Ann Thorac Surg 2007;84:1717-1723. doi:10.1016/j.athoracsur.2007.05.071
© 2007 The Society of Thoracic Surgeons
Original Articles: General Thoracic
Surgical Considerations for the Management and Resection of Esophageal Gastrointestinal Stromal Tumors
Matthew G. Blum, MDa,*,
Karl Y. Bilimoria, MDb,
Jeffrey D. Wayne, MDb,
Alberto L. de Hoyos, MDa,
Mark S. Talamonti, MDb,
Brian Adley, MDc
a Department of Thoracic Surgery, Northwestern Memorial Hospital, Chicago, Illinois
b Department of Surgical Oncology, Northwestern Memorial Hospital, Chicago, Illinois
c Department of Pathology, Northwestern Memorial Hospital, Chicago, Illinois
Accepted for publication May 29, 2007.
* Address correspondence to Dr Blum, 201 E. Huron, Galter 10-105, Chicago, IL 60611 (Email: mblum{at}nmh.org).
 |
Abstract
|
|---|
Background: Gastrointestinal stromal tumors (GISTs) are the most common sarcoma of the gastrointestinal tract, but occur rarely in the esophagus. Use of tyrosine kinase inhibitors has revolutionized therapy for GISTs but complete resection remains the treatment of choice. Esophageal GISTs require special consideration regarding perioperative treatment, evaluation, and conduct of operation.
Methods: We report our experience (from 2000 to 2003) with four patients who underwent resection of esophageal GIST, reviewed consensus management guidelines, and performed a literature review for reported cases. Query of the National Cancer Database was made to assess national experience with esophageal GISTs.
Results: All four patients had esophageal GISTs successfully differentiated from leiomyomata by immunohistochemistry on fine needle aspirates obtained during endoscopic ultrasound. At the time of resection all tumors were noted to be adherent to tissues other than submucosa that required resection of esophageal muscularis, periesophageal tissue, or mucosa. All patients were thought to have complete resection and were alive at last contact (median, 33 months). Two of four patients recurred, one of whom underwent esophagectomy; the other patient is alive on imatinib. Seven case reports with median follow-up of three years reported no recurrence with either enucleation or esophagectomy. The National Cancer Database contains 33 cases of esophageal GIST reported since 1999. Nine of 24 patients died within six months of diagnosis.
Conclusions: Esophagectomy is the treatment of choice for esophageal GISTs but requires attention to operative details that emphasize complete en bloc excision of surrounding tissue. Preoperative imatinib may improve resectability.
|
Introduction
|
|---|
Because esophageal gastrointestinal stromal tumors (GISTs) are rare, their natural history, operative findings, and technical issues pertaining to resection remain poorly defined. Although pathologic series of esophageal GISTs have been published [1], surgical cases have been limited to single patient case reports or cases included within series of GISTs from other primary sites [2–11].
Substantial experience with resection of serosal-lined intraabdominal organs, where wedge or segmental resection is possible, has led to the development of consensus guidelines by a panel of experts in the field. However, GISTs of the esophagus are more difficult to manage because of the lack of tumor confinement by a serosal layer and the relative contraindication to segmental esophageal resections. Even diagnosing esophageal GISTs may be difficult due to a clinical appearance that overlaps with the far more common esophageal leiomyoma. In this study we report the largest surgical series to date of esophageal GISTs and discuss their management in the context of current clinical guidelines and our own experience. We also report the experience with esophageal GISTs reported in the American College of Surgeons National Cancer Database and a comprehensive summary of published case reports.
|
Patients and Methods
|
|---|
Within the last four years, four patients have undergone surgical resection of esophageal GISTs at our institution. Clinicopathologic factors, endoscopic findings, radiographic appearance, details of operative resection, operative findings, and clinical course of each patient were obtained from the medical records. The histologic and immunohistologic characteristics of the tumors, including CD34 and c-kit (CD117) expression were reviewed. This study was approved by our Institutional Review Board, all uniquely identifiable clinic data were blinded, and waiver of consent was obtained.
We performed a Medline search to identify cases of reported esophageal GISTs. We examined the bibliography of all identified case reports for additional studies not captured by our Medline search. Data from these reports were tabulated. Authors of case reports were contacted by e-mail for additional follow-up.
The National Cancer Database (NCDB) is a program of the American College of Surgeons that has been collecting data on incident cancers since 1985. Over 1,440 hospitals report to the NCDB. This accounts for approximately 75% of all cancer (new) diagnoses in the United States annually. Based on the American Cancer Society estimates, the NCDB currently captures approximately 81% of all esophageal cancers in the United States. Using the National Cancer Data Base (1985 to 2004), patients were identified based on the International Classification of Diseases for Oncology, Second and Third Edition (ICD-O-2/3) for tumors of the esophagus: C15.0 to C19.0. Prior to the use of c-kit, GISTs may have been classified as leiomyosarcomas. To help evaluate this possibility and gather information on known GISTs, histology ICD-O-2/3 codes were used to select patients with esophageal GISTs (ICD-O code 8936) and esophageal leiomyosarcomas (ICD-O code 8890). Survival data for patients who underwent resection compared with those who did not was estimated by the Kaplan-Meier method and compared using the log rank test.
|
Results
|
|---|
Report of Cases
Four patients underwent surgical resection of esophageal GIST. All occurred in the lower third of the esophagus. Esophagoscopy and endoscopic ultrasound (EUS) demonstrated smooth, submucosal lesions clinically indistinguishable from leiomyoma. All patients underwent EUS-guided fine needle aspiration (FNA) of their tumors and cells from each were immunohistochemically positive for CD117. All patients initially underwent esophageal sparing, wide local excision of the tumor and tissues to which it was adherent including mucosa and muscularis. Subtle but definitively different features that distinguished GIST resection from typical leiomyomata resection were as follows: (1) GISTs had a translucent waxy appearance on gross examination; (2) GISTs were difficult to enucleate without violating the tumor due to poor tumor coherence and lack of a true capsule; (3) all had some adhesion to the submucosa or muscularis; and (4) adequacy of resection was difficult to establish at the time of operation due to lack of substantial capsule and adhesion to surrounding structures. Two patients had involvement of the gastroesophageal junction. Both underwent transmural resection with closure of the esophageal defect and fundoplication reconstruction to support the closure. No patients were treated with imatinib in the neoadjuvant setting. The relevant clinicopathologic features are noted in Table 1. No patient had nodal metastasis and all had complete resections. After three years of follow-up, two patients suffered relapse of their disease. One patient had local recurrence only. He underwent an Ivor-Lewis esophagectomy and reconstruction with a gastric pull-up and is without evidence of disease 19 months after his second resection. The second patient with recurrence developed both periesophageal and intrapleural tumor and is currently alive with disease on imatinib therapy.
Cancer Registry Cases
The NCDB contained 33 cases of esophageal GIST with the first case reported in 1999. Since then there has been a relative increase in the number of reported GISTs and a decrease in the number of leiomyosarcomas reported (Fig 1). The demographic information and treatments used are listed in Table 2. Observed survival for all 33 patients is shown in Figure 2. Patients who underwent resection have substantially better survival.
View larger version (51K):
[in this window]
[in a new window]
|
Fig 1. Reported cases of esophageal leiomyosarcoma and gastrointestinal stromal tumors (GISTs). (Black area = GIST; grey area = leiomyosarcoma.)
|
|
Published Case Report Findings
Review of the literature found eight case reports, the details of which are listed in Table 3. Of seven cases noting a reason for presentation, three patients had dysphagia and three were noted as incidental. Tumors occurred in the lower third of the esophagus and were resected by several different techniques. Three patients had enucleation with a fourth undergoing enucleation in addition to resection of a hepatic metastasis. Only the patient with hepatic metastasis was treated with imatinib. There have been no reported recurrences in a median three years of follow-up.
|
Comment
|
|---|
Management guidelines for GISTs have been defined by consensus of the National Comprehensive Cancer Network (NCCN) and the European Society of Medical Oncology [12, 13]. Although some basic surgical guidelines are enumerated by these panels, there have been no studies specifically addressing the issue of surgical resection for esophageal GISTs. In large part this may be due to the relative rarity of the tumors, such that no one center has developed a substantial clinical experience. Only 1% of 1,458 GIST cases reported from the Surveillance Epidemiology and End Results database were esophageal in origin [14].
Esophageal GISTs present unique challenges. First, the tumor must be recognized correctly as a GIST. Due to the similar clinical, endoscopic, and radiographic appearance as the far more common esophageal leiomyoma, a GIST may not be identified as such until after resection. Because GISTs are fluorodeoxyglucose (FDG) avid, FDG-positron emission tomography (PET) scanning may be used to differentiate them from leiomyoma [10]. As our series showed, GISTs can be reliably identified preoperatively by EUS-guided FNA. While all GISTs in our series had physical characteristics noted at the time of resection that suggested they were not benign leiomyomata (poor integrity, no capsule, waxy appearance), frozen section was unreliable for definitively diagnosing a GIST as this requires immunohistochemistry.
Although they share similar clinical, endoscopic, and radiographic characteristics, leiomyoma, schwannoma, leiomyosarcoma, and GIST are distinct pathologic entities. A pathologic study combining all esophageal mesenchymal tumors from the Armed Forces Institute of Pathology and Hartman Institute in Helsinki, reported that 25% (17 of 68) of tumors were GISTs [1]. They noted several pathologic characteristics that distinguished GISTs from leiomyomata, the most important of which was the expression of c-kit and CD34 by immunohistochemistry. Prior to the use of these markers, esophageal GISTs were often grouped with leiomyosarcomas, leiomyomata, and schwannomas [15]. Cancer registry evaluation corroborated this idea by showing a relative increase in the number of GISTs reported and a corresponding decrease in leiomyosarcomas, where the total number of tumors reported remained relatively stable. In the past this may have led to underdiagnosis in historical series and may account for some cases of "recurrent esophageal leiomyoma."
Traditionally, it has been suggested that smooth, well-circumscribed, submucosal lesions identified on esophagoscopy with ultrasound should not undergo biopsy [16]. Such lesions are usually leiomyomata and scarring due to biopsy has been suggested to complicate otherwise simple enucleation. Due to the presence of expertise with EUS and FNA at our institution, most patients with esophageal tumors now undergo needle biopsy prior to surgical consultation. Despite this, in two patients undergoing enucleation of true leiomyomata after biopsy, we did not encounter adhesion of tumor to muscularis or mucosa that complicated thoracoscopic enucleation. In contrast, all of the GISTs encountered in this series had some adhesion to the mucosa or the muscularis. Whether this was a consequence of biopsy or due to inherent tumor behavior remains unclear. Nevertheless, we recommend that tumors larger than 2 cm, those that are enlarging on serial examination, or those with activity on PET scan should undergo EUS with FNA as these are not typical features of typical leiomyomata.
Although small intestinal and gastric GISTs may be resected with segmental or wedge resections, esophageal GIST resections are essentially limited to either simple enucleation or esophagectomy. In one of our cases, enucleation of a large tumor resulted in recurrence and therefore this cannot be recommended. The NCCN guidelines state that enucleation of small (<2 cm) esophageal GIST may be acceptable and that small intraabdominal tumors might be resected laparoscopically, but we would also suggest that the poor integrity of esophageal GISTs makes thoracoscopic enucleation unadvisable. One of our patients with a larger GIST developed local recurrence after full thickness, partial circumference, resection of the distal esophagus and the gastroesophageal junction. Reconstruction of the gastroesophageal junction in a second patient after tumor resection was difficult and resulted in leak. Although three of eight case reports have not noted recurrence after local enucleation, follow-up is relatively short, the patients were younger than ours or those in the database and at least one had favorable histology (no mitosis or necrosis), and one was treated with imatinib. Therefore, we recommend esophagectomy for resection of larger tumors and those involving the gastroesophageal junction. For small lesions (less than 2 cm) confined to the wall of the esophagus, particularly in patients unable to tolerate esophagectomy, an open local resection may be an acceptable alternative if a margin negative resection can be obtained.
Successful surgical treatment of GISTs depends on complete local resection. The approach to esophagectomy for GISTs should minimize blunt or blind dissection as this will not reliably include maintaining the thin potential barrier of pleura that may overlie extramucosal tumor. Additionally, poor tumor integrity and lack of esophageal serosa increase the risk of tumor rupture with blunt dissection. Transhiatal esophagectomy would likely violate tumors of the distal and midesophagus that extend beyond the muscularis and cannot be recommended. A transthoracic en bloc resection of the pleura overlying the esophagus and any involved surrounding tissues, including diaphragm, is advisable to avoid microscopically or macroscopically incomplete resection. A left thoracoabdominal approach is advocated for larger tumors at the gastroesophageal junction as this will allow excellent visualization of the parahiatal tissue. This should facilitate en bloc resection of pleura and peritoneum overlying the tumor and the diaphragm surrounding tumor at the hiatus if needed.
Because of the potentially high morbidity of esophagectomy and the relative lack of a substantial barrier to local extension that makes complete resection difficult, imatinib should be considered as neoadjuvant therapy for larger tumors. Cytoreduction may decrease the risk of tumor rupture and increase the likelihood of potentially curative complete resection. Whether to use imatinib as a neoadjuvant agent prior to planned resection or to use resection as salvage therapy after imatinib failure has yet to be established, but is a secondary goal of the Radiation Therapy Oncology Group trial S-0132.
Postresection monitoring with interval computed tomographic scan is the mainstay of follow-up even for small, apparently completely resected tumors [12]. Those at higher risk for recurrence (> 5 cm tumors or with > 5 mitoses/50 high powered fields) should be considered for more frequent scans (every three to four months). Although not recommended as routine follow-up screening, PET is sensitive for evaluating the extent of tumor recurrence and differentiating tumor from scar. Patients with unresectable recurrence should receive a trial of imatinib [12]. If local resection was attempted initially and recurrence is confined to the esophagus, these patients may benefit from esophagectomy.
Outcomes after esophageal GIST resection have not been reported in large numbers but recurrence and mortality rates are high. The five-year survival after diagnosis of esophageal GIST was 14% from the Surveillance Epidemiology and End Results database [14]. No breakdown by therapy was given but patients with partial or total organ removal had significantly lower mortality rates when all patients were included. Miettinen and colleagues [1] noted that seven of their patients underwent "transmural esophageal resection," and ten had mucosal sparing operations, but they did not examine the impact of type of resection on outcome. Although none of the individual case reports have had a recurrence to date, 9 of 17 patients in the series by Miettinen and colleagues died of disease within five years. Median survival has not been reached with the few patients in the NCDB but only slightly over half were alive at two years. Based on the available data complete surgical resection is still most likely to provide the best chance of survival.
Diagnosis and surgical treatment of esophageal GISTs present a number of unique clinical challenges. Further reporting of surgical experiences, as well as the results of ongoing clinical trials including Z-9000 and Z9001 from the American College of Surgeons Oncology group, will hopefully improve long-term outcomes with these difficult tumors. Treatment of patients with these rare lesions is most likely to be advanced if care of these patients can be regionalized to multidisciplinary sarcoma centers where there also exists an experienced group of esophageal surgeons.
|
References
|
|---|
- Miettinen M, M Sarlomo-Rikala M, L Sobin L, J Lasota J. Esophageal stromal tumors: a clinicopathologic, immunohistochemical, and molecular genetic study of 17 cases and comparison with esophageal leiomyomas and leiomyosarcomas Am J Surg Pathol 2000;24:211-222.
- Padula A, Chin N, Azeez S, Resetkova E, Andriko JA, Miettinen M. Primary gastrointestinal stromal tumor of the esophagus in an HIV-positive patient Ann Diagn Pathol 2005;9:49-53.
- Lee JR, Anstadt MP, Khwaja S, Green LK. Gastrointestinal stromal tumor of the posterior mediastinum Eur J Cardiothorac Surg 2002;22:1014-1016.
- Ertem M, Baca B, Dogusoy G, Erguney S, Yavuz N. Thoracoscopic enucleation of a giant submucosal tumor of the esophagus Surg Laparosc Endosc Percutan Tech 2004;14:87-90.
- Gouveia AM, Pimenta AP, Lopes J, et al. Esophageal GIST: therapeutic implications of an uncommon presentation of a rare tumor Dis Esophagus 2005;18:70-73.
- Feakins RM, Mears L, Atkinson P, Hughes F. Oesophageal gastrointestinal stromal tumour masquerading as neuroendocrine carcinoma Histopathology 2005;47:320-334.
- Manu N, Richard P, Howard S. Bleeding esophageal GIST Dis Esophagus 2005;18:281-282.
- Basoglu A, Kaya E, Celik B, Yildiz L. Giant gastrointestinal stromal tumor of the esophagus presenting with dyspnea J Thorac Cardiovasc Surg 2006;131:1198-1199.
- Huang CS, Hsu WH, Wu YC, Chau GY, Tsay SH, Huang MH. Enucleation of an advanced esophageal gastrointestinal stromal tumor with liver metastasis J Gastroenterol Hepatol 2006;21:482-487.
- Chang WC, Tzao C, Shen DHY, Cheng CY, Yu CP, Hsu HH. Gastrointestinal stromal tumor (GIST) of the esophagus detected by positron emission tomography/computed tomography Dig Dis Sci 2005;50:1315-1318.
- Wada Y, Kadokura M, Kitami A, Nakajima H, Inoue H, Shiokawa A. Esophageal gastrointestinal stromal tumor surrounding the middle esophagus with dysphagia for 8 years; report of a case Jpn J Thorac Surg 2004;457:1250-1253.
- Demetri GD, Baker LH, Benjamin R, et al. DNCCN, soft tissue sarcomaClinical practice guidelines in oncology. National Comprehensive Cancer Network; 2006v.3.
- Blay JY, Bonvalot S, Casali P, et al. Consensus, meeting for the management of gastrointestinal stromal tumors: report of the GIST consensus, conference of 20-21 March 2004, under the auspices of ESMO Ann Onc 2005;16:566-578.
- Tran T, Davila JA, El-Serag HB. The epidemiology of malignant gastrointestinal stromal tumors: an analysis of 1,458 cases from 1992 to 2000 Am J Gastroenterol 2005;100:162-168.
- Ueyama T, Guo KJ, Hashimoto H, Daimaru Y, Enoji M. A clinicopathologic and immunohistochemical study of gastrointestinal stromal tumors Cancer 1992;69:947-955.
- Cattaneo SM, Yang SC. Benign tumors of the esophagusIn: Yang SC, Cameron DC, editors. Current Therapy in Thoracic and Cardiovascular Surgery. Philadelphia, PA: Mosby; 2004. pp. 339-340.
Related Article
-
Invited commentary
- Norman Snow
Ann. Thorac. Surg. 2007 84: 1723.
[Extract]
[Full Text]
[PDF]
This article has been cited by other articles:
|
|
|
|
 
N. Snow
Invited commentary
Ann. Thorac. Surg.,
November 1, 2007;
84(5):
1723 - 1723.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
