Annals of Surgical Oncology Sign the Guestbook
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Howe, J. R.
Right arrow Articles by Scott-Conner, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Howe, J. R.
Right arrow Articles by Scott-Conner, C.
Related Collections
Right arrow Other Malignancies
Annals of Surgical Oncology 8:496-508 (2001)
© 2001 Society of Surgical Oncology

ORIGINAL ARTICLES

Small Bowel Sarcoma: Analysis of Survival From the National Cancer Data Base

James R. Howe, MD, Lucy H. Karnell, PhD and Carol Scott-Conner, MD, PhD

From the Departments of Surgery (JRH), Otolaryngology (LHK), and Surgery (CS-C), University of Iowa College of Medicine, Iowa City, Iowa.

Correspondence: Address correspondence and reprint requests to: James R. Howe, MD, Department of Surgery, University of Iowa College of Medicine, 200 Hawkins Drive, Iowa City, IA 52242; Fax: 319-356-9378; E-mail: james-howe{at}uiowa.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Background: Small bowel sarcomas (SBS) are rare, accounting for 10% of small bowel cancers. As a result, few studies of SBS have had enough patients to accurately define their natural history and to determine the factors that have an impact on patient survival. The objective of this study was to examine patient and tumor factors in SBS and to determine prognostic factors for disease-specific survival (DSS) using the National Cancer Data Base.

Methods: Data from the National Cancer Data Base for patients diagnosed with primary SBS between 1985 and 1995 were analyzed. The {chi}2 statistic was used to determine significant differences between groups of patient, tumor, and treatment factors. DSS was calculated for patients diagnosed between 1985 and 1990. Significant differences in survival were determined using the Wilcoxon statistic for univariate analyses and by Cox regression in multivariate analyses.

Results: Of 14,253 small bowel tumors diagnosed between 1985 and 1995, sarcomas represented 10.1%. Overall, 5-year DSS was 38.9%, with a median survival of 34.1 months (n = 590). By univariate analysis, patient age, sex, tumor size, tumor grade, histologic type, general summary stage, nodal status, and whether cancer-directed surgery was performed were significantly correlated with DSS. In multivariate analysis, tumor size <5 cm, leiomyosarcoma histology, and localized disease were found to be significant favorable prognostic factors for DSS.

Conclusions: SBS are rare tumors that are challenging in terms of their histopathologic classification, grading, and staging. Patients with SBS were treated predominantly by surgery, with a minority receiving adjuvant therapy. Tumor size, histologic type, and general summary stage were independent prognostic factors for 5-year DSS in patients with SBS, which is improved relative to 5-year DSS seen in patients with small bowel adenocarcinoma.

Key Words: Small bowel • Gastrointestinal sarcoma • National Cancer Data Base • Disease-specific survival


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
The small intestine is the longest portion of the gastrointestinal (GI) tract, yet is the site of origin for only 2% to 3% of GI tumors.1 In a study of 14,253 small bowel tumors in the National Cancer Data Base (NCDB) between 1985 and 1995, 35% were adenocarcinomas, 28% were carcinoids, 21% were lymphomas, and 10% were sarcomas.2 In the year 2000, there were an estimated 4,700 new cases and 1,200 deaths from small bowel cancer in the United States,3 of which 10% to 13% would have been primary small bowel sarcomas (SBS).1,2,4,5 The Surveillance, Epidemiology, and End Result (SEER) program has estimated the annual incidence of SBS to be 1.2 cases per million persons.4

Sarcomas of the GI tract account for approximately 2% of all sarcomas in adult patients6; in a review of 725 malignant GI sarcomas, 5% were found to arise in the esophagus, 47% in the stomach, 35% in the small bowel, and 12% in the colorectum.7 It was previously believed that the majority of GI mesenchymal tumors were derived from smooth muscle, but the recent routine use of immunohistochemical and electron microscopic techniques have revealed a more complex picture.8 Current nomenclature refers to these tumors as gastrointestinal stromal tumors (GIST), gastrointestinal autonomic nerve tumors (GANT), mixed smooth muscle-neural tumors, or GIST-NOS (not otherwise specified) tumors.9 Although further pathologic evaluation of these tumors may more accurately reflect their cell of origin, it is unclear whether these changes in classification will be of prognostic significance.

Because SBS are rare tumors, there have been few studies with enough patients to define their natural history or to determine factors that have an impact on patient survival. Most larger series have been reports of sarcomas from all GI sites, some of which have demonstrated diminished survival in patients with sarcomas of small bowel origin,6,10,11 but few have contained details specifically related to SBS. In this study, the large numbers of patients available in the NCDB was used to determine which patient, tumor, and treatment factors influence the survival of patients with SBS. Having recently performed a similar review in adenocarcinomas of the small bowel, this report represents the second review of the four most common histologic types of small bowel tumors and, to our knowledge, is the largest series of SBS to date.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
National Cancer Data Base
The NCDB collects data from a broad spectrum of hospitals in the United States on a voluntary basis and is a joint project of the American College of Surgeons Commission on Cancer and the American Cancer Society. These data are provided from cancer registries in a standardized format using coding schema from the Data Acquisition Manual,12 the American Joint Commission on Cancer (AJCC) Manual for Staging of Cancer,13 and the World Health Organization’s International Classification of Diseases for Oncology (ICD-O-2) system for coding site and histologic type.14 An algorithm based on patient and disease characteristics (including patient sex, site, date of birth, and zip code) was used to identify and remove duplicate records to ensure that patients seen at multiple institutions for the same cancer were not included within the database more than once.

SBS were extracted from the NCDB on the basis of anatomic site (ICD-O-2 codes C17.0–17.9) and histologic type (ICD-O codes 8680, 8800–8804, 8810, 8814, 8830, 8832, 8840, 8850–8858, 8890, 8891, 8894–8896, 8897, 8900–8902, 8910, 8920, 8990, 9050, 9120, 9140, 9150, 9180, 9260, 9500, 9540, and 9560) for those patients first diagnosed and/or treated between January 1985 and December 1995. Small bowel subsites included the duodenum, jejunum, ileum, and Meckel’s diverticulum, as well as overlapping lesions of the small intestine (those involving two subsites) and NOS lesions of the small intestine. Periampullary tumors and those of the ampulla of Vater (ICD-O code C24.0) were not included. The method of data collection by the NCDB involves an annual submission of data from cancer registries for the most current year (for which complete data are available) and for the year 5 years previous to this.15 Therefore, up to 10 years of follow-up data were available for the patients diagnosed in 1985, up to 5 years available for those diagnosed between 1986 and 1990, and 1 year for those diagnosed between 1991 and 1995.

Individual states were grouped into six geographic regions: the Northeast, Southeast, Midwest, South, Mountain, and Pacific. Race/ethnicity was grouped into White non-Hispanic, African American non-Hispanic, Hispanic (of any racial origin), and other (e.g., Asian, Pacific Islander, Native American). Income reflected the average family income of the patient’s residential zip code. Income was grouped according to approximations of the top and bottom 10% of all NCDB cancer cases, with the low-income group being less than $20,000 (actually representing 11.2%) and the high-income group being $47,000 or more (actually representing 10.3%).

Because there is no AJCC staging system for sarcomas, general summary stage was used. This system classifies patients as having local, regional (extension into adjacent tissues or nodal involvement), or distant disease. The World Health Organization’s standard grading system was used with four separate categories (well, moderately well, poorly differentiated, and undifferentiated). With this grading system, if a tumor is assigned two different degrees of differentiation, the lesser differentiation is recorded.

Treatment represents the first planned course of cancer-directed therapy used to manage the primary cancer and excludes treatments initiated for recurrence or relapse of disease. Patients were classified as having had no surgery, non–cancer-directed surgery (including biopsy, exploration, bypass, and unknown/NOS non–cancer-directed procedures), or cancer-directed surgery (radical or simple excision with or without nodal dissection, fulguration, cryosurgery, laser ablation, surgery at regional or distant sites only, and unknown/NOS cancer-directed procedures). The data in the NCDB did not include whether surgery was performed with curative or palliative intent, nor was there information regarding the completeness of resection. Specific details regarding the amount of radiation received or the agents and dosages used for chemotherapeutic regimens were also not available in the NCDB.

Statistical Analysis
All analyses were performed using SPSS software version 9.0 (Chicago, IL). When appropriate, variables were further stratified by disease characteristics that influence treatment and outcome. The {chi}2 test was used to determine statistically significant differences between variables. Using the actuarial life table method, annual survival rates represented disease-specific rates, with the end point being death with cancer. Pairwise comparisons were performed using the Wilcoxon statistic to determine whether selected stratifications had significantly different survival rates. Variables with significantly different survival rates as demonstrated in univariate analysis (P <= .05) were further analyzed by Cox regression analysis to identify independent influences on survival.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Patient Characteristics
The mean age of patients with SBS was 59.3 years (median, 61.0 years), with a range of 8 to 94 years (Table 1). There was a slight male predominance (53.7%), and a higher than expected incidence of White non-Hispanics (85.8%) than in the U.S. population (75.6% in the June 1990 census [http://www.census.gov]). There was also a lower than expected incidence in Hispanic (3.0%, vs. 9% of the population in the1990 census) and African-American patients (7%, vs. 12% of the population in the 1990 census).


View this table:
[in this window]
[in a new window]
  		TABLE 1. Patient characteristics for small bowel sarcomas, 1985–1995
 
Tumor Characteristics
The distribution of various tumor characteristics is listed in Table 2. Tumors in Meckel’s diverticula, overlapping sites, or NOS sites accounted for 44.4% of SBS. When these subsites were removed, nearly half of SBS were classified as jejunal (46.8%), with 27.7% in the duodenum and 25.5% in the ileum. There was a similar percentage of tumors <5 cm in size (29.9%) as there was of tumors >10 cm in size (29.3%). Tumor grade was only reported in 47.5% of SBS, and nodal status data was only available in 48.3%. General summary stage information was not available in 11.7% of cases. Regarding general summary stage, 42.1% of tumors were localized at presentation, whereas 30.5% exhibited regional spread and 27.4% showed distant disease. The most common histologic subtypes were leiomyosarcoma (75.1%), followed by epithelioid leiomyosarcoma (6.9%), Kaposi’s sarcoma (4.1%), NOS sarcoma (4.0%), and spindle cell sarcoma (2.8%).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Tumor characteristics for small bowel sarcomas, 1985–1995
 
Treatment
Surgery was the primary treatment for most patients with SBS (Table 3). Cancer-directed surgery was performed in 87.9% of patients, non–cancer-directed surgery in 6.5%, and no surgery was performed in 5.6%. Non–cancer-directed surgery predominantly included procedures involving exploration, biopsy, or bypass (either alone or in combination). Surgical codes in the Data Acquisition Manual12 have no code for complete or incomplete resection, which has been shown to be of prognostic value in some studies of GI sarcoma.6,11 Radiotherapy was performed in a total of 6.0% of patients, and chemotherapy in 16.1%. Radiotherapy and/or chemotherapy were used as an adjuvant to surgery in 15.5% of patients, and 3.8% of patients received no cancer-directed treatment.


View this table:
[in this window]
[in a new window]
  		TABLE 3. Cancer-directed treatment received for small bowel sarcomas, 1985–1995
 
Survival
The overall 5-year disease-specific survival (DSS) of patients with SBS was 38.9% (median, 34.1 months) (Table 4), with a median time to death or follow-up of 22.5 months (mean, 32.5 months; n = 590). The median follow-up of the patients alive at last contact was 63.0 months (mean, 50.0 months; n = 228), and 17.0 months for those who died (mean, 21.5 months; n = 362). Of the 590 patients with SBS diagnosed between 1985 and 1990 used in these survival analyses, 44 (7.5%) were lost to follow-up at 1 year, 85 at 2 years (14.4%), 93 at 3 years (15.8%), and 131 (22.2%) at 5 years. In the univariate analysis, advancing age was a significant risk factor for cancer-related death (P = .0015), with the survival of patients aged >75 years being significantly less than of those aged <60 years (P = .0004) and 60 to 75 years (P = .0042). There was no significant difference in DSS between patients aged <60 years and 60 to 75 years (P = .3117). The 5-year DSS of male patients was significantly less than that of female patients (P = .0242), but there were no survival differences based on race/ethnicity or income.


View this table:
[in this window]
[in a new window]
  		TABLE 4. Five-year disease-specific survival for small bowel sarcomas, 1985–1990: Overall and by select patient/tumor factors
 
The anatomic site of the sarcoma was not correlated with survival, but only 337 patients (57.1%) had the specific site recorded as duodenal, jejunal, or ileal, as compared with other/NOS sites. In subset analysis, the survival of patients with duodenal SBS was lower than that of patients with jejunal tumors (P = .0462), but there were no significant differences between duodenal and ileal sites or jejunal and ileal sites. Tumor size was also a significant prognostic factor (P < .0001) (Fig. 1), with significant differences in 5-year DSS between patients with tumors <5 and 5–10 cm (P = .0050) and tumors <5 and >10 cm (P < .0001), but not for tumors 5 to 10 cm vs. tumors >10 cm (P = .0574). The rate of distant metastases increased with increasing tumor size (P = .0050), from 15.6% in patients with tumors <5 cm in size, to 19.1% for those with tumors 5 to 10 cm, and 28.4% for patients with tumors >10 cm.



View larger version (25K):
[in this window]
[in a new window]
  		FIG. 1. Disease-specific survival by tumor size for sarcoma of the small bowel, 1985–1990 (P < .0001; 95% confidence intervals are provided for the 5-year survival rate).

 
General summary stage was a significant predictor of DSS (Fig. 2), with significant differences between localized and regional disease (P < .0001), regional and distant disease (P < .0001), and local and distant disease (P < .0001). Patients with negative nodes had reduced 5-year DSS compared with those with involved nodes (P = .0028), but data on nodal status was only available from 48.3% of patients. Patients with well- and moderately well-differentiated tumors had higher 5-year DSS compared with those with poorly or undifferentiated tumors (P < .0001; Fig. 3), but the number of cases for analysis was limited to only 41.9% of patients. Patients having cancer-directed surgery had significantly higher 5-year DSS compared with those not having cancer-directed surgery (P < .0001) (Fig. 4). There were also significant differences in DSS regarding the three most common SBS histologic subtypes (P < .0001), with 5-year DSS being highest in patients with leiomyosarcoma (40%), followed by patients with epithelioid leiomyosarcoma (21%). There were no 5-year survivors with Kaposi’s sarcoma. The differences in survival were significant between all three groups (P = .0007, leiomyosarcoma vs. epithelioid; P = .0003, leiomyosarcoma vs. Kaposi’s; P = .0142, epithelioid vs. Kaposi’s).



View larger version (26K):
[in this window]
[in a new window]
  		FIG. 2. Disease-specific survival by general summary stage for sarcoma of the small bowel, 1985–1990 (P < .0001; 95% confidence intervals are provided for the 5-year survival rate).

 


View larger version (22K):
[in this window]
[in a new window]
  		FIG. 3. Disease-specific survival by grade for sarcoma of the small bowel, 1985–1990 (P < .0001; 95% confidence intervals are provided for the 5-year survival rate).

 


View larger version (24K):
[in this window]
[in a new window]
  		FIG. 4. Disease-specific survival by type of surgery for sarcoma of the small bowel, 1985–1990 (P < .0001; 95% confidence intervals are provided for the 5-year survival rate).

 
In univariate analysis, those factors found to be significantly correlated with DSS were age, sex, size, histology, general summary stage, nodal status, grade, and surgery. Multivariate analysis was then performed using these factors, but because of low numbers of observations for grade (247) and nodal status (293), these two factors were not included in the multivariate analysis calculations (Table 5). When grade and nodal status were included, the total number of evaluable cases dropped to 89, and neither variable was found to be a significant independent prognostic factor. When grade and nodal status were removed, the number of evaluable cases for multivariate analysis increased to 308. By this analysis, age, sex, and cancer-directed surgery were not found to be independent risk factors for DSS, whereas size, histology, and general summary stage remained independent predictors of survival. The relative risk of death from epithelioid leiomyosarcoma and Kaposi’s sarcoma was 2.1 and 77.9 times higher, respectively, than the risk of death from leiomyosarcoma. Compared with tumors < 5 cm in size, the risk of dying was 1.5 times higher in patients with tumors 5–10 cm, and 2.4 times higher for patients with tumors >10 cm. Patients with regional general summary stage were 3.4 times more likely to die than patients with localized disease, which increased to a 9.2-fold higher risk in patients with distant metastases.


View this table:
[in this window]
[in a new window]
  		TABLE 5. Multivariate analysis of independent prognostic factors for disease-specific survival for small bowel sarcomas, 1985–1990 (n = 308)
 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
Tumors of the small bowel are rare compared with other regions of the GI tract, despite the fact that the small bowel accounts for 60% of GI tract length. This finding has been ascribed to a variety of factors, including higher levels of detoxification enzymes, improved immunity to infectious agents, rapid transit time of potentially injurious substances, and increased cell turnover rate.16–18 Data from the NCDB has demonstrated that adenocarcinomas are the most common malignant tumors of the small bowel, followed by carcinoid tumors, lymphomas, and sarcomas.2

Although the NCDB is not population based and is therefore subject to ascertainment bias, a large number of hospitals across the United States (525 to 1324 hospitals per year between 1985 to 1995) provide data. The principal limitations of cancer registry data are that they are collected retrospectively, do not contain specific details on presenting symptoms or the specifics of the treatments received, and are entered by cancer registrars. Also, unlike single-institution studies that pay attention to detail and can perform pathologic review when needed, the accuracy of staging and pathologic diagnosis within a national registry may vary widely across the institutions that submit data. The advantages of the NCDB data are large numbers, representation from a broad spectrum of hospitals, and reduced selection bias for accrual of cases.

Having large numbers is of particular importance for analysis of rare tumors such as SBS, where it is nearly impossible for a single institution to collect enough cases to make meaningful predictions regarding important prognostic factors. Therefore, most previous series have examined sarcomas of the GI tract in general, which is problematic in that the survival for patients with tumors derived from various GI sites may be very different. As an example, Ng et al.11 found that the median survival of patients with resected large bowel sarcomas was 39 months, which was reduced to 25 months for stomach sarcomas and 18 months for SBS. Since several other studies suggest decreased survival for SBS,6,10,19,20 we focused here on trying to identify factors of biologic relevance specific to SBS.

Patient and Tumor Factors
In this study, the mean age of the 1441 patients with SBS was slightly higher than that seen for 252 patients with benign and malignant small bowel smooth muscle tumors (57 years) collected by the Armed Forces Institute of Pathology (AFIP),20 200 patients with GI sarcomas at Memorial Sloan Kettering Cancer Center (MSKCC) (mean age, 58 years),21 191 patients with GI sarcomas from M. D. Anderson Cancer Center (MDACC) (mean age, 54 years),11 and 28 patients with small bowel leiomyosarcoma from the Cleveland Clinic (mean age, 53 years).22 The slight male predominance noted in this series was higher than that seen in 232 patients reported from the SEER database on small bowel tumors (51% male),4 but less than that from MDACC (58%),11 the Cleveland Clinic (61%),22 and MSKCC series for all sites of GI sarcomas (56%).21 Skandalakis and Gray7 found an equal incidence of these tumors between the sexes. Comparing the incidence of SBS to small bowel adenocarcinoma by race during this same time period in the NCDB,22 the percentage of cases in White non-Hispanics was 81% and 86% (White non-Hispanics made up 76% of the U.S. population in the 1990 census), in African Americans 14% and 7% (12% in the census), and in Hispanics 3% and 3% (9% in the census) for small bowel adenocarcinoma and SBS, respectively. Therefore, there was an even higher increase in the percentage of cases in White non-Hispanics, a similar percentage in Hispanics, and a 50% reduction in African Americans for cases of SBS relative to small bowel adenocarcinoma. In the SEER study of SBS, African Americans were found to have a decreased incidence of SBS (0.9 cases per million annually) compared with Whites (1.3 cases per million annually).4

The small bowel is the second most common site of GI sarcoma. Skandalakis and Gray’s review7 of 725 leiomyosarcomas of the GI tract found that 48% arose in the stomach, 33% in the small bowel, 12% in the colorectum, and 5% in the esophagus. The AFIP study20 of 1004 benign and malignant GI smooth muscle tumors found that 52% arose within the stomach, 25% in the small bowel, 11% in the colorectum, 5% in the esophagus, and 7% in other sites (omentum, peritoneum, and mesentery). Pooling the results of six other recent series encompassing 628 GI sarcomas, 40% arose in the stomach, 36% from the small bowel, 15% from the colorectum, and 8% from other/NOS sites.10,11,21,23–25 Within the small bowel, we found the distribution of sarcomas in the duodenum (28%), jejunum (47%), and ileum (25%) to be similar to that in the SEER study in terms of jejunal tumors (48%), but the NCDB series had a 10% higher incidence of duodenal and 10% lower incidence of ileal tumors.4 In their literature review including 213 small bowel leiomyosarcomas, Skandalakis and Gray7 found 31% to arise in the duodenum, 37% in the jejunum, and 32% in the ileum. The AFIP study20 found 31% to originate within the duodenum, 46% in the jejunum, and 23% in the ileum, which is very similar to the results of the current study.

Pathologic Classification of SBS
The pathologic analysis of SBS presents several problems. Because they are spindle cells and arise in the muscularis propria, they were historically considered to be of smooth muscle origin26 and were diagnosed as leiomyomas (benign) and leiomyosarcomas (malignant). However, when immunoperoxidase markers specific for smooth muscle origin were found to be negative in a great number of cases, doubt was cast on their origin from smooth muscle. Recent evidence suggests that most such sarcomas originate from the interstitial cells of Cajal (the pacemaker cells in the lamina propria), on the basis of positive staining for the c-kit proto-oncogene protein (CD117).27,28 Most pathologists now refer to SBS as either GIST (previously referred to as leiomyosarcoma) or GANT. GIST are much more common than GANT, and the latter are clearly of neural origin. Some have suggested subclassification of GI mesenchymal tumors into smooth muscle neoplasms (spindle, epithelioid, and mixed leiomyosarcomas), GANT, mixed leiomyosarcoma/neural tumors, and GIST-NOS.9 The data in the older literature describing GI leiomyomas and leiomyosarcomas are still relevant in light of this new classification, for the majority of these tumors are GIST, and it is not clear whether CD117 positivity has an influence on prognosis.29 However, it has been shown that STI 571, a compound used to treat chronic myelogenous leukemia, can inhibit the tyrosine kinase activity of c-kit and seems to have activity in suppressing the proliferation of c-kit positive tumors.30,31 This agent is currently being investigated in GI stromal tumors and represents new hope for the treatment of these chemotherapy-resistant tumors.

The literature is further confused by grouping stromal tumors from all GI sites, despite evidence that tumors arising in the esophagus differ significantly in their behavior from those arising in the stomach, which differ from those arising in the small bowel. The collection of this NCDB data predates the widespread use of this new terminology, and it remains to be seen whether reclassification of SBS into GIST and GANT will provide improved prognostic information in the future. There were 30 different ICD-O-2 codes found in this series, pointing out the complexity of classifying individual tumors, which may vary from pathologist to pathologist. In this study, the general categories of leiomyosarcoma, epithelioid leiomyosarcoma, and Kaposi’s sarcomas accounted for 86% of all cases, with the less common histologic subtypes representing the minority of cases.

Two other areas that present difficult diagnostic challenges for pathologists are distinguishing benign from malignant mesenchymal tumors and the grading of these tumors. Invasion into surrounding structures or metastases are unequivocal features of malignancy, but these are only evident in a subset of these tumors. For this reason, some have suggested that these tumors be described as having either low or high malignant potential.32 Characteristics such as tumor size, cellularity, mitotic rate, and necrosis dominate the criteria used for grading and malignant potential,8,19,23,25,33 but there is incomplete agreement on specific details, such as the critical size or mitotic rate. Counting mitoses is notoriously variable and depends on the size of the microscopic field, completeness of fixation, and the subjective interpretation of the pathologist. The inevitable result of this lack of strict criteria was reflected in this study with data collected from over 1000 hospitals, where the majority of tumors were not graded between 1985 and 1995. Because tumor grade has been shown to be an important prognostic factor in GI mesenchymal tumors,11,23,25 it would be helpful for future studies if a consensus panel of pathologists could agree on acceptable criteria for the grading of these tumors.

Staging of SBS
There is no AJCC staging system for gastrointestinal sarcomas, as "sarcomas arising in parenchymatous organs and from hollow viscera" are excluded in this classification.13,34 Ng et al.11 have proposed a TGM staging system for GI sarcoma. They defined T1 tumors as being localized and <5 cm in size, T2 as localized and >5 cm, T3 as invading contiguous organs or with peritoneal implants, and T4 as tumor rupture; G1 refers to low-grade tumors, G2 to high-grade tumors; M0 to no metastases, and M1 to the presence of metastases. Patients with stage I lesions (T1G1M0) had a 75% 5-year overall survival (n = 10), stage II (T2G1M0) a 52% 5-year overall survival (n = 12), stage III (T1–2G2M0 or T3GanyM0) a 28% 5-year overall survival (n = 47), stage IVA (M1 tumors or residual disease) a 12% 5-year overall survival (n = 46), and stage IVB (any T4 tumor) a 7% 5-year overall survival (n = 24).11 The NCDB data does not have fields for peritoneal implants or for tumor rupture, so this system was not applicable. The current study used the general summary stage classification, which groups patients into those with localized tumors, those with regional extension, or those with distant metastases. Although these two staging systems are not directly comparable and the Ng study contained GI sarcomas from all sites, the results were similar. We found a 72% 5-year DSS with localized SBS compared with their 75% and 52% 5-year overall survival for stage I and stage II tumors, respectively. Those with regional disease had 31% 5-year DSS (vs. 28% overall survival for stage III tumors in the study by Ng et al.), which decreased to 7% 5-year DSS in those with distant disease (vs. Ng et al.’s 12% and 7% 5-year overall survival for stage IVA and IVB tumors, respectively).11 In this proposed staging system, the single worst prognostic factor was tumor rupture, which, along with incomplete resection, was found to be a significant adverse prognostic factor by multivariate analysis.11 This proposed staging system conveys more detailed information than the general summary stage classification and could help to determine the prognosis of patients with greater precision.

Survival
The overall 5-year DSS of NCDB patients with SBS was 39%, and the median survival was 34.1 months. Factors significantly correlated with decreased DSS were advancing age, male sex, epithelioid or Kaposi’s histology, positive nodes, high grade, tumor size >5 cm, regional or distant spread of disease, and not having cancer-directed surgery. By multivariate analysis, only tumor size, histology, and generally summary stage were found to be significant independent prognostic factors for DSS in SBS. These results are compared with other series of SBS in Table 6.


View this table:
[in this window]
[in a new window]
  		TABLE 6. Summary of studies examining prognostic factors in small bowel sarcomas
 
Shiu et al.19 reported the results of 38 patients with myosarcomas of the large (n = 20) and small intestine (n = 18) treated at MSKCC between 1949 and 1974 . They found that tumor size >5 cm, free perforation or local invasion, and grade were correlated with patient survival, but formal statistical analysis was not performed. Patients with tumors <5 cm in size had a 71% 5-year survival, which decreased to 27% for those with tumors >5 cm in size. The 5-year survival of those with local invasion or perforation was 28%, as compared with 81% for those with tumors confined to the small or large intestine. Patients with low-grade tumors had a 62% 5-year disease-free survival rate, as compared with only 12% for those with high-grade tumors. They proposed a staging scheme in which the stage was equal to the total number of these three adverse factors (size, invasion/perforation, grade) that were positive. They found 5-year survival rates of 100% for patients with stage 0, 44% for stage I, 31% for stage II, and 0% for stage III tumors.

The AFIP study20 found significant differences in survival of patients with GI smooth muscle tumors with respect to their site of origin. The 5-year survival of patients with small bowel tumors was the lowest, at approximately 58%, vs. 74% for esophageal tumors. The higher survival rates seen in this study were likely related to the inclusion of benign smooth muscle tumors.20 Multivariate analysis for all sites of GI smooth muscle tumors revealed that tumor size, site, and patient age were significant independent prognostic factors for survival. Mitotic index was also found to be significant prognostic factor in the entire group of tumors, but was not predictive of survival at the small bowel site.

Ng et al.11 from MDACC grouped all leiomyosarcomas of the GI tract and found that the resection without tumor rupture, extent of disease (localized, resectable implants, or contiguous organ), and tumor grade were all significant prognostic factors by multivariate analysis, whereas tumor site and size were significant by univariate, but not by multivariate analysis. Actuarial 5-year survival for all sites of GI leiomyosarcomas was 28% with a median survival of 29 months, and in those having complete resection, the median survival increased to 48 months. The median survival of 44 patients with SBS having complete resection was 48 months (n = 44), compared with 24 months in patients with incomplete resection (n = 18) and 19 months for those with complete resection but tumor rupture (n = 12). Tumor size and grade were not significantly associated with survival in tumors at the small bowel site.11

In the MSKCC series of GIST tumors, factors associated with diminished survival by multivariate analysis were male sex, tumor size >5 cm, and unresectable lesions. This included 93 patients with primary lesions, 94 patients with metastases at presentation to MSKCC, and 13 patients presenting with local recurrence. In those patients who presented with primary GIST who had complete resection (n = 80), the only significant independent risk factor was tumor size >10 cm. The overall 5-year survival for the entire group was 35%, vs. 54% in patients with primary tumors having complete resection.21 An earlier study of 38 patients from MSKCC with GI sarcomas (9 from the small bowel) found that small bowel site, abdominal pain, high-grade tumors, and incomplete resection were all significant prognostic variables for decreased survival by univariate analysis, but multivariate analysis was not performed.6

Chiotasso and Fazio’s series22 of 28 small bowel leiomyosarcomas from the Cleveland Clinic revealed that duration of symptoms >12 months, distant metastases, and tumor diameter >9 cm predicted for decreased survival by univariate analysis. In a series of 45 GI leiomyosarcomas from Taiwan, male sex, tumor size >5 cm, incomplete resection, advanced (nonlocalized) tumors, and high-grade tumors predicted for decreased survival by univariate analysis. Patient age, GI bleeding, tumor necrosis, tumor ploidy, and S-phase fraction were not found to be significant by univariate analysis. By multivariate analysis, only advanced tumors (5-year survival of 35% and 0%, for localized and advanced tumors, respectively) and high grade (5 year survival of 36% and 0%, for low- and high-grade tumors, respectively) were found to be independent adverse prognostic factors .25 McGrath et al.23 from the Medical College of Virginia found that high grade, tumor size >10 cm, and nonlocalized tumors were associated with decreased survival by univariate analysis.

In future prospective studies of small bowel and other GI mesenchymal tumors, information on tumor size, localized versus locally invasive or perforated tumors, and whether complete resection was performed will be important factors to include in databases. Clearly, tumor grade is also very important but will not be of practical use until a standardized grading system is developed. The poor survival of patients with tumors >5 cm, high grade, perforation, or incomplete resection suggest that this group will be the most likely to benefit from novel treatment modalities. As shown in this study, there has been little enthusiasm for adjuvant therapy in SBS, with only 18% of patients receiving adjuvant chemotherapy or radiotherapy. The lack of response of these tumors to chemotherapy and the complications associated with radiation to the bowel are likely explanations for this observation and highlights the need for new ideas in the treatment of GI sarcomas.

When compared with other small bowel tumor types from the NCDB during this same time period, the overall 5-year DSS for SBS (38.9%) was improved compared with adenocarcinomas (5-year DSS of 30.5%).2 However, patients with carcinoid tumors (5-year DSS of 72.7%) and lymphomas (5-year DSS of 56.4%) of the small bowel had even higher survival than those with SBS. Continuing studies of these tumors will be helpful to determine which patient, tumor, operative, and treatment factors define the differing natural histories and prognosis of these malignant neoplasms of the small bowel.


   Acknowledgments
 
The authors thank Dr. Frank A. Mitros for his review and comments regarding the pathologic features of SBS.


   Footnotes
 
Presented at the 53rd Annual Meeting of the Society of Surgical Oncology, New Orleans, Louisiana, March 16–19, 2000.

Received for publication January 3, 2001. Accepted for publication April 3, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Martin RG. Malignant tumors of the small intestine. Surg Clin North Am 1986; 66: 779–85.[Medline]
  2. Howe JR, Karnell LH, Menck HR, Scott-Conner C. The American College of Surgeons Commission on Cancer and the American Cancer Society. Adenocarcinoma of the small bowel: review of the National Cancer Data Base, 1985–1995. Cancer 1999; 86: 2693–706.[CrossRef][Medline]
  3. Greenlee RT, Murray T, Bolden S, Wingo PA. Cancer statistics, 2000. CA Cancer J Clin 2000; 50: 7–33.[Abstract]
  4. Weiss NS, Yang CP. Incidence of histologic types of cancer of the small intestine. J Natl Cancer Inst 1987; 78: 653–6.
  5. DiSario JA, Burt RW, Vargas H, McWhorter WP. Small bowel cancer: epidemiological and clinical characteristics from a population-based registry. Am J Gastroenterol 1994; 89: 699–701.[Medline]
  6. Conlon KC, Casper ES, Brennan MF. Primary gastrointestinal sarcomas: analysis of prognostic variables. Ann Surg Oncol 1995; 2: 26–31.[Abstract]
  7. Skandalakis JE, Gray SW. Smooth-muscle tumors of the alimentary tract.In: Ariel IM, ed. Progress in Clinical Cancer, vol 1. New York: Grune & Stratton, 1965: 692–708.
  8. Lewin KJ, Riddell RH, Weinstein WM. Gastrointestinal Pathology and Its Clinical Implications. New York: Igaku-Shoin Medical Publishers, 1992.
  9. Erlandson RA, Klimstra DS, Woodruff JM. Subclassification of gastrointestinal stromal tumors based on evaluation by electron microscopy and immunohistochemistry. Ultrastruct Pathol 1996; 20: 373–93.[Medline]
  10. Dougherty MJ, Compton C, Talbert M, Wood WC. Sarcomas of the gastrointestinal tract: separation into favorable and unfavorable prognostic groups by mitotic count. Ann Surg 1991; 214: 569–74.[Medline]
  11. Ng EH, Pollock RE, Munsell MF, Atkinson EN, Romsdahl MM. Prognostic factors influencing survival in gastrointestinal leiomyosarcomas: implications for surgical management and staging. Ann Surg 1992; 215: 68–77.[Medline]
  12. Cancer Co. Data Acquisition Manual. 2nd ed. Chicago: American College of Surgeons, 1994.
  13. Beahrs O, Henson DE, Hutter RVP, Kennedy BJ. Manual for Staging of Cancer. 4th ed. Philadelphia: JB Lippincott, 1992.
  14. World Health Organization. International Classification of Diseases for Oncology. 2nd ed. Geneva, Switzerland: WHO, 1992.
  15. Jessup JM, Steele GD, Winchester DP. Introduction.In: Steele GD, Jessup JM, Winchester DP, Menck HR, Murphy GP, eds. National Cancer Data Base: Annual Review of Patient Care 1995. Atlanta: American Cancer Society, 1995: 1–11.
  16. Wattenberg LW. Carcinogen-detoxifying mechanisms in the gastrointestinal tract. Gastroenterology 1966; 51: 932–5.[Medline]
  17. Lowenfels AB. Why are small-bowel tumours so rare? Lancet 1973; 1: 24–6.[Medline]
  18. Wilson JM, Melvin DB, Gray GF, Thorbjarnarson B. Primary malignancies of the small bowel: a report of 96 cases and review of the literature. Ann Surg 1974; 180: 175–9.[Medline]
  19. Shiu MH, Farr GH, Egeli RA, Quan SH, Hajdu SI. Myosarcomas of the small and large intestine: a clinicopathologic study. J Surg Oncol 1983; 24: 67–72.[Medline]
  20. Emory TS, Sobin LH, Lukes L, Lee DH, O’Leary TJ. Prognosis of gastrointestinal smooth-muscle (stromal) tumors: dependence on anatomic site. Am J Surg Pathol 1999; 23: 82–7.[CrossRef][Medline]
  21. DeMatteo RP, Lewis JJ, Leung D, Mudan SS, Woodruff JM, Brennan MF. Two hundred gastrointestinal stromal tumors: recurrence patterns and prognostic factors for survival. Ann Surg 2000; 231: 51–8.[CrossRef][Medline]
  22. Chiotasso PJ, Fazio VW. Prognostic factors of 28 leiomyosarcomas of the small intestine. Surg Gynecol Obstet 1982; 155: 197–202.[Medline]
  23. McGrath PC, Neifeld JP, Lawrence W Jr, Kay S, Horsley JS, Parker GA. Gastrointestinal sarcomas: analysis of prognostic factors. Ann Surg 1987; 206: 706–10.[Medline]
  24. He LJ, Wang BS, Chen CC. Smooth muscle tumours of the digestive tract: report of 160 cases. Br J Surg 1988; 75: 184–6.[CrossRef][Medline]
  25. Chou FF, Eng HL, Sheen-Chen SM. Smooth muscle tumors of the gastrointestinal tract: analysis of prognostic factors. Surgery 1996; 119: 171–7.[CrossRef][Medline]
  26. Golden T, Stout AP. Smooth muscle tumors of the gastrointestinal tract and retroperitoneal tissues. Surg Gynecol Obstet 1941; 73: 784–810.
  27. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 1998; 279: 577–80.[Abstract/Free Full Text]
  28. Kindblom LG, Remotti HE, Aldenborg F, Meis-Kindblom JM. Gastrointestinal pacemaker cell tumor (GIPACT): gastrointestinal stromal tumors show phenotypic characteristics of the interstitial cells of Cajal. Am J Pathol 1998; 152: 1259–69.[Abstract]
  29. Miettinen M, Sarlomo-Rikala M, Lasota J. Gastrointestinal stromal tumors: recent advances in understanding of their biology. Hum Pathol 1999; 30: 1213–20.[CrossRef][Medline]
  30. Wang WL, Healy ME, Sattler M, et al. Growth inhibition and modulation of kinase pathways of small cell lung cancer cell lines by the novel tyrosine kinase inhibitor STI 571. Oncogene 2000; 19: 3521–8.[CrossRef][Medline]
  31. Heinrich MC, Griffith DJ, Druker BJ, Wait CL, Ott KA, Zigler AJ. Inhibition of c-kit receptor tyrosine kinase activity by STI 571, a selective tyrosine kinase inhibitor. Blood 2000; 96: 925–32.[Abstract/Free Full Text]
  32. Franquemont DW. Differentiation and risk assessment of gastrointestinal stromal tumors. Am J Clin Pathol 1995; 103: 41–7.[Medline]
  33. Evans HL. Smooth muscle tumors of the gastrointestinal tract: a study of 56 cases followed for a minimum of 10 years. Cancer 1985; 56: 2242–50.[CrossRef][Medline]
  34. Fleming ID, Cooper JS, Henson DE, et al. AJCC Cancer Staging Manual. 5th ed. Philadelphia: Lippincott, Williams & Wilkins, 1998.




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Howe, J. R.
Right arrow Articles by Scott-Conner, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Howe, J. R.
Right arrow Articles by Scott-Conner, C.
Related Collections
Right arrow Other Malignancies

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS