ABSTRACTBackground: Gastric GISTs account for more than

half of all gastrointestinal stromal tumors and representless than 5% of all gastric tumors. The peak age forharboring GIST of the stomach is around 60 years and aslight male preponderance is reported. These tumors areidentified by expression of CD117 or CD34 antigen.Symptoms at presentation usually include bleeding, ab-dominal pain or abdominal mass. Endoscopically, theytypically appear as a submucosal mass with or withoutulceration and on CT scans an extragastric mass is usuallyseen. Complete surgical resection provides the only chancefor cure, with only 1-2cm free margins needed. However,local recurrence and/or metastases supervene in almosthalf the patients treated with surgery alone, even whenno gross residual is left. Thereby imatinib mesylate wasadvocated as an adjuvant to surgery, which appears tohave improved disease-free and overall survival.

Aim of the Work: The aim of this work was to assessclinico-pathological features of gastrointestinal stromaltumors (GIST) of the stomach and to appraise the resultsof treatment by surgery in patients treated at the NationalCancer Institute (NCI) of Cairo between January 2002and December 2007.

Patients and Methods: Nineteen patients with histo-logically and immuno-histochemically proven GIST ofthe stomach were treated by surgery at the NCI duringthe 6-year study period. Preoperative assessment includeddetailed history, clinical examination, full laboratory tests,endoscopy, abdominal ultrasound and CT. General medicalassessment included chest X-ray, ECG and echocardio-graphy.

Results: The patients’ age ranged from 26 to 77 yearswith a median of 51 years. Obvious male/female prepon-derance was noticed (68.4% to 31.6%). Tumors werelocated at the upper 1/3 in 42.1%, at the middle 1/3 in31.6% and at the lower 1/3 in 26.3%. The most commonclinical presentation was related to bleeding (hematemesis,melena or anaemia) and was seen in 63.2%. No tumors

Journal of the Egyptian Nat. Cancer Inst., Vol. 20, No. 1, March: 80-89, 2008

Gastrointestinal Stromal Tumors (GIST) of the Stomach:Retrospective Experience with Surgical Resection at the NationalCancer Institute

SHERIF F. NAGUIB, M.D.; ASHRAF S. ZAGHLOUL, M.D. and HAMDY El MARAKBY, M.D.

The Department of Surgical Oncology, National Cancer Institute.

80

were labeled as very low or low risk while there were52.6% intermediate risk and 47.4% high risk. Wedgeresection was carried out in 15.8%, partial gastrectomyin 37.8%, total gastrectomy in 5.2%, extended gastricresection in 21.1% and only biopsy in 5.2%. Lymphadenec-tomy was carried out in 5/19 patients to reveal negativelymph nodes in all five. Complications occurred in 73.7%of patients and only 1 case of early postoperative mortalitywas recorded. Two patients were lost to follow-up. Theremaining 16 patients were followed-up for a periodranging from 6-34 months with a mean of 19.5±5.6 monthsand they were all alive by the end of the study, 10 werefree of disease and 6 showed disease recurrence.

Conclusion: Gastric GIST can present with vagueand non specific clinical picture. Therefore, thoroughclinical and radiological evaluation and preoperativeendoscopy and biopsy are essential to reach the diagnosisand to assess the risk for metastasis. The clinical outcomeof these tumors is influenced by completeness of tumorextirpation while avoiding tumor rupture, and by the tumormalignant potential. Accordingly for tumors with adversefactors, multimodal therapy with adjuvant imatinib or oneof its successors should be considered in order to improveoverall and disease-free survival.

Key Words: Gastric – GIST – Surgical treatment.

INTRODUCTION

The term GIST was first introduced in 1983by Mazur and Clark [1] to describe non-epithelialtumors of the GI tract that are thought to arisefrom the interstitial cells of Cajal (ICC) whichare components of the intestinal autonomicnervous system and that act as pacemakersregulating peristalsis [2]. In 1998, Hirota andcolleagues [3] demonstrated gain-of-functionmutations of the KIT proto-oncogene in the vastmajority of GISTs. This lead to identificationof these tumors by the universal expression(~95%) of the CD117 antigen, part of the KITreceptor. About 60% to 70% of GISTs may showimmuno-positivity for CD34 which is morecommon in the colorectum and esophagus [4].

Correspondence: Dr. Sherif Naguib, 7 B El Bostan St.,Tahrir Square, Cairo, Egypt, ashrafsz@hotmail.com

81

Gastrointestinal stromal tumors (GISTs) arethe most common mesenchymal tumors of theGI tract [4] with an annual incidence in theUnited States ranging between 10-20 per million[5,6]. In Egypt the relative incidence reportedby the National Cancer Institute is 2.5% of allGI tumors and 0.3% of all malignancies [7].These tumors can arise throughout the GI tract,and the stomach is the most common site, ac-counting for 50%-70% of GISTs [5,6,8] butrepresenting <5% of all gastric tumors [9], withthe proximal stomach being involved in abouttwo thirds of the patients [10]. Less frequently,other GI organs may be affected such as thesmall intestine (25%-30%), the rectum (5%),the esophagus (2%) and other abdominal loca-tions (5%) [11,12].

Gastric GISTs have been identified in pa-tients ranging from 8 to 95 years of age. How-ever, the peak age of diagnosis is around 60years in most series, with less than 10% of thetumors found before the age of 40 years. Thereis a slight male predominance in adult patients,whereas most pediatric GISTs arise in girls [8].

Metastases are found at presentation in 10%-47% of cases, the most frequent locations beingthe liver (65%) and omentum (21%) whilelymph nodes are rarely affected (2-6%). Thepattern of metastatic spread is almost entirelyintra-abdominal; hence bones and lungs areaffected in only 6% and 2% respectively [13].

Many GISTs are asymptomatic and they arediscovered accidentally upon imaging or lap-arotomy for other reasons. Symptomatic GISTstend to be large, with an average size of 6cm,versus 2cm for asymptomatic tumors and 1.5cmfor those detected at autopsy [14]. GISTs arisingfrom the stomach most frequently present withGI bleeding and anaemia (54.4%), abdominalpain (16.8%), or a palpable mass; but sponta-neous tumor rupture very rarely takes place(1.7%) [8].

Histologically, most GISTs are composedof spindle-shaped cells (70%). However, someGISTs are dominated by epithelioid cells (10%-20%) or contain a mixture of spindle and epi-thelioid morphologies (10%-20%) [5]. GISTsbehavior ranges from benign to malignant andthese tumors are now classified according totheir size and mitotic count (MC) into one ofthe following categories: very low risk, low

risk, intermediate risk and high risk [15]. Lesions<2cm and those with low mitotic index onlyoccasionally demonstrate metastases [16] whilethose 2-5cm exhibit metastatic behavior in upto 20% of cases [17].

Upon endoscopy, the appearance of a prima-ry GIST is that of a submucosal GI tract lesion,with or without ulceration [18]. Because of theirsubmucosal location, FNAC or core needlebiopsy with endoscopic ultrasound guidance isfrequently required for tissue diagnosis [16]. CTscans (Fig. 1) are significantly helpful to deter-mine the exact extent of the tumor and assistin both staging and grading of the tumor forpre-operative planning. An exophytic patternof growth (Fig. 2), together with the absenceof associated lymphadenopathy help to differ-entiate GISTs from other gastric tumors suchas lymphomas or adenocarcinomas [19].

Surgical resection remains the standard treat-ment of choice for all resectable non-metastatictumors since it provides the only chance forcure at the present time [13,20,21]. A 1-2cmmargin was advocated to achieve adequate re-section [22]. However, more recently, DeMatteoet al. [13] demonstrated that tumor size (and notwide negative microscopic surgical margins) ismore important in determining survival. Positiveresection margins have not be proven to com-promise survival but they may result only in ahigher risk of peritoneal relapse [13,20]. There-fore, the surgeon’s goal should be a completeresection with only negative gross margins.Accordingly, wedge resection has been advo-cated by many investigators for the majority ofgastric GISTs [13,20,21]. In some cases, however,tumor size and location may dictate a moreextensive surgery, including partial or even totalgastrectomy [23,24]. Locoregionally advancedtumors requiring total gastrectomy or extendedresections should be considered for neoadjuvanttreatment with imatinib mesylate to be subse-quently re-evaluated for more limited resection[25].

Laparoscopic resection of GISTs should belimited to tumors less than 2cm [23]. It providesthe advantage of minimal tissue manipulationand effective diagnosis and treatment of obscurecases presenting with acute bleeding [26].

Avoidance of tumor rupture is imperativeto avoid complications such as bleeding or intra-

Gastrointestinal Stromal Tumors (GIST) of the Stomach

Sherif F. Naguib, et al.82

abdominal dissemination of tumor cells withsubsequent high risk of tumor recurrence [27].Therefore, intra-abdominal open biopsy is dis-couraged by most experts because of the riskof tumor spillage [20].

Lymphadenectomy is not routinely requiredsince lymph node involvement is rare with GIST[13]. Nevertheless, LN dissection should be con-sidered for patients with any suspicion of nodalmetastasis [28].

Until the year 2000, chemotherapy and ra-diation treatments had no proven effective rolein the treatment of advanced disease and theonly known ''effective'' treatment for metastaticGISTs was surgery [29]. However, disease re-curred in all patients after a very short disease-free interval and 5 and 10-year survival ratesafter potentially curative surgery alone werereported to be 32-78% and 19-63% respectively[13,30-33].

Imatinib mesylate was recently introducedas an integral component of GIST treatment. Itis a potent and specific inhibitor of the KIT-protein tyrosine-kinase and it has been approvedfor the treatment of KIT (CD 117)-positiveirresectable or metastatic GISTs. However, theuse of adjuvant imatinib after complete resectionof primary GIST is still being evaluated [34].

PATIENTS AND METHODS

Patients:Nineteen patients with primary or recurrent

gastric stromal tumor (GIST) were operatedupon at the department of surgery of the NationalCancer Institute of Cairo in the period betweenJanuary 2002 and December 2007. This retro-spective descriptive clinical study was basedon studying the patients’ data retrieved fromtheir medical records over that 6-year period.

The studied parameters included patients'demographic characteristics such as age, genderand occupation as recorded in patient’s files(Table 1). Clinical data retrieved from patients'records included symptoms, tumor’s location,size and histology (Table 2). Information re-garding type of operation, resection margin andlymph node status (Table 3).

Preoperative assessment included CBC, liverand kidney function tests, chest X-ray and car-diac function assessment. Preoperative diagnosis

was based on endoscopic, ultrasound (US) orCT imaging (Fig. 1) and guided biopsy; whilemetastatic work up included abdominal USand/or CT, bone scan and chest X-ray.

Evaluation:All tumors were reviewed by experienced

pathologists for histological confirmation ofthe diagnosis of GIST and evaluation of themorphological and immuno-histochemical char-acteristics. All included tumors were c-KIT(CD117) or (CD34) positive. Tumor size wasevaluated on fresh specimens. The mitotic ratewas assessed by counting the number of mitosisper 50 high power fields (HPF) in all patients.The tumors were classified according to therisk assessment suggested by Fletcher et al. [15]into one of 4 categories: very low, low, inter-mediate or high risk.

Treatment:Different types of gastric resections were

carried out and they were classified into thefollowing categories; RO was defined as com-plete resection. R1 included complete resectionof accompanying peritoneal seeding, omentalor liver deposits. R2 entailed leaving grossresidual [35]. Imatinib mesylate was administeredto patients having distant metastases at diagno-sis, undergoing an incomplete (R2) resectionor experiencing recurrence after surgical resec-tion.

Evaluation of treatment results:All patients were carefully followed-up clin-

ically at three-monthly intervals and endoscop-ically, by ultrasound and by CT at six-monthlyintervals. The follow-up period was calculatedfrom the date of surgery to the last month offollow-up, tumor recurrence or death. Postop-erative hospital stay, complications, site andtime of recurrence and survival were also noted(Table 4).

RESULTS

Nineteen patients were considered eligiblefor this study. They were 13 males and 6 femalesand their ages ranged from 26 years to 77 years(median: 51 years). Their tumor size rangedfrom 5 to 50 cm with a median size of 18.5 cm.Tumors measured 5-10 cm in 15 patients and10-50 cm in 4 patients. The mitotic counts ofthe tumors were found to be <5/50 HPF in 10cases, 5-10/50 HPF in 6 cases and >10/50 HPF

83

in 3 cases. All included tumors were c-KIT(CD-117) positive (16 cases) or CD-34 positive(3 cases). The histological types of the tumorswere Spindle cell type in 13 cases, Epithelioidtype in 2 cases, mixed type in 2 cases andunclassified in 2 cases. Patients were subdividedinto 4 groups according to Fletcher’s risk as-sessment rules [15]. None were assigned to thevery low or to the low risk groups, 10 caseswere labeled as intermediate and 9 cases fell inthe high risk category (Table 1).

All patients were symptomatic at presenta-tion. The main presenting symptoms were ab-dominal pain in 10 patients, a palpable abdom-inal mass in 7 cases and gastrointestinal bleeding(hematemesis or melena) in 2 cases. In addition,10/19 patients were found anaemic with a He-moglobin of <12g/dl.

Only 14/19 patients were diagnosed as GISTpre-operatively, Three had been previouslyoperated outside the NCI and were admitted forexcision of abdominal recurrence. Five patientswere diagnosed through upper GI endoscopicbiopsy, two because of upper GI bleeding andthree because of persistent anemia. The remain-ing six were diagnosed by CT guided coreneedle biopsy. The undiagnosed casespreoperatively were explored with the suspicionof gastric lymphoma or adenocarcinoma (in 3cases), retroperitoneal sarcoma (in 1 case) andsplenic hilar mass (1 case).

All 19 patients underwent surgery, wherewedge resection of the stomach was carried outin 3 patients (Figs. 2,3), partial gastrectomy in7, total gastrectomy in 1, extended resection in4 (including splenectomy in 2 cases and omen-tectomy in 2 cases). Excision of local recurrencewas carried out in 3 and only exploration andbiopsy was performed in 1 patient who hadextensive liver metastasis. Laparoscopic assistedantrectomy was undertaken in only 1 patientwith a small (5 cm) tumor (Table 2).

On the whole, fourteen patients had R0resection; three had R1 and two had R2 resec-tion. During follow-up, liver metastases occurredin 3 patients undergoing R0 resection, peritonealand liver metastases supervened in one of threepatient undergoing R1 resection; while Localrecurrence occurred in both patients undergoingR2 resection (Fig. 4). Regional lymphadenec-tomy was undertaken in 5 cases because ofgrossly enlarged lymph nodes and all were

proved to be LN–ve on pathological examina-tion. Two patients were found to harbor me-tastases at laparotomy, one in the liver and theother in the greater omentum. In the first patientno resection was done and only biopsy wastaken, while in the other additional omentectomywas carried out together with hemi-gastrectomy.Tumor rupture occurred in 6/19 patients. Amongthese, 4 patients developed recurrence 2 locally,1 in the liver and 1 in both (Fig. 5).

The duration of surgery ranged from 100-290 minutes with a mean operative time of190±65 minutes. Blood transfusion was requiredin 12 patients with a median of 3 units/patient(range: 1-5 units). The median length of hospitalstay was 10 days (range 5-26 days).

Early postoperative complications occurredin 9 patients from whom 1 patient died of fatalmassive pulmonary embolism on postoperativeday 6 and five cases developed transient atelecta-sis and/or pneumonia. Minor anastomotic leak-age occurred in 3 cases while a major gastro-cutaneous fistula developed in 1 case. Theformer three were treated conservatively; whilethe latter was successfully treated surgically.Late postoperative complications consisted ofincisional hernia in 4 cases. Three of those wererepaired with prolene mesh while the fourthpatient refused surgery (Table 2).

Two of our patients were lost to follow-upand we had one early (D6) postoperative mor-tality. The rest (16 patients) were followed-upfor a period ranging from 6-34 months with amean of 19.4±3.6 months. By the end of thestudy, none of the 16 followed patients had diedof their disease and 10/16 (62.5%) were aliveand free of disease. The remaining 6/16 (37.4%)of the followed cases showed treatment failureconsisting of liver metastases in 3 patients,local recurrence in 2 and both peritoneal recur-rence and liver metastases in 1.

Fifteen patients had tumors smaller than 10cm, 3 of them developed recurrence (2 in theliver and 1 locally) while four had tumors largerthan 10 cm and 3 of them showed disease re-currence (1 in the liver, 1 locally and 1 in bothliver and tumor bed) (Fig. 6). Among the 10patients with <5/50 mitoses per HPF, only 1patient (10%) showed disease recurrence in theliver, while among the 6 cases with 5-10 mitoses/HPF, 3 patients (50%) developed recurrencetwo locally and one in the liver and among the

Gastrointestinal Stromal Tumors (GIST) of the Stomach

Sherif F. Naguib, et al.84

3 cases with >10/50 mitoses per HPF, 2 patients(66.7%) developed recurrence one in the liverand the other both on the peritoneal surface andliver (Fig. 7 and Table 2).

Fig. (3): Same tumor after wedge gastric resection withgrossly free margin from Stomach wall.

Fig. (2): Same case of exophytic gastric GIST at laparot-omy.

Fig. (1): Abdominal CT showing exophytic GIST arisingfrom gastric body.

Fig. (6): Recurrence in relation to size of tumor.

12

10

8

6

4

2

0<10cm >10cm

No recurrence

Recurrent cases

Fig. (5): Recurrence in relation to tumor rupture.

12

10

8

6

4

2

0No rupture With rupture

No recurrence

Recurrent cases

Fig. (4): Recurrence in relation to type (R) of gastricresection.

12

10

8

6

4

2

0R0 R1 R2

No recurrence

Recurrent cases

85

DISCUSSION

Gastric GISTs account for more than halfof all gastrointestinal stromal tumors and rep-resent less than 5% of all gastric tumors. Thesetumors are identified by expression of CD117antigen which is a part of KIT receptor. Symp-toms at presentation usually include bleedingand anaemia, abdominal pain and/or abdominalmass. Complete surgical resection provides theonly chance for cure, with only 1-2 cm freemargins needed. Nevertheless, local recurrence

Gastrointestinal Stromal Tumors (GIST) of the Stomach

Table (3): Surgical treatment and outcome in 19 patientswith gastric GIST.

Surgical intervention:Wedge GastrectomyPartial Gastrectomy*Total GastrectomyExtended GastrectomyExcision of local recurrenceExploration & Biopsy

Microscopic margin**:FreeCloseInvolved

* 1/7 patients underwent laparoscopic antrectomy.** Only 18 patients were included because 1 patient was irresec-

table and only wedge biopsy was taken.

371431

1611

Numberof patients

15.836.85.321.115.85.3

88.95.65.6

%

Table (4): Results of treatment in 19 patients with gastricGIST.

Complications:Chest infectionPulmonary embolismAnastomotic leakageIncisional hernia

Recurrence*:SiteLocalLiverBoth

Time (months):0-67-1213-1819-24

* Only 16 patients were included because 1 patient died on D6and 2 patients were lost to follow-up.

5144

231

0213

Numberof patients

26.35.321.121.1

12.518.86.3

012.56.318.8

%

Fig. (7): Recurrence in relation to mitotic count (No ofmitoses/50 HPF).

No recurrence

Recurrent cases

<5/50 >10/50

10

8

6

4

2

05-10

Table (2): Clinico-pathological features of tumors.

Site:Upper1/3Middle 1/3Lower1/3

Size:5-10cm10-50cm

Mitotic count:<5/50 HPF5-10/50 HPF>10/50 HPF

Histology:Spindle cellEpithelioid cellMixed cellUnclassified

Malignant potential:Very lowLowIntermediateHigh

42.131.526.3

78.921.1

47.436.815.8

68.510.510.510.5

0052.647.4

%

865

154

973

13222

00109

Number of patients

HPF: High power field.

Table (1): Demographic features of 19 patients with gastricGIST.

Gender:MaleFemale

Age (years):21-3031-4041-5051-6061-7071-80

68.431.5

5.310.526.331.515.810.5

%

136

125632

Number of patients

Sherif F. Naguib, et al.86

and/or metastases supervene in almost half thepatients treated with surgery alone, even whenno gross residual is left [4,5].

The present study reviewed the clinico-pathological features of gastrointestinal stromaltumors (GIST) of the stomach in 19 patientstreated surgically at the National Cancer Instituteof Cairo during the 6-year period between Jan-uary 2002 and December 2007. We also triedto appraise the results of treatment in relationto patients’ criteria, tumors’ characteristics andextent of surgery and their effect on diseaserecurrence.

Sex distribution among our patients showeda clear male predominance (2.2: 1) comparedto most reported series which found no appre-ciable sex difference in adults [8,10,11]. Thisobservation needs to be studied more thoroughlyin a larger series.

Tumor location was found in the upper 1/3of the stomach in 8 cases (42.1%), the middle1/3 in 6 cases (31.5%) and the lower 1/3 in 5cases (26.3%). These figures correlate withother studies reporting that gastric GIST usuallyaffects the proximal stomach in over two thirdsof cases [10].

All our patients were symptomatic at pre-sentation. This could be explained by the largesize of our tumors ranging from 5-50 cm (me-dian: 18.5 cm). In contrast, most western studiesreported that only 50-70% of patients are symp-tomatic [8,11,36]. Three studies done one at Mayoclinic [36], another at Cleveland [37] and a thirdat Irvine [38] in the USA, reported that theirtumors’ size ranged from (1.5 to 7.0 cm), (0.5to 10.5 cm) and (2.8 to 7.1 cm) respectively. Inthe first study, not a single patient presentedwith symptoms. Kindblom [14] found that symp-tomatic GISTs tend to be larger (average sizeof 6 cm) versus 2 cm for asymptomatic GISTs.In a study published by Miettinen and colleagues[8], 54.4% of patients presented with symptomsrelated to GI bleeding (most commonly anemia)and a smaller fraction of patients (16.8%) pre-sented with upper abdominal pain. Slightlyhigher figures were found in our study where63% of the patients presented with symptomsrelated to bleeding such as hematemesis, melenaor anaemia. This could probably be explainedby more frequent mucosal ulceration caused byour larger tumors.

Upon histological examination of the spec-imens, 68.5% of the tumors, were found to bespindle cell, 10.5% were epithelioid, 10.5%were mixed (spindle and epithelioid) and 10.5%were unclassified. This compares favorablywith the described incidence in other studies.Levy et al., reported that 20-30% of gastricGISTs have epithelioid morphology and thatsome show both elements [39]. In this study, nocases were assigned to the very low or to thelow risk groups, while 52.6% of the cases werelabeled as intermediate risk and 47.4% as highrisk. Our cases showed higher risk when com-pared to other studies dealing with gastric GIST.Berindoague et al. [9] reported the low, interme-diate and high risk groups to account for 44.4%,33.3% and 22.2% respectively. This could alsobe due to the large tumor size at presentationin our cases.

It has been reported that 10%-47% of pa-tients with GIST harbor distant metastases atpresentation [13,35]. In this study, only 2/19patients (10.5%) presented with metastatic dis-ease. This low incidence in association withhigh prevalence of high risk cases needs to befurther explored. Nevertheless, relative meta-static distribution was close to other reports[13,33,35,40] since it involved the liver in 5.3%and the omentum in 5.3%.

Five of our patients underwent nodal dissec-tion but no lymph node metastases were histo-logically found in any of them. Again, thisfinding compares favorably with other studies[13,20,41] where positive nodes were rarely re-ported. Our results consolidate the generalconsensus that lymphadenectomy is warrantedonly when evident nodal involvement is found[20].

Most studies on GIST advocate that vitalstructures such as the stomach should not besacrificed if grossly free margins can beachieved, since the status of microscopic mar-gins does not seem to affect survival [13,42]. Awedge resection of the stomach with tumor-freemargins is satisfactory in most gastric GISTswhile gastrectomy is reserved for tumors in-volving the pylorus or the esophago-gastricjunction [40]. In accordance with this principleof organ preservation, total gastrectomy wasundertaken, in this study in only one case (5.2%)and this was due to the large size of the tumorinvolving approximately the whole stomach.

87

Microscopic margins were infiltrated in 1 caseand close in 1 case. The former patient devel-oped local tumor bed recurrence after 8 monthswhile the latter stayed free for 13 months (tillthe end of the study). Among the remaining 13patients with free surgical margins, only 2 pa-tients developed local recurrence. This matcheswith the finding by Gold and DeMatteo [43] thatthe presence of residual tumor is significantlyrelated to early recurrence and short survival.Similarly, Pierie et al. [42] and DeMatteo et al.[13] reported a significantly longer 5 year sur-vival rate when GISTs were completely removed(42% and 54% respectively, versus 9% forincomplete removal according to the formerstudy).

Treatment failures are known to affect almosthalf of GIST patients treated by surgery alone[4,5] and tend to be found in the liver in 65%,the peritoneal surfaces in 50% and in both inabout 20% [13]. In agreement with these find-ings, tumor recurrence occurred in 6/16 of ourfollowed patients (37.5%). It was in the liverin 50%, local on the peritoneal surface in 33.3%and it occurred synchronously in both in 16.7%.We probably had slightly smaller figures becauseof our smaller number of patients and shorterfollow-up period. On the other hand, this studyinvolved only gastric GISTs which have beenreported to follow a less aggressive coursecompared to small bowel tumors of the samesize [39,44].

Tumor rupture occurred in 6/19 cases.Among these 66.7% developed disease recur-rence (Fig. 5). This finding confirms the recom-mendation of Mochizuki et al. [27] to avoidtumor rupture since it was associated with intra-abdominal dissemination of tumor cells andsubsequent high risk of local tumor recurrence.

Lillemoe et al. [44] found that recurrencewas predicted by tumor size as well as mitoticcount. Comparable findings were found by Boniet al. [45]. Our results are in accordance withthese findings. In the present study, patientswith tumors >10 cm in diameter developeddisease recurrence more frequently than thosewith smaller tumors measuring <10 cm (75%and 20% respectively) (Fig. 6). Also, low num-ber of mitotic count was found to correlate withless risk of recurrence. The recurrence rate forpatients with <5/50, 5-10 and >10/50 mitoses

per HPF, was 10%, 50% and 66.7% respectively(Fig. 7).

In our study macroscopically complete re-section was undertaken in 17/19 cases (89.5%).One patient was found inoperable and was justbiopsied and another patient showed an involvedresection margin. The latter patient developedlocal peritoneal recurrence 8 months postoper-atively in spite of adjuvant treatment with ima-tinib. This is in agreement with findings byBoni et al. [45] that the presence of residualtumor is significantly related to early recurrenceand short survival. The negative effect of mac-roscopic residual tumor is well known sincemany authors [13,42] found a significantly longer5 year survival rate when GISTs were complete-ly removed (42% versus 9%).

The median time to recurrence was 15.8months (range: 8-24 months) which is shorterthan other studies reporting median intervalsto recurrence of 19 and 25 months [13,42]. Thiscan be explained by the relatively large numberof high risk cases in our study.

It is generally agreed that neoadjuvant ima-tinib is not warranted unless the expected de-crease in size by the tumor will noticeably affectsurgery [20]. Following this strategy, none ofour patients received neoadjuvant treatmentwith imatinib since all but one patient weretreated with gastric preservation. The only pa-tient who undertook total gastrectomy had notbeen diagnosed as GIST preoperatively. Post-operatively, Imatinib was given only to patientswho had liver metastases at surgery (10.5%),those who underwent an incomplete resection(10.5%) or those experiencing recurrence aftersurgical resection (31.5%).

Conclusion:The clinical outcome of gastric GIST is

influenced by many factors, the most importantof which are tumor malignant potential andcompleteness of tumor extirpation. Gastrointes-tinal stromal tumors of the stomach can presentby vague and obscure symptoms; thereforeprompt diagnosis by abdominal CT and endos-copy is required to detect small tumors amenableto cure. Radical gastrectomy is not required,but only complete tumor resection without lym-phadenectomy and with only 1-2 cm free mar-gins is sufficient. However, local recurrenceand/or metastases supervene in almost half the

Gastrointestinal Stromal Tumors (GIST) of the Stomach

Sherif F. Naguib, et al.88

patients treated with surgery alone, hence, mul-timodal therapy warrants consideration for tu-mors with adverse factors. It is thought thatlarger prospective randomized studies are need-ed to precisely identify significant prognosticfactors and to clarify whether adjuvant treatmentwith imatinib or one of its more recent succes-sors can improve overall and disease-free sur-vival in high risk GISTs.

REFERENCES

1- Mazur MT, Clark HB, Hashimoto K, Nishida T,Ishiguro S, et al. Gastric stromal tumors: Reappraisalof histogenesis. Am J Surg Path. 1983, 7: 517-9.

2- Graadt van Roggen JF, van Velthuysen ML, Hogen-doorn PC. The histopathological differential diagnosisof gastrointestinal stromal tumors. J Clin Pathol. 2001,54: 96-102.

3- Hirota S, Isozaki K, Moriyama Y, Hashimoto K,Nishida T, et al. Gain of function mutations of c-kitin human gastrointestinal stromal tumors. Science.1998, 279: 577-80.

4- Miettinen M, Majidi M, Lasota J. Pathology anddiagnostic criteria of gastrointestinal stromal tumors(GISTs); a review. Eur J Cancer. 2002, 38 (5): 39-51.

5- Corless C, Fletcher J, Heinrich M. Biology of gas-trointestinal stromal tumors. J Clin Oncol. 2004, 18:3813-25.

6- Miettinen M, Lasota. Gastrointestinal stromal tumors-definition, clinical, histological, immunohistochemicaland molecular genetic features and differential diag-nosis. Virchows Arch. 2001, 438 (1): 1-12.

7- Mokhtar N, Gouda I, Adel I. Digestive system malig-nancies. In Cancer Pathology Registry (2003-2004),p. 67. NCI, Cairo University.

8- Miettinen M, Sobin LH, Lasota J. Gastrointestinalstromal tumors of the stomach: A clinicopathologic,immunohistochemical and molecular genetic studyof 1765 cases with long-term follow-up. Am J SurgPathol. 2005, 29: 52-68.

9- Berindoague R, Targarona EM, Feliu X, Artigas V,Balague C, et al. Laparoscopic resection of clinicallysuspected gastric stromal tumors. Surgical innovation.2006, 13 (4): 231-7.

10- Matthews BD, Joels CS, Kercher KW, Heniford BT.Gastrointestinal stromal tumors of the stomach. Min-erva Chir. 2004, 59: 219-31.

11- Nilsson B, Bümming P, Meis-Kindblom JM, Odén A,Dortok A, et al. Gastrointestinal stromal tumors: Theincidence, prevalence, clinical course and prognosti-cation in the pre-imatinib mesylate era-a population-based study in western Sweden. Cancer. 2005, 103:821-9.

12- Debiec-Rychter M, Wasag B, Stul M, De Wever I,Van Oosterom A, et al. Gastrointestinal stromal tumors

(GISTs) negative for KIT (CD117 antigen) immunore-activity. J Pathol. 2004, 202: 430-8.

13- DeMatteo RP, Lewis JJ, Leung D, Mudan SS, WoodruffJM, Brennan MF. Two hundred gastrointestinal stromaltumors: Recurrence patterns and prognostic factorsfor survival. Ann Surg. 2000, 231: 51-8.

14- Kindblom LG. Gastrointestinal stromal tumors; diag-nosis, epidemiology, prognosis. ASCO Annual Meet-ing, Chicago. 2003.

15- Fletcher CD, Berman JJ, Corless C, Gorstein F, LasotaJ, et al. Diagnosis of gastrointestinal stromal tumors:A consensus approach. Hum Pathol. 2002, 33: 459-65.

16- Pannu HK, Hruban RH, Fishman EK. CT of gastricleiomyosarcoma: Patterns of involvement. AJR AmJ Roentgenol. 1999 Aug, 173 (2): 369-73.

17- Rossi CR, Mocellin S, mencarelli R, Foletto M, PilatiP, et al. Gastrointestinal stromal tumors, from a surgicalto a molecular approach. Int J Cancer. 2003, 107:171-6.

18- Bertagnolli MM. Gastrointestinal stromal tumors in:Maingot Abdominal Operation, 11th ed. Ed. ZinnerMG and Ashley WS. Pub Mac Graw Hill and Co.2007, p 439-51.

19- Rader AE, Avery A, Wait CL, McGreevey LS, FaigelD, Heinrich MC. Fine needle aspiration biopsy diag-nosis of gastrointestinal stromal tumors using mor-phology, immunohistochemistry and mutational anal-ysis of c-kit. Cancer. 2001, 93: 269-75.

20- Blay JY, Bonvalot S, Casali P, Choi H, Debiec-RichterM, et al. GIST consensus meeting panelists: Consensusmeeting for the management of gastrointestinal stromaltumors. Report of the GIST consensus conference of20-21 March 2004, under the hospices of ESMO. AnnOncol. 2005, 16: 566-78.

21- Heinrich MC, Corless CL. Gastric GI stromal tumors(GISTs): The role of surgery in the era of targetedtherapy. J Surg Oncol. 2005, 90: 195-207; discussion207.

22- Matthews BD, Walsh RM, Kercher KW, Sing RF,Pratt BL, et al. Laparoscopic Vs open resection ofgastric stromal tumors. Surg Endoc. 2002, 16: 803-7.

23- Demetri GD, Blanke CD. NCCN Task Force Report.Optimal management of patients with gastrointestinalstromal tumors (GIST): Expansion and update ofNCCN clinical guidelines. J Natl Comp Cancer Net-work. 2004, 2 (Suppl): 1-26.

24- Fujimoto Y, Nakanishi Y, Yoshimura K, Shimoda T,et al. Clinicopathologic study of primary malignantgastrointestinal stromal tumors of the stromach, withspecial reference of prognostic factors: Analysis ofresults in 140 surgically resected patients. GastricCancer. 2003, 6: 39-48.

25- Hohenberger P, Wardelmann E. Surgical considerationsfor gastrointestinal stroma tumor. Chirurg. 2006, 77:33-40.

89

26- Cueto J, Vasqez-Frias JA, Castaneda-Leeder P, Baque-ra-Heredia J, Weber-Sanchez A. Laparoscopic-assistedresection of a bleeding gastrointestinal stromal tumor.JSLS. 1999, 3: 225-8.

27- Mochizuki Y, Kodera Y, Ito S, Yamamura Y, KanemitsuY, et al. Treatment and risk factors for recurrence aftercurative resection of gastrointestinal stromal tumorsof the stomach. World J Surg. 2004, 28: 870-5.

28- Canda AE, Ozsoy Y, Nalbant OA, Sagol O. Gastrointes-tinal stromal tumor of the stomach with lymph nodemetastasis. World J Surg Oncol. 2008, 6: 97-103.

29- Dematteo RP, Heinrich MC, El-Rifai WM, DemetriG. Clinical management of astrointestinal stromaltumors: Before and after STI-571. Hum Pathol. 2002,33: 466-477.

30- Ng EH, Pollock RE, Munsell MF, Atkinson EN,Romsdahl MM. Prognostic factors influencing survivalin gastrointestinal leiomyosarcomas. Implications forsurgical management and staging. Ann Surg. 1992,215: 68-77.

31- Ng EH, Pollock RE, Romsdahl MM. Prognostic im-plications of patterns of failure for gastrointestinalleiomyosarcomas. Cancer. 1992, 69: 1334-1341.

32- Shiu MH, Farr GH, Papachristou DN, Hajdu SI.Myosarcomas of the stomach: Natural history, prog-nostic factors and management. Cancer. 1982, 49:177-187.

33- McGrath PC, Neifield JP, Lawrence WJ, Kay S,Horsley JS, et al. Gastrointestinal sarcomas. Analysisof prognostic factors. Ann Surg. 1987, 206: 706-10.

34- Berman J, O’Leary TH. Gastrointestinal stromal tumorworkshop. Hum Pathol. 2001, 32: 578-82.

35- An JY, Choi MG, Noh JH, Sohn TS, Kang WK, ParkCK, Kim S. Gastric GIST: A single institutional ret-rospective experience with surgical treatment forprimary disease. Eur J Surg Oncol. 2007 Oct, 33 (8):1030-5.

36- Huguet KL, Rush RM Jr, Tessier DJ, Schlinkert RT,Hinder RA, et al. Laparoscopic gastric gastrointestinalstromal tumor resection: The mayo clinic experience.Arch Surg. 2008 Jun, 143 (6): 587-90.

37- Walsh RM, Ponsky J, Brody F, Matthews BD, HenifordBT, et al. Combined endoscopic/laparoscopic intra-gastric resection of gastric stromal tumors. J Gas-trointest Surg. 2003, 7: 386-92.

38- Nguyen NT, Jim J, Nguyen A, Lee J, Chang K, et al.Laparoscopic resection of gastric stromal tumor: Atailored approach. Am Surg. 2003, 69: 946-50.

39- Levy AD, Remonti HE, Thompson WM, Sobin LH,Miettinen M. Gastrointestinal stromal tumors: Radio-logic features with pathologic correlation. Radiograph-ics. 2003, 23: 283-304.

40- Bucher P, Villiger P, Egger JF, Buhler LH, Morel P.Management of gastrointestinal stromal tumors fromdiagnosis to treatment. Swiss Med Wkly. 2004, 134:145-53.

41- Loong HHF. Gastro-intestinal stromal tumors: Areview of current management options. Hong KongMed J. 2007, 13: 61-5.

42- Pierie JP, Choudry U, Muzikansky A, Yeap BY, SoubaWW, et al. The effect of surgery and grade on outcomeof gastrointestinal stromal tumors. Arch Surg. 2001,136: 383-9.

43- Gold JS, DeMatteo RP. Combined surgical and mo-lecular therapy: The gastrointestinal stromal tumormodel. Ann Surg. 2006 August, 244 (2): 176-184.

44- Lillemoe KD, Efron DT. Gastrointestinal stromaltumors. In Current surgical therapy. Cameron JL, GeryL, editor. Mosby Inc, USA. 2001, pp. 112-117.

45- Boni L, Benevento A, Dionigi G, Rovera F, DionigiR. Surgical resection for gastrointestinal stromaltumors (GIST): Experience on 25 patients. World JSurg Oncol. 2005, 3: 78.

Gastrointestinal Stromal Tumors (GIST) of the Stomach