lowest for synovial (10%), figure 1B.

Figure 1:

Model results for logistic regression of A) factors associated with receipt of regional lymph node surgery and B) adjusted probabilities of lymph node surgery by sarcoma histology.

Factors Predictive of Positive Lymph node

Among the 631 patients with complete information about nodal status, clinically node negative disease, and who underwent RLNS, 45 (7.1%) had a pathologically positive lymph node. We examined factors associated with pathologic nodal positivity in a multivariate model (Figure 2A). On multivariate logistic analysis, there was no association of age with nodal positivity, despite age being significantly associated with patient selection for RLNS. Tumor differentiation was not associated with nodal positivity, but size was associated with nodal positivity with most significant effect in tumors greater than 15 cm (OR 21.2, CI 5.32–84.2, p<0.001). Sex and tumor site were not significantly associated with nodal positivity. Adjusted probability of a positive node was highest for clear cell, 21.8% (95% CI 13.2 –30.3%), followed by epithelioid (9.8%, 95% CI 4.5–15.2), and angiosarcoma (9.4%, 95% CI 1.92–17.0%) (Figure 2B). Rhabdomyosarcoma (4.9%) and synovial sarcoma (2.1%) had low adjusted rates of nodal positivity.

Figure 2:

Model results for logistic regression of factors associated with nodal positivity among those who A) underwent regional lymph node evaluation and B) adjusted probabilities of regional lymph node disease by sarcoma histology.

Survival

For the entire cohort of patients, 5-year overall survival was 70.6% (95% CI 69.1–72.1) (Figure 3). Patients with negative RLN had a 5-year survival rate of 76.3% (95%CI: 72.2, 79.9) compared to 70.3% (95%CI: 68.6, 71.9) in patients without nodal staging and 37.3% (95%CI: 22.9, 51.7) in patients with a positive RLN. On multivariable analysis, positive RLN was significantly associated with increased mortality (HR 2.45, CI 1.64–3.64, p<0.001), table 2. Histology was also significantly associated with survival. Other factors associated with 5-year overall survival were age, gender, presence of comorbidities, and poor/undifferentiated histology.

Figure 3:

Kaplan-Meier survival curves evaluating overall survival for patients with negative regional lymph nodes, positive regional lymph nodes, and no regional node evaluation by sarcoma histology. Abbreviations: RLNS: regional lymph node surgery; SCARE: Synovial, Clear cell, Angiosarcoma, Rhabdomyosarcoma, Epithelioid.

Table 2:

Cox proportional hazards multivariate model examining factors associated with overall survival in soft tissue sarcoma SCARE histologies

Characteristic*	HR (95% CI)	P-Value
Histology
Synovial	Ref
Clear cell	1.98 (1.56,2.52)	<0.001
Angiosarcoma	2.23 (1.90,2.63)	<0.001
Rhabdomyosarcoma	1.46 (1.22,1.76)	<0.001
Epithelioid	1.27 (1.04,1.55)	0.018
Regional Lymph Node Status
No RLNS	Ref
Negative RLN	0.84 (0.7,1.02)	0.072
Positive RLN	2.45 (1.64,3.64)	<0.001
Age
<18 years	Ref
18 – 39 years	1.81 (1.34,2.45)	<0.001
40 – 64 years	2.60 (1.93,3.48)	<0.001
>=65 years	5.63 (4.12,7.69)	<0.001
Male vs Female	1.27 (1.13,1.42)	<0.001
White vs Nonwhite	0.93 (0.81,1.08)	0.35
0 vs >0 Comorbidities	0.79 (0.68,0.91)	0.001
Poorly/Undifferentiated
No	Ref	<0.001
Yes	1.97 (1.55,2.51)	<0.001
Unknown	1.67 (1.33,2.10)	<0.001

RLNS: Regional lymph node surgery; RLN: Regional lymph node;

^*Tumor size, receipt of chemotherapy, and receipt of radiation therapy violated the proportional hazards assumption and these results are stratified on those variables

Impact of nodal status on histology specific survival

We analyzed survival by histology and nodal status to determine the histology specific prognostication of pathologic nodal status. There was significant stratification of survival in all histologies except for rhabdomyosarcoma. The absolute difference in 5-year survival between node positive and node negative patients was most pronounced in patients with clear cell sarcoma, 82.5% for node negative patients and 20% for node positive, figure 3. Patients with angiosarcoma similarly had a large stratification of survival based on nodal status, 59.1% for node negative patients and 16.7% for node positive. This effect was smaller in epithelioid (88.4% vs 74.1%) and synovial (70.1% vs 42.9%).

DISCUSSION

The role of pathologic lymph node assessment in soft tissue sarcoma has not been clearly elucidated. In this manuscript we report factors associated with receipt of regional lymph node surgery, factors associated with nodal positivity, and describe how lymph node status stratifies patients for histology specific survival. Using this data, we establish a group of patients that is at sufficiently high risk of nodal metastasis and in whom that information provides meaningful prognostication, to warrant consideration for pathologic nodal evaluation in clinically node negative patients.

We used a designation of “regional lymph node surgery” rather than sentinel lymph node biopsy because that distinction is not present in the NCDB data until 2012. A sentinel lymph node biopsy (SLNB) may provide a less invasive way to detect regional micrometastatic disease in high risk STS as in other cancers such as melanoma or breast cancer.^21,22 However, the performance characteristics of sentinel lymph node biopsy for sarcoma are not well described. One previous study has shown that performance of sentinel lymph node biopsy did not accurately identify patients at risk for regional failure.¹⁵ The adoption of regional lymph node evaluation for STS relies on the accuracy of SLNB since the prognostic value is more difficult to justify against the morbidity of complete lymph node dissection. Further study is needed to evaluate the accuracy of SLNB to identify occult nodal metastasis in STS.

We identify discordance between factors associated with the receipt of pathologic nodal staging and factors associated with nodal positivity. In fact, the decision for nodal staging appears to be primarily influenced by age, whereas the probability of a positive node was not associated with age. Nodal positivity is strongly associated with tumor size, whereas there was a weak relationship between size and selection for nodal evaluation. This, perhaps, represents selection of patients with longer life expectancy in whom prognostication may be perceived as more meaningful, or of patients more likely to tolerate or be recommended for chemotherapy. This also suggests that selection of patients for pathologic nodal staging can undergo evidence-based optimization. An understanding of both the probability of a positive node, as well as the value of that information to determine risk of death, must be accounted for when counseling patients regarding nodal evaluation.

There exists mounting evidence in other cancers, most notably melanoma and breast cancer, that nodal positivity is a marker of probability of recurrence and that aggressive attempts at surgical eradication of nodal disease are not associated with improved survival.^22–26 This has led to an increasing selectivity of clinically node negative patients for anatomic lymph node dissections, but has also influenced consideration for sentinel lymph node biopsy.^27,28 Herein, we present that identification of regional lymph node disease in patients with soft tissue sarcoma of the extremities can similarly stratify patients at risk of mortality. However, this effect is not uniform by histology.

The NCCN guidelines for melanoma²⁹ recommend that sentinel lymph node biopsy be considered for an expected nodal positivity rate of 5–10%, and that SLNB be omitted if the risk of a positive node is less than 5%. Using these thresholds and an absolute difference in overall survival at 5 years of 25% between node positive and node negative patients, we identify patients that should be considered for pathologic nodal evaluation. Clear cell and angiosarcoma meet these criteria for consideration for pathologic nodal evaluation. Ideally the prognostication obtained from an additional procedure such as lymph node surgery can be used to stratify patients for adjuvant therapy. To date, chemotherapy regimens have shown little efficacy in STS,^30–33 although this is an area of ongoing investigation. The recently presented phase II SWOG S1609 demonstrated an objective response rate of 25% in patients with metastatic or unresectable angiosarcoma who received combination ipilimumab and nivolimumab.³⁴ Whether these therapies are effective in the adjuvant setting are not known. However, the use of pathologic nodal evaluation for patients with clear cell and angiosarcoma identifies patients at extremely high risk of death at five years and represent an ideal high-risk cohort to investigate improved strategies of adjuvant therapy.

This study has several limitations. Local and regional recurrence are not captured in the NCDB and we are unable to assess the effect of regional lymph node surgery on those outcomes. This study only applies to patients without preoperative evidence of nodal disease. Facility information is also not provided for patients under 40, limiting the ability to evaluate the association of practice setting or patient volume with RLNS. Patients with clinical evidence of nodal disease are at high risk for distant recurrence and for whom resection may be required for management of symptoms and/or control of regional disease. This study uses observational data and there are many factors associated with the selection of patients for nodal surgery that may not appear in the data. For example, there is a relatively high rate of lack of residual tumor or data about residual tumor at time of resection, leading to a rate of 17% of unknown tumor size. We elected to include those patients in the analysis because our primary focus was on the histology at biopsy affecting surgical decision-making. Similarly, with regard to clinical staging, we rely on coding of clinical nodal status which may not be a true representation of clinically node positive/negative disease. The presence of those patients could bias results towards increased rates of nodal positivity and worse outcome.

CONCLUSION

In this study we describe the receipt of surgical lymph node staging, factors associated with nodal positivity, and the effect of nodal positivity on survival by histologic subtype in soft tissue sarcoma. Discordance exists between selection of patients for pathologic nodal evaluation and factors associated with nodal positivity. Utilizing a threshold of at least 5–10% positivity and significant stratification of overall survival between pathologically node negative and node positive patients, clinically node negative patients with clear cell and angiosarcoma should be considered for pathologic nodal evaluation. Node positive clear cell and angiosarcoma patients represent a high-risk group ideal to investigate and improve adjuvant systemic therapy strategies.

Supplementary Material

Supplementary Figure 1

Supplemental Figure 1: CONSORT diagram of patients with extremity sarcoma from 2004–2017 in the National Cancer Database included in the study.

Synopsis:

Using the National Cancer Database, we evaluate nodal positivity rates and stratification of survival by regional lymph node status in patients with clinically node negative soft tissue sarcoma of the extremities undergoing regional nodal evaluation. Nodal status was most predictive of survival in clear cell sarcoma and angiosarcoma and pathologic nodal evaluation should be investigated to predict survival.

Data Availability Statement:

Due to the data use agreement with the American College of Surgeons, the data are not shared by the authors.