GENERALTHORACIC

Novel Diagnosis and Treatment of EsophagealGranular Cell Tumor: Report of 14 Cases andReview of the LiteratureWen-shu Chen, MD, Xiao-ling Zheng, MD, Long Jin, MD, Xiao-jie Pan, MD, andMing-fan Ye, MDDepartments of Thoracic Surgery, Gastrointestinal Endoscopy Center, and Pathology, Fujian Provincial Hospital, Fujian MedicalUniversity, Fuzhou, China

Background. Granular cell tumors (GCT), especially inthe esophagus, are rare neoplasms originating from thenervous system. There is still some controversy regardingthe diagnosis and treatment of esophageal GCT.

Methods. We report 14 cases of esophageal GCTdiagnosed and treated from January 2004 to March 2013.Their clinical manifestations, endoscopic image, endo-scopic ultrasonography (EUS) appearance, pathology,immunohistochemistry, treatment, and prognosis werereviewed.

Results. The typical images of EUS were hypoechoic,homogenous, and smooth-edged tumors restricted todeep mucosal and submucosal layers. However, therewere 2 cases with tumors invading muscular layer.Endoscopic ultrasonography was valuable to assess thetumor size, location, depth of invasion, and nature. Ac-cording to EUS manifestation, 11 cases with lesions 3 cm

Accepted for publication Aug 19, 2013.

Address correspondence to Dr Pan, Department of Thoracic Surgery,Fujian Provincial Hospital, Fujian Medical University, No. 134 East St,Fuzhou, Fujian 350001, China; e-mail: doctorpxj@163.com.

� 2014 by The Society of Thoracic SurgeonsPublished by Elsevier Inc

or less in diameter without muscular layer invasion un-derwent endoscopic resection without complication andthe other 3 cases underwent surgical resection. A newtechnique of submucosal tunnel endoscopic resectionwas performed in 3 submucosal cases with lesionsranging from 2 cm to 3 cm in diameter. All of these caseswere benign and histology was necessary for differentialdiagnosis.Conclusions. Endoscopic ultrasonography plays an

important guiding role in the diagnosis and treatment ofesophageal GCT. Submucosal tunnel endoscopic resec-tion is safe and effective. Further study is needed todetermine whether this technique can be expanded intoother applications.

(Ann Thorac Surg 2014;97:296–302)� 2014 by The Society of Thoracic Surgeons

ranular cell tumors (GCTs), previously called gran-

Gular cell myoblastomas or Abrikossoff’s tumors, areuncommon neoplasms first described in the tongue byAbrikossoff in 1926.

Most GCTs are benign lesions occurring in the softtissues throughout the body, most commonly in the skin,tongue, subcutaneous tissue, and skeletal muscle, but alsoin the nervous system, respiratory tract, female genitaltract, and all segments of the gastrointestinal tract.Approximately 5% to 11% of GCTs occur in the gastro-intestinal tract, with approximately one third in theesophagus [1–3]. Most of the esophageal GCT cases havebeen reported either from small series or single reports.Therefore, there is still some controversy regarding thediagnosis and treatment of esophageal GCTs. With theprogress of endoscopic technology and wide use ofendoscopic ultrasonography, now there is a new oppor-tunity for the development of the diagnosis and treatmentof this rare tumor. Previous studies have proved thesafety and efficiency of endoscopic resection for lesions

confined to mucosa and submucosa with diameters of 10mm or less [4–7].In this article, we present a clinical and pathologic

study of 14 cases of esophageal GCTs, including theirclinical manifestations, endoscopic image, endoscopicultrasonography (EUS) appearance, pathology, immuno-histochemistry, treatment, and prognosis. In our cohort,endoscopic resection was performed in 11 cases with di-ameters of 3 cm or less. We applied the technique ofsubmucosal tunnel endoscopic resection for submucosallesions ranging 2 cm to 3 cm in diameter. To ourknowledge, this is the first report of applying this tech-nique on the treatment of esophageal lesions. The ulti-mate aim of our study is to assess the guideline value ofEUS for the diagnosis and treatment of this rare diseaseand to emphasize adopting individual therapy accordingto different conditions.

Material and Methods

Patients with esophageal GCT were reviewed fromJanuary 2004 to March 2013. Chest computed tomography(CT) scan, endoscopy, EUS, and abdominal ultrasonog-raphy were performed before treatment in all patients toevaluate their tolerance to treatment and biological

0003-4975/$36.00http://dx.doi.org/10.1016/j.athoracsur.2013.08.042

297Ann Thorac Surg CHEN ET AL2014;97:296–302 ESOPHAGEAL GRANULAR CELL TUMORS

GENERALTHORACIC

characteristics of the tumor. Patient medical records,including basic information, clinical manifestations,endoscopic image, EUS appearance, pathology, immu-nohistochemistry, therapy, and prognosis, were reviewedand analyzed (Table 1).

Criteria for Treatment SelectionThe criteria for treatment selection were (1) uppergastrointestinal symptoms such as dysphagia; (2) lesions 1cm or greater in maximum diameter, or lesions 1 cm orless in patients with strong demand for treatment; (3)rapid recent growth; and (4) lesions that were difficult todistinguish from other malignant neoplasms.

Endoscopic ResectionIdeal lesions for endoscopic resection should be 3 cm orless in maximum diameter confined to mucosal or sub-mucosal layers. Resection methods included endoscopicmucosal resection (EMR) and submucosal tunnel endo-scopic resection. The EMR was performed for lesionslocated in the mucosa and tumors less than 2 cm indiameter confined to the submucosa.

Submucosal tunnel endoscopic resection (Fig 1) wasespecially suitable for lesions ranging from 2 to 3 cm indiameter located in the submucosa. The initial mucosalincision was made by endoscopic mucosectomy at about 5cm above the tumor to provide mucosal entry to thesubmucosal space. Epinephrine or similar solution wasinjected into submucosa to separate the superficial mu-cosa from muscularis propria and create a submucosalcavity. And then, a submucosal tunnel was created, asdescribed previously [8]. Through the submucosal tunnel,the tumor was dissected from the overlying mucosa or

Table 1. Characteristics of 14 Esophageal Granular Cell Tumors

CaseNo.

Age(Years), Sex

InitialSymptoms

TumorLocationa (cm)

1 40, male Dysphagia 282 52, male Asymptomatic 383 47, female Asymptomatic 254 44, male Dysphagia, weight

loss38

5 42, male Weight loss 386 57, male Dysphagia, retrosternal

pain, weight loss32–40

7 45, female Dysphagia 228 45, female Dysphagia 24

9 57, female Dysphagia 20

10 56, female Epigastric pain 3311 39, female Dysphagia 2912 60, male Dysphagia 3513 50, female Asymptomatic 3714 54, female Epigastric pain 27

a Distance from cutting teeth.

EMR ¼ endoscopic mucosal resection; STER ¼ submucosal tunnel endosc

submucosa and then carefully removed with aninsulated-tip knife. After the specimen was extractedthrough the tunnel, EUS was performed to ensure com-plete resection of the tumor. After complete hemostasis inthe submucosal tunnel, the entry orifice was tightly closedwith clips. The whole process was performed undergeneral anesthesia with intravenous administration.

SurgeryFor lesions with a high suspicion of malignancy orgreater than 3 cm in maximum diameter or originatingfrom the muscularis propria, we chose surgical resection,including traditional open surgery and minimally inva-sive surgery (video-assisted thoracoscopic surgery). Sim-ple tumor enucleation was chosen for lesions confinedto the muscularis propria or located in the cervicalesophagus. Subtotal esophagectomy was selected for le-sions invading the adventitia with a high suspicion ofmalignancy.

Follow-Up ProtocolFollow-up visits were performed at the end of treatmentevery 6 months for the first 2 years and then annually foran additional 8 years. All patients were monitored withthoracic CT, endoscopy, and EUS to evaluate treatmenteffectiveness.

Results

Fourteen patients with esophageal GCT were identified,with a male-to-female ratio of 6:8. The average age ofpatients was 49.14 years (�6.93; range, 39 to 60). The mostcommon initial symptoms for endoscopy were dysphagia

InvadingLayer

MaximumDiameter (cm)

Method ofExcision

Length ofFollow-Up

Mucosa 2.5 EMR 5 yearsSubmucosa 1.5 EMR 5 yearsSubmucosa 0.8 EMR 4 yearsMucosa 2.1 EMR 4 years

Submucosa 2.8 STER 4 yearsMuscularis

propria8 Subtotal

esophagectomy2 years

Submucosa 2.5 STER 2 yearsSubmucosa 4.0 Transthoracic

excision14 months

Muscularispropria

2.3 Transcervicalexcision

12 months

Mucosa 1.3 EMR 8 monthsMucosa 3 EMR 10 yearsSubmucosa 2.7 STER 3 yearsMucosa 1.1 EMR 3 yearsSubmucosa 1.5 EMR 8 years

opic resection.

Fig 1. Operational flow charts of submuco-sal tunnel endoscopic resection. (A) Submu-cosal tunneling: a submucosal tunnel 5 cmabove the tumor is created by endoscopicmucosectomy. Epinephrine or similar solu-tion is injected into submucosa to separatethe superficial mucosa from muscularispropria and create a submucosal cavity.(B) Tumor separation and resection: thesubmucosal tumor is separated from sur-rounding tissue and dissected from muscu-laris propria and mucosal. (C) Removal ofthe submucosal tumor: the submucosal tumoris totally extracted and carefully removedwith an insulated-tip knife through the sub-mucosal tunnel. (D) Closure of the mucosalentry orifice: after complete hemostasis in thesubmucosal tunnel, the mucosal entry orificeis tightly closed with hemostatic tips.

298 CHEN ET AL Ann Thorac SurgESOPHAGEAL GRANULAR CELL TUMORS 2014;97:296–302

GENERALTHORACIC

(8 cases), weight loss (3 cases), epigastric pain (2 cases),

and retrosternal pain (1 case). Three patients wereasymptomatic and diagnosed incidentally during endo-scopic examination. All patients just had a solitary lesion.Among these 14 patients, 6, 5, and 3 cases were found tohave the lesions in the distal, middle, and proximalesophagus, respectively.

The endoscopy in these patients presented a variableappearance. Nine cases exhibited a typical yellowishappearance (Fig 2A). Four showed a white appearance(Fig 2B). One appeared a convex reddish lump with anulcer in the center (Fig 2C). All lesions were covered byintact normal mucosa.

Endoscopic ultrasonography was performed in all pa-tients. The EUS could not only measure the size but alsoreveal the origin and depth of invasion of the tumor.According to EUS, 12 cases were confined to mucosa(Fig 3A) or submucosa (Fig 3B), whereas 2 cases originatedfrommuscularis propria (Fig 3C), with 1 showing invasionof the esophageal adventitia (Fig 3D). The typical EUSimages showed hypoechoic, homogenous, and smooth-edged lesions confined to mucosal and submucosal

Fig 2. Endoscopic images of esophageal granular cell tumor. (A) A yellowireddish neoplasm with an ulcer in the center.

layers. The tumors ranged in size from 0.8 cm to 8 cm inmaximum diameter, with an average of 2.51 cm.Different treatment methods were adopted according

to the tumor size, location and invading layer. Elevencases with lesions 3 cm or less in maximum diameterunderwent endoscopic resection without complication,including EMR in 8 cases (tumors less than 2 cm inmaximum diameter; Fig 4) and submucosal tunnelendoscopic resection in 3 cases (submucosal tumorsranging from 2 cm to 3 cm in maximum diameter; Fig 1).All lesions of the 11 cases were confined to esophagealmucosa or submucosa. The other 3 cases with lesionsgreater than 3 cm in maximum diameter or originatingfrom muscularis propria underwent surgical excision,including simple tumor enucleation in 2 patients andsubtotal esophagectomy in 1. The patient with tumormeasuring 2.3 cm in diameter located in the cervicalesophagus underwent transcervical simple tumorenucleation. The patient with the largest lesion (8 cm indiameter), originating from muscularis propria layer andinvading adventitia, underwent traditional open surgery.Postoperatively, gastroesophageal reflux, which was

sh colored neoplasm. (B) A white colored neoplasm. (C) A convex

Fig 3. Typical endoscopic ultrasonographyimages of esophageal granular cell tumor,showing hypoechoic, homogenous, andsmooth-edged lesions confined to (A)mucosal layers and (B) submucosal layers.(C) A tumor confined to muscularis propria(arrow). (D) A tumor originating frommuscularis propria and invading theadventitia (arrow).

299Ann Thorac Surg CHEN ET AL2014;97:296–302 ESOPHAGEAL GRANULAR CELL TUMORS

GENERALTHORACIC

confirmed by endoscopy, occurred in this patient andseriously reduced the quality of life. Heartburn and acidregurgitation were typical symptoms, especially wheneating and in the supine position. These symptoms werealleviated by proton pump inhibitors, antacids, and life-style adjustment. However, these symptoms were notfound in other patients in posttreatment follow-up. Allpatients in this cohort underwent successful excisionwithout death. No residual lesion or recurrence wasfound in long-term follow-up ranging from 8 months to10 years. The median follow-up time was 4 years.

Pathology and immunohistochemistry were performedin all specimens. Surgical margins were negative in allspecimens. The gross appearance showed a single andfirm nodule with a gray or yellowish cross-section.Microscopically, GCTs were shown to have sheets ornests of plump, round or polygonal cells with eosinophilicgranular cytoplasm and small, round, pyknotic centraluniform nuclei (Fig 5A). The cells were often separated bycollagen fiber bundles. There was no necrosis in all tumortissues. Two cases showed invasive growth, whereas theothers displayed expansive growth. But none of them met

Fig 4. Endoscopic images of an esophagealgranular cell tumor of 2.0 cm in diameterconfined to submucosa (A) before and (B)after endoscopic mucosal resection.

Fig 5. The classical histologic characteristics of esophageal granular cell tumor. (A) Histopathologic examination revealed sheets or nests of plump,round or polygonal cells with eosinophilic granular cytoplasm and small, round, pyknotic, central, uniform nuclei. The cells were often separatedby collagen fiber bundles (hematoxylin and eosin staining, original magnification �200). (B) Both the cytoplasm and nuclei of the tumor cellsshowed positive immunohistochemical staining for S100 protein (original magnification �200). (C) The cytoplasm of the tumor cells showedpositive immunohistochemical staining for neuron-specific enolase (original magnification �200).

300 CHEN ET AL Ann Thorac SurgESOPHAGEAL GRANULAR CELL TUMORS 2014;97:296–302

GENERALTHORACIC

the diagnostic criteria for malignant GCT. Immunohis-tochemical staining showed that tumor cells were 100%strongly positive for S100 protein (Fig 5B) and neuron-specific enolase (Fig 5C), partially positive for CD68,vimentin, and CD56, and negative for CD117, Dog1,desmin, and smooth muscle actin.

Comment

EpidemiologyGranular cell tumor is a rare soft-tissue neoplasm origi-nating from the nervous system. Most of them are benign,and only 1% to 2% are malignant in previous reports [9].Since Abrikossoff [10] described the first esophageal GCTin 1931, approximately 350 cases of esophageal GCTshave been reported in the English literature [4, 11]. Mostof these cases are presented in case reports. According toprevious studies, approximately 65% of esophageal GCTsare located in the distal, 20% in the middle, and 15% inthe proximal esophagus [2, 11].

Clinical FeaturesPatients with lesions less than 20 mm in diameter areusually asymptomatic and the lesions are found inciden-tally during upper gastrointestinal endoscopy or radiog-raphy [12]. For patients with lesions 20 mm or greater indiameter, the most common presenting symptom isdysphagia. Other symptoms, such as gastroesophagealreflux disease, dyspepsia, chest pain, cough, nausea, andhoarseness, are less reported. Generally, symptomaticlesions are remarkably larger than asymptomatic tumors,as shown in Table 1 (maximum diameter). In our cohort,all patients had a single tumor. In previous studies,multifocal esophageal GCTs and multifocal GCTs simul-taneously affecting various organs have been reported ina few cases [4, 13].

Endoscopic and EUS AppearanceThe typical endoscopic appearance of esophageal GCT isa yellowish, smooth surfaced, firm lesion covered byintact overlying mucosa. However, some tumors show

white or red appearance, as in our cohort. Similar toprevious reports, it is rare that ulcerous tumors areobserved on the mucosal surface [14, 15], just as in case 6in our group. When tumors are superficial, endoscopicbiopsy can be performed to make a definitive diagnosis.In the diagnosis and treatment of esophageal GCT,

EUS plays an important role. It is valuable to assess thesize, location, and invading layer of the tumor, and todetermine the specific treatment. In our cohort, thetypical EUS images were hypoechoic, homogenous, andsmooth-edged lesions located in mucosa or submucosa,which are similar to previous reports [4, 11]. However,lesions located in muscularis propria have been rarelyreported (2 cases in our cohort) [4]. The typical EUS im-ages, in addition to the classical endoscopic appearance,are highly suggestive of esophageal GCT. AlthoughEUS can differentiate esophageal GCT from malignanttumors, we still cannot obtain a definite diagnosis.Based on EUS and endoscopy features, GCT can be

distinguished from esophageal cyst, inflammatory polyp,and lipoma [16, 17]. However, EUS and endoscopy fea-tures alone cannot reliably differentiate between neo-plasms, such as leiomyoma, located in the muscularispropria or muscularis mucosa, and GCT located in themuscularis propria [18]. Therefore, histology and immu-nohistochemical staining are essential to distinguishthese lesions.

Histology and Immunohistochemical StainingHistology is necessary for the diagnosis of this rare tumor.Most tumors are located in mucosa or submucosa. Theyare solid, firm, and nonenveloped with a yellow oryellowish cross section. The histologic characteristic isthat the tumors are composed of sheets or nests of roundor polygonal large cells with abundant eosinophilicgranular cytoplasm and small, round, central uniformnuclei [19]. The tumor cells are often separated bycollagen fiber bundles. According to previous studies,immunohistochemical staining shows that GCTs arepositive for S100 protein, neuron-specific enolase,vimentin, and various myelin proteins, and negative for

301Ann Thorac Surg CHEN ET AL2014;97:296–302 ESOPHAGEAL GRANULAR CELL TUMORS

GENERALTHORACIC

smooth muscle actin and desmin [11, 20], as in our series.These data support the neural system origin and aSchwann cell derivation. Conversely, esophageal leio-myomas are positive for smooth muscle actin and desminand negative for S100 protein [21, 22].

The malignant GCTs are rare, approximately 2% of alllesions in the literature [23]. Fanburg-Smith and associ-ates [9] have proposed histological criteria for malignantGCT, including necrosis, spindling, vesicular nucleiwith large nucleoli, a high nuclear-to-cytoplasmic ratio,increased mitotic activity (more than 2 mitoses per 10high-power fields at �200 magnification), and pleomor-phism. Tumors fulfilling at least three of these criteria areclassified as malignant, those meeting one or two criteriaare classified as atypical, and those showing only focalpleomorphism but meeting none of the other criteria areclassified as benign. Immunohistologically, malignantGCTs are highly sensitive to p53 and Ki67 protein [24],especially the former. However, sometimes histology andimmunohistology are not enough for the diagnosis ofmalignant GCTs. Clinical manifestations are also essen-tial. Rapid recent growth, local recurrence, and remotemetastasis after complete resection are believable clinicalevidence. Chatelain and colleagues [24] concluded thatthe infiltrative feature of the esophageal GCTs, by itself,cannot be considered as a malignant feature, which iscontrary to other opinions [20]. The diagnosis of malig-nant esophageal GCTs relies on the detection of metas-tasis. In our cohort, there was a case with neoplasminvading adventitia. But all cases did not meet the diag-nostic criteria for malignant GCT and there was no localrecurrence or remote metastasis in the follow-up.

There are various methods for tissue diagnosis,including endoscopic biopsy, EUS- or CT-guided fine-needle biopsy. Andrade and coworkers [25] have pro-posed that biopsy is not advised because of the risk ofbleeding, ulceration, fistula formation, secondary infec-tion, and mistaken diagnosis of squamous cell carcinoma.Therefore, periodic follow-up is advised for asymptom-atic patients with benign clinical manifestations. It ismeaningless to distinguish the specific type of thesebenign tumors. However, if it is difficult to distinguishfrom malignancy, the tumor must be resected.

TreatmentEndoscopic ultrasonography is not only of great help inthe identification of the size and location of the lesion, butalso in determining the choice of treatment. Small (lessthan 1 cm in diameter) and asymptomatic neoplasms withbenign clinical course indicate conservative approachwith only regular endoscopic and EUS follow-up, thusavoiding potential surgical or endoscopic resection com-plications [5, 11, 19]; whereas endoscopic or surgicalresection should be performed for tumors sized 1 cm orgreater, rapid recent growth, suspicion of malignancy, ortransmural infiltration, or producing symptoms such asdysphagia [11, 19, 26]. Endoscopic resection includesEMR and submucosal tunnel endoscopic resection; EMRis applied in patients with lesions 2 cm or less in diameterand nonattachment to the muscularis propria [14], and

submucosal tunnel endoscopic resection is applied inpatients with lesions ranging from 2 cm to 3 cm andlocated in submucosa. When direct EMR is difficult toperform, epinephrine or a similar solution is injected intosubmucosa to separate the superficial mucosa frommuscularis propria, raising the neoplasm, to avoidbleeding and perforation during the procedure. Submu-cosal tunnel endoscopic resection is a new, promisingtechnique for selected submucosal tumors of the esoph-agus and cardia, but not safe for neoplasms located inmuscularis propria [27]. It is more advantageous thanEMR for en bloc resection of large tumor, avoidingesophageal perforation and secondary infection. All casesof endoscopic resection in our cohort were safe withoutcomplications.Surgery, including video-assisted thoracoscopic sur-

gery and conventional open surgery, is the most invasivetreatment option with the highest incidence of com-plications. Indications for surgery are suspicion of ma-lignancy, deep layer invasion, contraindication forendoscopic dissection, or multiple symptoms [11].In conclusion, esophageal GCTs are rare neoplasms

originating from the nervous system. Yellowish or whiteappearance is the classical endoscopic image. Endo-scopic ultrasonography plays an important role inguiding the diagnosis and treatment of this rare diseaseas it not only can assess the exact size, location, andinvading layer of the tumor, but also judge the nature ofthe neoplasm. The typical EUS images are hypoechoic,homogenous, and smooth-edged lesions located in mu-cosa or submucosa. Histology is necessary for the diag-nosis of esophageal GCTs. Surgery should be chosenprudently. Endoscopic mucosal resection can be per-formed in mucosal tumors 2 cm or less in diameter.Submucosal tunnel endoscopic dissection can be per-formed in submucosal neoplasms 3 cm or less in diam-eter. However, owing to short-term application and thesmall number of cases, the safety and efficacy of this newtechnique still remains to be verified. Further study isneeded to determine whether the technique of submu-cosal tunnel endoscopic dissection can be expanded intoother medical applications.

The authors thank Professor Guo-zhang Zhu, PhD (Departmentof Biological Sciences at Marshall University) for Englishrevision.

References

1. Parfitt JR, McLean CA, Joseph MG, Streutker CJ, Al-Haddad S, Driman DK. Granular cell tumours of thegastrointestinal tract: expression of nestin and clinicopatho-logical evaluation of 11 patients. Histopathology 2006;48:424–30.

2. Lack EE, Worsham GF, Callihan MD, et al. Granular celltumor: a clinicopathologic study of 110 patients. J Surg Oncol1980;13:301–16.

3. Johnston MJ, Helwig EB. Granular cell tumors of thegastrointestinal tract and perianal region: a study of 74 cases.Dig Dis Sci 1981;26:807–16.

302 CHEN ET AL Ann Thorac SurgESOPHAGEAL GRANULAR CELL TUMORS 2014;97:296–302

GENERALTHORACIC

4. Zhong N, Katzka DA, Smyrk TC, Wang KK, Topazian M.Endoscopic diagnosis and resection of esophageal granularcell tumors. Dis Esophagus 2011;24:538–43.

5. Fotiadis C, Manolis EN, Troupis TG, Gorgoulis VG,Sechas MN. Endoscopic resection of a large granular celltumor of the esophagus. J Surg Oncol 2000;75:277–9.

6. Battaglia G, Rampado S, Bocus P, Guido E, Portale G,Ancona E. Single-band mucosectomy for granular cell tumorof the esophagus: safe and easy technique. Surg Endosc2006;20:1296–8.

7. Wehrmann T, Martchenko K, Nakamura M, Riphaus A,Stergiou N. Endoscopic resection of submucosal esophagealtumors: a prospective case series. Endoscopy 2004;36:802–7.

8. Inoue H, Minami H, Kobayashi Y, et al. Peroral endoscopicmyotomy (POEM) for esophageal achalasia. Endoscopy2010;42:265–71.

9. Fanburg-Smith JC, Meis-Kindblom JM, Fante R,Kindblom LG. Malignant granular cell tumor of soft tissue:diagnostic criteria and clinicopathologic correlation. Am JSurg Pathol 1998;22:779–94.

10. Abrikossoff AI. Weitere untersuchungen €uber myoblastenmyome [Further studies on myoblasts fibroids]. VirchowsArch Pathol Anat 1931;280:723–40.

11. De Rezende L, Lucendo AJ, Alvarez-Arguelles H. Granularcell tumors of the esophagus: report of five cases and reviewof diagnostic and therapeutic techniques. Dis Esophagus2007;20:436–43.

12. Huang AT, Dominguez LM, Powers CN, Reiter ER. Granularcell tumor of the cervical esophagus: case report and litera-ture review of an unusual cause of dysphagia. Head NeckPathol 2012 Nov 11 [E-Pub ahead of print].

13. John BK, Dang NC, Hussain SA, et al. Multifocal granular celltumor presenting as an esophageal stricture. J GastrointestCancer 2008;39:107–13.

14. Narra SL, Tombazzi C, Datta V, Ismail MK. Granular celltumor of the esophagus: report of five cases and review of theliterature. Am J Med Sci 2008;335:338–41.

15. Maekawa H, Maekawa T, Yabuki K, et al. Multiple esoph-agogastric granular cell tumors. J Gastroenterol 2003;38:776–80.

16. Buscarini E, Stasi MD, Rossi S, et al. Endosonographicdiagnosis of submucosal upper gastrointestinal tract lesionsand large fold gastropathies by catheter ultrasound probe.Gastrointest Endosc 1999;49:184–91.

17. Hizawa K, Matsumoto T, Kouzuki T, Suekane H, Esaki M,Fujishima M. Cystic submucosal tumors in the gastrointes-tinal tract: endosonographic findings and endoscopicremoval. Endoscopy 2000;32:712–4.

18. Xu GQ, Zhang BL, Li YM, et al. Diagnostic value of endo-scopic ultrasonography for gastrointestinal leiomyoma.World J Gastroenterol 2003;9:2088–91.

19. Percinel S, Savas B, Yilmaz G, et al. Granular cell tumor ofthe esophagus: three case reports and review of the litera-ture. Turk J Gastroenterol 2008;19:184–8.

20. Hulagu S, Senturk O, Aygun C, et al. Granular cell tumor ofesophagus removed with endoscopic submucosal dissection.Turk J Gastroenterol 2007;18:188–91.

21. Fei BY, Yang JM, Zhao ZS. Differential clinical and patho-logical characteristics of esophageal stromal tumors andleiomyomata. Dis Esophagus 2013 Feb 5 [E-Pub ahead ofprint].

22. Terada T. A clinicopathologic study of esophageal 860 benignand malignant lesions in 910 cases of consecutive esophagealbiopsies. Int J Clin Exp Pathol 2013;6:191–8.

23. Ordonez NG. Granular cell tumor: a review and update. AdvAnat Pathol 1999;6:186–203.

24. Chatelain D, Terris B, Molas G, Belghiti J, Degott C, Flejou J.Infiltrating granular cell tumor of the esophagus: a descrip-tion of two cases. Ann Pathol 2000;20:158–62.

25. Andrade J de S, Bambirra EA, de Oliveira CA, et al. Granularcell tumor of the esophagus: a study of seven cases diag-nosed by histologic examination of endoscopic biopsies.South Med J 1987;80:852–4.

26. Buratti S, Savides TJ, Newbury RO, Dohil R. Granular celltumor of the esophagus: report of a pediatric case and liter-ature review. J Pediatr Gastroenterol Nutr 2004;38:97–101.

27. Inoue H, Ikeda H, Hosoya T, et al. Submucosal endoscopictumor resection for subepithelial tumors in the esophagusand cardia. Endoscopy 2012;44:225–30.