Impact of endoscopic ultrasound for evaluation of submucosal lesions in gastrointestinal tract

DIGESTIUEENDDSCDPY

UIGEST LlUER GIS 2002;34:290-7

Impact of endoscopic ultrasound for evaluation of submucosal lesions in gastrointestinal tract 6. Brand L. Oesterhelweg K. F. Binmoeller PV. J. Sriram S. Bohnacker S. Seewald A. De Weerth N. Soehendra

Background. Endoscopic ultrasound is widely used following endoscopy for evaluation of suspected submucosal lesions and may guide further management of patients. Patients and Method. A total of 181 consecutive patients with suspected submucosal lesion in the upper gastrointestinal tract were diagnosed by endoscopic ultrasound between 199097. We evaluated: I) the potential of endoscopic ultrasound criteria to predict histological type of submucosal lesions in 69 patients with available histology, 2) the ability of endoscopic ultrasound alone or with clinical presentation, to predict malignancy in 86 patients with available histology or follow-up of > 12 months. Results. Sensitivity and specificity for diagnosing 44 gastrointestinal stromal tumours were 95 and 72%, respectively while 25 miscellaneous lesions were diagnosed correctly in only 56% by endoscopic ultrasound. Diagnosis of malignancy using any two of three endoscopic ultrasound criteria [heterogeneous echotexture, size >3 cm, irregular margins] showed a sensitivity of 80% and specificity of 77%, giving accurate endoscopic ultrasound diagnosis in 16/20 malignant and 5 ‘l/66 benign submucosal lesion. Heterogeneous echotexture, size >3 cm, and irregular margins showed a relative risk of 7.2, 5.4 and 4.6, respectively for presence of malignancy The presence of symptoms, potentially suggesting malignancy (dysphagia, gastrointestinal bleeding, pain and weight loss], had a relative risk of 4.2, however this did not increase the accuracy of diagnosing malignancy based on endoscopic ultrasound criteria alone. Conclusion. The accuracy of endoultrasound is high in diagnosing gastrointestinal stromal tumours, which show a significant potential of malignancy Endoscopic ultrasound morphology appears to be helpful in selection of patients for surgical or conservative treatment. The accuracy of endoscopic ultrasound in differential diagnosis of non-gastrointestinal stromal tumour lesions is limited.

Digest

From Depwtment of Interdisciplinary Endoscopy, University Hospital Eppendorf, Hamburg, Germany. Addfwetiw wbnce Dr. B. Brand, Department of Interdisciplinary Endoscopy, University Hospital Eppendorf Martinistrasse 52, 2L7246 Hemburg, Germany. Fax: +49-40-42BO3-4420. E-mail: brand@uke. uni-hamburg. de Submitted April 2, 2001. Accepted after revision Ju/y 25, 2001.

290

Liver

Key words:

Ilk

2002;34:280-7

endoscopic

ultrasound;

GI-malignancy;

GIST; submucosal

gastrointestinal

tumour

Introduction Submucosal lesions (SML) are usually incidental findings, detected on contrast radiography or during endoscopy. The differential diagnosis of SML includes a wide variety of benign and malignant non-epithelial and epithelial tumours as well as non-neoplastic lesions. The highest incidence was reported in the stomach in 0.36% of routine endoscopies performed for nonspecific symptoms ‘. The majority of SMLs are smooth muscle tumours, usually appearing as rounded, elevated lesions with normal overlying mucosa*. Since endoscopic tissue sampling is superficial, the diagnostic yield is low 3-5.

8. Brand et al.

Due to its ability to visualize the five-layer structure of the gastrointestinal (GI) wall, endoscopic ultrasound (EUS) was shown to be very useful in the detection and differential diagnosis of intrinsic lesions as well as the differentiation from extrinsic compressions 6m’o.Compared to computed tomography (CT) and other imaging modalities, EUS has the highest sensitivity in the detection of SMLs ‘“. We report our results of EUS in the evaluation of SMLs over a seven-year period. In our study, we address the following questions: Can EUS be used for accurate prediction of tissue type? Are there EUS signs or clinical symptoms which may be useful for prediction of malignancy in a wide variety of unselected SMLs? How to guide further treatment of suspected SML (no further work up, EUS follow-up or surgery) using EUS and clinical criteria?

Patients and Methods Between July 1990 and July 1997, EUS was performed by four experienced investigators (>lOOO EUS examinations each) in 181 patients with endoscopic impression of submucosal turnout-s (Fig. 1). All the procedures were performed on an outpatient basis using iv

Fig. 1. Diagnostic algorithm of present study.

sedation. Olympus GF-UM 3 and GF-UM 20 echoendoscopes (7.5/12 MHz) or Olympus miniprobes UM 2R and UM 3R (12/20 MHz) (Olympus Optical Co., Tokyo, Japan), all providing a 360” image perpendicular to the axis of the echoendoscope were used. Following written consent, detailed EUS was performed to identify and localize the lesion as intramural or extrinsic, to differentiate the aetiology and characterize as benign or malignant. In gastric lesions, the stomach was routinely filled with water, using an endo-water-jet (Pauldrach Medical Inc., Garbsen, Germany). A water tilled balloon or balloon catheter for miniprobes (Olympus MH246R) was used, if required, to improve image quality. Abnormalities were documented by thermoprints or video recording. Tissue sampling was performed optionally, by forceps biopsy or EUS-guided fine-needle-aspiration (EUS-FNA), if the clinical condition of the patient would have allowed surgery to be performed, and the decision for such treatment had not yet been made. The technical details of echoendoscopic FNA have been reported elsewhere ‘I. Prospective evaluation was performed to determine the aetiology of SML in patients with available tissue diagnosis (group 1): lesions were characterized as gastrointestinal stromal tumours (GIST) if they were hypoechoic and seen to arise from the mucosa, suggesting the origin of the lesion to be in the muscularis mucosae (layer 2) or muscularis propria (layer 4) and as non-stromal tumours if located in the tela submucosa. Lipomas were differentiated from other tumours such as fibromas on account of their higher echogenicity due to fat and from stromal tumours on account of their localization within the tela submucosa. Cystic lesions were diagnosed in the presence of an anechoic appearance and distal echo enhancement and were differentiated from vessels by imaging the course of the structures. Other EUS diagnoses were based on characteristic appearances I2 13. Extrinsic lesions (extramural compressions by physiological structures or extramural lesions) were diagnosed from the anatomical location and typical morphology. On retrospective analysis, we compared EUS features as well as clinical presentation in benign vs malignant SML (group 2). Videoendoscopic features always resembled the corresponding EUS morphology (endoscopic estimates and EUS measure, intact surface on endoscopy and regular margins) and, therefore, were not assessed as independent diagnostic criteria. A total of 3 echo criteria (margin, echotexture, size) and four clinical features (pain, bleeding, dysphagia, weight loss) were retrospectively correlated with the biological behaviour of the SML. A final diagnosis of malignancy required cytology or histology, while lesions were classified as benign only on surgical histology 291

EUS for submucosal lesions

(criteria according to Miettinen et al. 14)or a consistent clinical and EUS follow-up of at least one year, showing no further change in size or morphology. Statistic analysis was performed by SPSS@statistic software using Fisher exact test, in addition relative risks (RR) with 95% confidence interval (95% CI) were calculated according to Pearson, Mendel and Haenszel. Results

Of the 181 patients (median age: 56 years, range: lo89: 100 females, 81 males), extrinsic compressions were found either due to an abnormality in 12 (7%), or due to one of the normal adjacent viscera in 35 (19%), whereas 134 had intramural lesions (Table I). The majority of the lesions was located in the stomach (n=116) followed by the oesophagus (n=57) and the duodenum (n=8). In 134 (74%) patients with intrinsic lesions, the muscularis propria (n=93) was the predominant layer of origin. In addition to GIST in 102 patients (76 benign and 26 malignant), a large variety of intrinsic lesions were diagnosed on EUS. Fifty patients were not eligible for analysis due to various reasons such as changes in EUS morphology, not followed by surgery because of advanced age and poor general condition (n=4), no available follow-up observation period of at least 12 months (n=34), death due to unrelated causes (n=7) and loss to follow-up (n=5). In a total of 62 patients, endoscopic sampling of material was performed in addition to EUS. Diagnosic material Tsbla 1. Aetiology of submucosal lesions on endosonography fn=l8’l I. h&Ml

Intramural GIST Benign Malignant

102 76 26

Hypoechoic turnoutof the submucosa

9

Lipoma

6

VesselIhaemangioma

6

Invasive process

4

Heterotopic pancreas

4

Lymphoma

1

cyst

1

Varix treated by Histoacryl GIST: gastrointastinal

292

stromal

1 turnour.

Emamurol

Inz471

Tumours

7

Haemangioma

3

Pancreatic pseudocyst

1

Bronchogenic cyst

1

Spleen

13

Splenic vessel

13

Gallbladder

6

Kidney

1

Pancreatic head

1

Aortic arch

1

was obtained in 17142 cases using forceps biopsy (37%) and in lo/30 cases (33%) using EUS guided FNA. Aetiology (group 1: n=69) Aetiology of the lesions confirmed by surgery (n=49) or biopsy/EUS-FNA (n=20) is shown in Table II. Forty-two of the 44 stromal tumours (96%), 3 of the 4 bronchial carcinoma (75%) and 11 of the 21 miscellaneous histological diagnoses (52%) were correctly identified on EUS. In the diagnosis of GIST, the sensitivity, specificity, PPV, NPV and the diagnostic accuracy were 95%, 72%, 86%, 90% and 87%, respectively. Thirty-nine of the 44 GIST originated from the muscularis propria, 3 from the muscularis mucosae while another 2 seemed to be of extramural origin (both false negative on EUS while a small stalk between the tumour and the muscularis propria was not detected). Seven lesions diagnosed as GIST on EUS had various other diagnoses on histology. Bronchial carcinomas presented as a mediastinal mass causing extrinsic impression or infiltration of the wall layers. One bronchial carcinoma, diagnosed as a malignant GIST on EUS, demonstrated a large, hypotechoic, heterogeneous mass in continuity with the muscularis propria. Of the miscellaneous diagnoses, 11 of the 21 lesions were diagnosed correctly on EUS.

Table II. Differential diagnosis of lesions confirmed by histology ln=691. II-

n

GIST Bronchial CA Granular cell tumour Sarcoidosis Liposarcoma Lymphoma Oesophageal CA Metastasis (Ovarian1 Lipoma Fibroma Brunneroma Accessory spleen Vessel Heterotopic pancreas Subepithelial scar tissue Dysontogenetic cyst Gastric cyst

44 4 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1

Inflammatory pseudotumour Bronchogenic cyst 1 Polyp 1 No abnormality detected 1 GIST

gas.trointesCinal

stromal

tumow.

42 (2 extramural tumoursl 3 II GIST1 2 1 [I heterotopic pancreas1 1 1 1 1 1 1 1 0 [GIST) 0 (GIST1 0 [GISTI Cl kwophageal malignancy] 0 fGl5Tl 0 fGlST1 : (GISTI 0 kubmucosal mass1 0 fcarcinoidl

8. Brand et al.

WI8 III. EUS features and symptoms associated with malignancy fn=861.

crlterlelt H&erogenous echotexture Size > 3 cm Irregular margins Internal structures* Symptoms”” * Calcifications,

Reletluerisk

llirtdooy Be&n a-06 I%) n2eli@mtIts20 Ml

cystic

areas,

17 17 8 7 16 vessels.

““Pain,

f85.01 f85.01 (40.01 f35.01 (80.01 GI bleeding,

21 27 3 11 26 dysphagia,

131.8) (40.91 f4.51 f18.71 139.41

weight loss.

Fisk exact test P

RR IB5% Cl1 7.16 5.41 4.55 0.49 4.19

RR: relative

f2.84-18.051 f2.06-14.171 [2.22-9.321 fO.22-1.091 ~1.72-10.811

risk; 95% Cl: 95%

confidence

< c < > <

0.001 0.001 0.001 0.05 0.001

interval.

Fig. 2. Histology proven benign gastric stromal tumour in continuity with muscularis propria, measuring 1.7 cm, with smooth margins and homogeneous echotexture Uvliniprobe UM 3R, Olympus, 20 MHz).

Fig. 3. Histology proven malignant gastric stromal tumour arising from muscularis propria, measuring 5.2 cm, with irregular margins and heterogeneous echotexture [Echoandoscopa LIM 20, Olympus, 7.5 MHz].

Malignant versus benign lesions (group 2: n=86) Eighty-six patients were eligible for evaluation of lesions confirmed by surgery (n=49), biopsy/EUS-FNA (n=6) or clinical follow-up (median 36 months, range: 12-83) (n=3 1). Twenty patients (23%) had proven malignancy while 66 were found to have benign lesions (77%). On univariate analysis, the established EUS criteria of heterogeneous echotexture, size >3 cm and irregular outline were significantly associated with malignancy (Table III) (Figs. 2,3). There was no malignant tumour without at least one of these EUS criteria. On the other hand, there was considerable overlap of the different criteria in benign and malignant cases (i.e., low specificity of each single sign). Internal structures (cystic spaces, vessel, calcification) or fusion of layer on EUS showed no association with malignancy. In 4 patients, the size of the lesion increased by at least

5 mm measured by EUS on follow-up, but malignancy was found in only one. While relevant symptoms (pain, weight loss, GI bleeding or dysphagia) were frequently found in patients with external compressions due to normal viscera (60% in our study), in the presence of SML they were significant predictors of malignancy (Table III). At least one of the clinical symptoms was present in 7 out of 8 malignant oesophageal or cardiac lesions, while symptoms in malignant lesions located in the stomach were present in 9 out of 12 cases (Table IV). Calculation of sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy for different criteria and combination of EUS morphology with clinical presentation is summarized in Table V A combination of 2 of any EUS criteria or 1 EUS criterion and symptoms showed the best results regarding sensitivity as well as specificity.

293

HIS for submucosal leoinns

Tabk~IY Symptoms in correlation with site and biological behaviour of lesions kt=861.

Tdla U Diagnostic value of EUS criteria and symptoms predicting malignancy.

Irregular margin HeteroganBous echotexture Size r3 cm Symptoms” 1 EUS criterion 1 EUS criterion + symptoms* 2 EUS criteria 2 EUS criteria + symptoms” 3 EUS criteria 3 EUS criteria + symptoms”

40.0 85.0 65.0 80.0 100.0 80.0 80.0 65.0 35.0 30.0

*Pain,Glbleeding,dysphegia.vvlalgbt lass.&bi-evi&ms:

95.5 60.2 59.1 60.6 55.0 75.8 77.3 86.4 93.9 97.0

ml

72.7 44.7 38.6 38.1 37.7 50.0 57.6 59.1 63.6 75.0

84.0 93.8 92.9 90.9 100.0 92.6 92.7 89.1 62.7 82.1

82.8 72.1 65.1 65.1 61.6 76.7 77.9 81.4

80.2 81.4

see list.

Discussion EUS for SML was described more than a decade ago and is now established as the most sensitive diagnostic procedure for detection of SML 6 7 9 lo. Compared to other diagnostic methods available, such as endoscopy, barium series, transabdominal ultrasound and CT, only EUS has the distinct advantage of visualizing the individual wall layers of the bowel wall and to reliably distinguish the lesions arising from one of the layers from extrinsic compression. Zhang and Nian I” compared extracorporeal US, CT and EUS, reporting an accuracy of 22%, 28% and lOO%, to diagnose SML. In the case of extrinsic compressions, when the area of interest can be completely imaged, there is hardly any doubt regarding exclusion of any underlying pathology by using EUS. Our results with 19% (35/181) anatomical structures are in agreement with several other studies showing a wide variety of adjacent structures mimicking SML I2 I5 16.It is important to underline that EUS can effectively relieve these patients from the anxiety of a serious disease, eliminating any further need for other investigations. 294

Pw

While Zhang and Nian ‘O reported no difficulties in the differentiation of extrinsic and intrinsic lesions on EUS, we misdiagnosed 2 GIST (judged as extramural compression), probably due to a small pedunculated continuity with the muscularis propria, which was not visualized on EUS. In addition, in three cases with suspected intramural lesions on EUS, histology showed one accessory spleen and one enlarged paragastric lymph node due to sarcoidosis, giving an accuracy of 94% (65/69) in differentiation of intrinsic and extrinsic pathology. If extrinsic compressions due to normal anatomic structures are included, the accuracy to differentiate intrinsic and extrinsic structures will reach an accuracy of 96%. Type of tissue and EUS diagnosis In 69 cases with histologically proven SML, several aetiologies could be differentiated with a predominance of GIST due to the characteristic echofeatures, as described by several authors 2 ‘* 13. The definition of GIST includes leiomyoma, leiomyosarcoma as well as benign and malignant schwannomas, as it is impossible to differentiate most of these lesions even on histology and immunocytochemistry 17.

B. Brand et al.

The sensitivity and accuracy of diagnosing GIST was satisfactory (95% and 87%, respectively). This high sensitivity underlines the importance of EUS in the selection of lesions requiring surgery. Due to the high potential for malignant transformation (17%-27%) 18-*0of GIST in contrast to all other types of intramural lesions, GIST-like morphology will influence therapeutic decisions independently or in addition to clinical and/or morphologic signs of malignancy. Malignant versus benign lesions Histopathology remains the gold standard of diagnosing the type of tissue as well as the biological behaviour of gastrointestinal disease. However, despite several techniques of tissue sampling, the diagnostic yield in submucosal lesions remains low 3-5*‘. As reported in the literature, EUS-guided FNA for SML is also less satisfactory than for peri-oesophageal lymph nodes, mediastinal masses or pancreatic lesions lo 22-24.While a positive diagnosis for malignancy can be accepted due to very low false positive results, absence of malignancy on FNA needs further confirmation due to the possibility of false negative results. In addition, in contrast to epithelial tumours, pathological diagnosis of malignant disease is less reliable in mesenchymal tumours, which applies not only to biopsies but also to surgical specimens I7 ‘*. Recurrence of tumours after surgical resection and even late metastases in histologically benign GIST have been described repeatedly in the literature, leading to the confusing diagnosis of a “benign metastasizing” leiomyoma *j 26.Especially for stromal tumours, several histopathologic parameters are used to determine the biological behaviour, none of them constituting definitive evidence of malignancy 17. Pathologists use the presence of two or more of the following criteria to diagnose malignancy: size, infiltration of adjacent structures, necrosis, increased nuclear/cytoplasmic ratio, increased mitotic rates and infiltration of the overlying mucosa r7. Interestingly most of these criteria are also used for EUS evaluation of GIST. Of the different echofeatures described in the literature 6 ‘O*‘-jr, heterogeneous echotexture, size ~3 cm and irregular outline were found to be significantly associated with malignancy, in our study. Echogenic foci and cystic spaces are reported to be predictive of malignancy in GIST by some authors 3’-33. However, in our study, not restricted to GIST alone, these features were not associated with malignancy. In addition, growth of tumour on follow-up, suggested as a sign of malignancy, was found only in one malignant GIST, presenting with rapid growth (from 2.0 cm to 3.6 cm) within a 6 months’ period. Thus, assessment of tumour growth appears to be a mandatory, but, however, non-specific criterion on malignancy in patients treated by echoendoscopic follow-up.

The combination of several EUS findings has been shown to improve the accuracy in detecting malignancy. Chak et al. reported on the ability of EUS to differentiate benign and malignant GIST showing accuracy rates between 56% and 77% and a fair to moderate interobserver agreement between five expert endosonographers who retrospectively judged EUS videotapes 3’. The sensitivity, in their study, ranged between 80% and 100% in detecting malignancy, if at least two features were present. Our data suggest, that instead of a combination of EUS criteria alone, clinical presentation may also add relevant information to diagnose malignancy in unselected SMLs. While the use of any single EUS criterion will have the highest sensitivity (lOO%), specificity will remain low (55%). The combination of EUS criteria with the clinical presentation will increase the specificity for prediction of malignancy. In conclusion, suspected SMLs may require resection, follow-up or no further attention. The management decisions are based on clinical presentation, suspected aetiology and likely biological behaviour. Extrinsic compression by normal neighbouring structures comprises a considerable percentage of cases referred for EUS evaluation of SML. In our study, aetiological differentiation by EUS was possible with a 95% sensitivity for diagnosing GIST. Combining the clinical presentation to EUS morphology may add relevant information increasing the specificity in the prediction of malignancy. It has to be concluded that, while EUS is highly sensitive in the detection of SMLs, it has neither 100% accuracy in the differentiation of intramural and extramural pathological structures nor 100% prediction of type of histology. However, all the intrinsic or extrinsic malignant lesions showed at least one morphological criterion of malignancy. Thus, in the absence of clinical or echoendoscopic signs of malignancy, endoscopy in combination with EUS may be safely used as a means of follow-up. In our opinion, surgery is definitely indicated in those patients presenting with 2 or more criteria, as the specificity of prediction of malignancy ranges from 77% (2 criteria) to 94% (3 criteria) in these individuals. If only one criterion is present, specificity of EUS signs reaches only 50%, so that the decision for surgical resection has to be made in context with the possible risks of surgery in the individual patient. In these patients, echoendoscopic follow-up as well as tissue sampling may be indicated, despite the limited yield of the currently available techniques. In the presence of at least one EUS sign of malignancy, clinical symptoms may be used similar to EUS criteria to predict the biological behaviour, However, symptoms such as pain and bleeding will suggest surgery of the lesion anyway. A 299

EUS for submucosal lesions

Fi#. 4. Suggested diagnostic and therapeutic algorithm of suspecte SML. Role of EUS to guide further treatment after echoendoscopi exclusion of extramural compressions by anatomical structures.

possible diagnostic shown in Figure 4.

r---Ltst of abbreviations

and therapeutic

algorithm

is

1

95% Cl: 95% confidence interval: CT computed tomography; EUS: endoscopic sonography; FNA: fine needle aspiration; GIST: gastroin1 testinal stromal tumour; NPV: negative predictive value; PPV: positive predictive value: RR: relative risk; SML: submucosal lesion

References ’ Hedenbro JL, Ekelund M, Wetterberg P. Endoscopic diagnosis of submucosal gastric lesions. The results after routine endoscopy. Surg Endosc 1991;5:20-3. ’ Fenoglio-Preiser CM, Lantz PE, Listrom MB, Davis M, Rilke FO. Gastrointestinal pathology. An atlas and text: Ch. 16. New York: Raven Press; 1989. p. 543-85.

296

3 Shimamoto M, Watanabe N, Tuneika K. Histologic diagnosis of submucosal tumour, with particular reference to newly devised needle biopsy. Prog Dig Endosc 1980;16:6-12. 4 Harasawa K, Asaki S, Nisimura T. Cytological diagnosis of submucosal tumour transendoscopic thorny needle biopsy. Gastroenterol Endosc 1984;26:596. 5 Kaneko E, Kumagai .I, Honda N, Nakamura S, Kino I. Evaluation of the new giant biopsy forceps in the diagnosis of mucosal and submucosal lesions. Endoscopy 1983;15:322-6. h Yasuda K, Nakajima M, Kawai, K. Endoscopic ultrasonographic imaging of submucosal lesions of the upper gastrointestinal tract. Gastrointest Endosc Clin N Am 1992;2:615-24. ’ Caletti GC, Fusaroli P, Bocus P. Endoscopic ultrasonography. Endoscopy 1998;30:198-221. x Wiersema MJ, Kochman ML, Cramer HM, Tao LC, Wiersema LM. Endosonography-guided real time fine needle aspiration biopsy. Gastrointest Endosc 1994;40:700-7. y Tio TL, Tytgat GNJ, den Hartog Jager FCA. Endoscopic ultrasonography for the evaluation of smooth muscle tumours in the upper gastrointestinal tract: an experience with 42 cases. Gastrointest Endosc 1990;36:342-50. lo Zhang QL, Nian WD. Endoscopic ultrasonography diagnosis in submucosal tumours of stomach. Endoscopy 1998;3O(Suppl I):A69-7 1. ” Binmoeller KF, Thul R, Rathod V, Henke P, Brand B, Jabusch HC, et al. Endoscopic ultrasound-guided IS-gauge, tine needle aspiration biopsy of pancreas using a 2.8 mm channel convex array echoendoscope. Gastrointest Endosc 1998;47: 121-7. ‘? Rosch T, Classen M. Gastroenterologic endosonography. Textbook and atlas: Ch. 9. Stuttgart - New York: Georg Thieme Verlag; 1992. p. 94-105. ” Kawamoto K, Yamada Y, Utsonomiya T, Okamura H, Mizuguchi N, Motooka M, et al. Gastrointestinal submucosal tumours. Evaluation with endoscopic US. Radiology 1997;205:733-40. ‘J Miettinen M, Blay JY, Sobin LH. Mesenchymal tumours of the stomach. In: Hamilton SR, Aaltonen LA, editors. Pathology and genetics of the digestive system, WHO classification of tumours. IARC Press, Lyon 2000. p. 62-5. I5 Motoo Y, Okai T, Otha H, Satomura Y, Watanabe H, Yamakava 0, et al. Endoscopic ultrasonography in the diagnosis of extraluminal compressions mimicking gastric submucosal tumours. Endoscopy 1994;26:239-42. I6 Schofl R, Potzi R, Gang1 A. Impression von auljen oder submukiiser Tumour? Beitrag der Endosonographie zur Abklarung. Z Gastroenterol 1992;30:272-5. I7 Suster S. Gastrointestinal stromal tumours. Seminar in Diagnostic Pathology 1996; 13:297-3 13. Ix Appelman HD, Helwig EB. Gastric epithelioid leiomyoma and leiomyosarcoma (leiomyoblastoma). Cancer 1976;38:708-28. I9 Lee JSY, Nascimento AG, Farnell MB, Carney JA, Harmsen, WS, Ilstrup DM. Epithelioid gastric stromal tumours (leiomyoblastoma): a study of fifty-five cases. Surgery 1995;118:653-61. x Pizzimbono CA, Hija E, Wise L. Leiomyoblastoma of the lesser sac: case report and review of literature. Am Surg 1973;39:692-7. ?’ Caletti GC, Bocchi E, Ferrari A, Bonora G, Santini D, Mazzoleni G, et al. Guillotine needle biopsy as a supplement to endosonography in the diagnosis of gastric submucosal tumours. Endoscopy 1991;23:251-4. ?2 Matsui M, Goto H, Niwa Y, Arisawa T, Hirooko Y, Hayakawa T. Preliminary results of fine needle aspiration biopsy histology in upper gastrointestinal submucosal tumours. Endoscopy 1998;30:750-5. *? Giovannini M, Seitz JF, Monges G, Perrier H, Rabbia I. Fineneedle aspiration cytology guided by endoscopic ultrasonography: Results in 141 patients: Endoscopy 1995;27:171-7.

8. Brand et al.

24 Yamao K, Ohashi K, Mizutani S, Furukawa T, Watanabe Y, Nakamura T, et al. Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) for the diagnosis of digestive diseases. Endoscopy 1998;3O(Suppl I):A176-8. Is Abel1 MR, Littler ER. Benign metastasizing uterine leiomyoma. Multiple lymph nodal metastases. Cancer 1975;36:2206-13. lh Ballerini C, Irma N, Ceretti AP, Prestipino F, Bianchi FM, Sparacio F, et al. Gastrointestinal stromal tumours: a benign tumour with hepatic metastasis after 11 years. Tumori 1998:84:78-81. ?’ Boyce GA, Sivak, MV, Rosch T, Classen M, Fleischer DE, Boyce HW et al. Evaluation of submucosal upper gastrointestinal tract lesions by endoscopic ultrasound. Gastrointest Endosc 1991;37:449-54. 2X Ueyama T, Guo, Hashimoto H, Daimaru Y, Enjoji. A clinicopathologic and immunhistochemical study of gastrointestinal stromal tumors. Cancer 1992;69:947-5.5.

I9 Riisch T, Lorenz R, Dancygier H, von Wichert A, Classen M. Endosonographic diagnosis of submucosal upper gastrointestinal tract tumours. Stand J Gastroenterol I992;27: l-8. j@Schumacher B, Frieling T, Ltibke H, Haussinger D. Diagnostic et surveillance de leiomyomes oesophagiens et gastriques par ultrasonographie endoscopique (USE). Acta Endosc 1995;25:425-3 1. j1 Chak A, Canto, MI, Rosch T, Dittler HJ, Hawes RW, Tio TL, et al. Endosonographic differentiation of benign and malignant stroma1 cell tumours. Gastrointest Endosc 1997;45:468-73. iZ Yoshikawa J, Matsumoto Y, Suekawa K, Arimura F. Yamamoto T, Arima T. Endoscopic ultrasonography is useful for diagnosis of gastric submucosal tumor (Abstract). Gastrointest Endosc 1997;45:AB 185. i’ Palazzo L, Landi B, Cellier C, Cuillerier E, Roseau G, Barbier JP. Endosonograhic features predictive of benign and malignant gastrointestinal stromal cell turnours. Gut 2000:46:88-92.

3RDWORLD CHINESE CONGRESS OF DIGESTOLOGY September 23-25,2002 The Third World Chinese Congress of Digestology (WCDD) will be co-sponsored by the World Journal of Gastroenterology (English), World Chinese Journal of Digestology, and Diagnosis and Treatment of Digestive Diseases, September 23-25, 2002 in China. Contact: Lian-Sheng Ma President of WCCD PO. Box 2345, Beijing 100230, China Fax: +89-65891893 l

l

E-mail: [email protected]