Indian Journal of Cancer Home 

[Download PDF]
Year : 2019  |  Volume : 56  |  Issue : 2  |  Page : 135--143

Analysis of clinicopathological and immunohistochemical parameters and correlation of outcomes in gastrointestinal stromal tumors

Vaibhav Kumar Varshney1, Rakesh Kumar Gupta2, Sundeep Singh Saluja1, Ila Tyagi2, Pramod Kumar Mishra1, Vineeta V Batra2,  
1 Department of Gastrointestinal Surgery, Govind Ballabh Pant Institute of Post Graduate Medical Education and Research, New Delhi, India
2 Department of Pathology, Govind Ballabh Pant Institute of Post Graduate Medical Education and Research, New Delhi, India

Correspondence Address:
Sundeep Singh Saluja
Department of Gastrointestinal Surgery, Govind Ballabh Pant Institute of Post Graduate Medical Education and Research, New Delhi


INTRODUCTION: Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. We analyzed the clinicopathological features, resectability, immunohistochemical markers, and various factors predictive of disease recurrence and survival. MATERIALS AND METHODS: Retrospective analysis of prospectively maintained database of GIST patients managed from 2005 to 2016 was done. Size, site, malignant potential, nuclear pleomorphism, histopathological variety, immunohistochemical markers, type of surgery, and adjuvant imatinib therapy were analyzed. RESULTS: Ninety-two patients with GIST were analyzed. Immunohistochemistry showed positivity for c-kit (82.4%), DOG1 (75%), and PDGFR-α (79%). Among 16 patients with c-kit-negative tumors, 10 patients were positive for DOG1, PDGFR-α, or both. The most common primary site was stomach (44, 47.8%) followed by small bowel (17, 18.5%) and duodenum (14, 15.2%). Of 92 patients, 80 (87%) underwent R0 resection with organ sparing resection in 56 (70%) patients. Seventeen (21.3%) patients showed recurrence at a median follow-up of 6 years. Median and 5-year overall survival (OS) was 36 months (12–120) and 75%, respectively, and 5-year RFS was 81.8%. On univariate analysis, size, mitotic activity, malignant potential, and nuclear pleomorphism were predictors of recurrence. However, on multivariate analysis, only nuclear pleomorphism was significant. CONCLUSIONS: GISTs had a wide spectrum of presentation, and immunohistopathological features with organ sparing resection were conceivable in maximum. Nuclear pleomorphism may be considered as an important variable to predict recurrence in addition to malignant potential of tumors.

How to cite this article:
Varshney VK, Gupta RK, Saluja SS, Tyagi I, Mishra PK, Batra VV. Analysis of clinicopathological and immunohistochemical parameters and correlation of outcomes in gastrointestinal stromal tumors.Indian J Cancer 2019;56:135-143

How to cite this URL:
Varshney VK, Gupta RK, Saluja SS, Tyagi I, Mishra PK, Batra VV. Analysis of clinicopathological and immunohistochemical parameters and correlation of outcomes in gastrointestinal stromal tumors. Indian J Cancer [serial online] 2019 [cited 2021 Jan 27 ];56:135-143
Available from:

Full Text


Per se gastrointestinal stromal tumors (GISTs) are rare tumors with an estimated incidence of 1.5/100,000 persons per year.[1] However, these are the most common mesenchymal tumors in the gastrointestinal tract.[2] They occur most frequently in stomach (60%) and small intestine (25%) as well as rarely in other gastrointestinal regions (esophagus, colon, rectum), and retroperitoneum.[3]

Earlier, c-kit gene mutation was found to be associated with the pathogenesis of GIST.[4] Further progress of immunohistochemistry (IHC) and development of CD117 (antibody against c-kit) made the accurate diagnosis of GISTs much easier. Later, it was recognized that c-kit mutation was found in 85%–90% of the cases; in the remaining 5%–10% of GIST cases, PDGFR-α mutation being noted. Also, CD117 did not show uniform expression in all the c-kit mutation induced tumors, and about 5% of the cases could be missed. Recently, a new specific immunohistochemical marker DOG1 (discovered on GIST) was discovered which when used in combination with CD117 usually resolved this issue.[5]

Surgical excision is the mainstay treatment modality for localized or potentially resectable GIST.[6] With the development of highly effective targeted therapies against GISTs, such as imatinib, the prognosis of these patients has significantly improved, which mandates a precise and accurate diagnosis of GISTs to be of utmost importance, especially for the wild-type (c-KIT and PDGFR-α negative) GISTs.

This study was undertaken in a tertiary care center in North India, to evaluate the clinicopathological features of GIST patients, assessment of resectability, response to imatinib therapy, and evaluation of various factors predictive of disease recurrence and survival.

 Materials and Methods

Data pertaining to patients of GIST, managed from January 2005 to December 2016, were analyzed retrospectively in the Departments of Pathology, and Gastrointestinal Surgery, Govind Ballabh Pant Institute of Post Graduate Medical Education and Research.

All patients with features suggestive of GIST on imaging and histology and underwent surgical resection or biopsy were subjected to a wide panel of immunohistochemical markers.

Inclusion criteria included tumors positive for c-kit or tumors negative for c-kit and for neuronal/smooth muscle markers but with morphological features suggestive of GIST.


The slides and pathology report from describing gross and microscopic features were retrieved. Hematoxylin and eosin (H and E) stained sections were reviewed independently by two pathologists and consensus arrived at.

Malignant potential of the tumors were graded as of low, intermediate, and high based on Fletcher's criteria including size of tumor and number of mitosis per 5 mm2.[7] The site of the tumor and greatest dimensions were noted on radiological imaging and gross examination. The tumors were classified based on the predominant cell type morphology as spindled (>50% spindle cells), mixed (10%–50% spindle cells) and epithelioid (<10% spindle cells) [Figure 1] and [Figure 2]. Specific microscopic features were noted including mitotic figures per 5 mm2, areas of necrosis, spread, and metastasis of tumor.{Figure 1}{Figure 2}


About 5-μm thick sections were cut for IHC. IHC studies were performed using antibodies directed against c-kit (K45, Dako, 1:2,000), desmin (D33, Thermo Scientific, 1:2,000), smooth muscle actin (SMA) (1A4, Thermo Scientific, 1:1,000), CD 34 (Thermo Scientific, ready to use), S100 (Dako, 1:2,000), PDGFR-α (C-20, Santa Cruz, 1:100), and DOG1 (SP-31, Thermo Scientific, 1:100). For IHC staining, streptavidin biotin conjugate immunoperoxidase method was used. Antigen retrieval was performed in a microwave oven using citrate buffer at pH 6.0 for all antibodies. For each batch, appropriate positive controls were taken. For negative controls, primary antibody was omitted. Immunostaining was regarded as positive if >10% of cells were found to be positive. Pattern of immunostaining was noted as cytoplasmic, nuclear, or membranous. IHC scoring for the different markers was done, taking into account both percentage of cells and intensity of staining as <10% positive (negative), 10%–25% positive (weak), 26%–75% positive (moderate), and >75% positive (strong).

Diagnosis of C-kit negative GIST

Cases that were negative for c-kit markers were subjected to DOG-1 and PDGFR-α markers along with other mesenchymal tumor markers, such as desmin, S-100, SMA, and CD-34 [Figure 3]. Diagnosis of GIST was made on the basis of morphological features, such as spindle cells arranged in fascicles, nuclear grooving, presence of mast cells, and exclusion of other mesenchymal tumors by suitable IHC.{Figure 3}

Work up

All patients were evaluated with detailed history and examination. Routine laboratory parameters included hemogram, liver and kidney function tests, and coagulation profile. Upper and lower gastrointestinal endoscopy were performed as per the symptoms. Contrast-enhanced computed tomography (CECT) was performed in all patients to assess the local extent and distant metastasis. Tumors in the duodenum were analyzed as separate site from those in the jejunum and ileum as the surgical approach and prognosis may differ in these patients.

Treatment policy

All patients deemed resectable underwent surgical resection. Patients with metastatic disease and those requiring adjacent organ resection were considered for neoadjuvant Imatinib. Therapy was given for 6 weeks followed by assessment of response to therapy. Patients with high-grade GIST were considered for adjuvant Imatinib therapy for 3 years.


Follow-up was done at every 3 months for first year, every 6 months for next 2 years, and yearly thereafter. At each follow-up, complete examination along with hemogram and ultrasound abdomen were performed. The CECT abdomen was performed at 1-year and then 2-year interval or when clinically indicated.

Parameters recorded

Demographic parameters, site and size of tumors, extent and type of surgery, and perioperative course, histopathological characteristics including malignant potential, IHC details for each patient were reviewed and analyzed.

Recurrences were classified as either local (tumor bed) or distant (liver, peritoneum, systemic) or both. Survival data were obtained from the date of the last out-patient visit or via personal communication with the patients. Deaths occurring up to 60th postoperative day were considered postoperative mortality and were excluded from the survival analysis. Survival has been expressed as a binary categorical variable, the categories being survived and expired.

Statistical analysis

Categorical variables were analyzed by Chi-square test and continuous variables with t-test or Mann–Whitney U-test, where appropriate. Survival probabilities were calculated using Cox-regression and Kaplan–Meier method. Statistical analysis was carried out using SPSS Version 21 (Chicago, IL, USA), and P value of ≤0.05 has been considered as significant.


In total, 92 cases of GIST were diagnosed and managed in the study. There was a wide age distribution range varying from 11 to 80 years with a median of 49 years. A male predominance with a male-to-female ratio of 1.5:1 was noted. The majority of cases were gastric in location (47.8%), followed by jejunoileal (18.5%) and duodenum (15.2%). We encountered two cases of neuroendocrine tumors of ampulla associated with neurofibromatosis. Both the cases were c-Kit positive. Extraintestinal group include liver, mesentery, and retroperitoneum that comprised of 16 patients (17.4%). The clinical presentation according to site of tumor is shown in [Table 1].{Table 1}

 Pathological Findings


The most common histopathological variant was spindle cell type 69/92 (75%). Thirty (32.6%) patients showed high mitotic activity (>5/5 mm2) and 37 cases (40.2%) had necrosis. Size-wise categorization showed that 34 (37%) patients each had tumor size 5–10 and >10 cm. Fifty-four (58.4%) patients had tumors with high malignant potential while 19 (20.7%) patients had intermediate and low malignant potential tumor, respectively. In the 10 c-KIT negative and PDGFR-α positive GISTs, 60% were nongastric, having size >5 cm and without necrosis. Morphologically, 80% were spindle cell with only 30% having high mitotic count IHC showed c-kit positivity in 75/91 (82.4%) cases, whereas CD34, DOG1, and PDGFR-α were positive in 70/87 (80.4%), 48/64 (75%), and 49/62 (79.0%) of the cases, respectively [Table 2]. Out of 75 c-kit positive cases, a concordant positivity for CD34, DOG1, and PDGFR-α was noted in 58 (77.3%), 38 (50.7%), and 39 (52%) cases, respectively. Of the 16 c-kit negative GISTs, 11 cases were positive for CD34 and 10 cases were positive for either DOG1 or PDGFR-α or both.{Table 2}

Management details

Gastric GIST (n = 44): Forty-one patients underwent resection with laparotomy in 38 patients, whereas laparoscopic wedge resection was performed in 3 patients. Only six patients required total, subtotal, or partial gastrectomy, whereas remaining patients could be managed with wide local excision (WLE). Five patients required adjacent organ resection: spleen in two cases, left lobe of liver, transverse colon and distal pancreas with spleen in one each. All except one patient had R0 resection.

Intestinal GIST (n = 31): Out of 14 duodenal GIST,11 underwent R0 section. Palliative surgery was performed in two patients for obstructing and bleeding tumor, whereas one had biopsy only. Pancreas preserving procedure was performed in six patients, whereas five patients required pancreatico-duodenectomy (PD). Adjacent organ resection in form of right hemicolectomy and nephrectomy was required in one patient.

Among 17 jejunoileal GIST, 16 were resected. Segmental resection was performed in all with three patients required adjacent organ resection; sleeve resection of colon, bladder cuff, and ovary and fallopian tube one each.

Extraintestinal sites (n = 16): Only 11 underwent R0 resection. Right trisectionectomy was performed for a GIST arising from the right lobe of liver. GIST arising from mesentery were resected completely in 77.8% patients, whereas only half of the patients with retroperitoneal GIST underwent complete excision [Table 3].{Table 3}

Overall two patients died during postoperative period. The patient with right trisectionectomy died due to liver failure on day 28, whereas patient who underwent palliative resection for retroperitoneal GIST died in postoperative period secondary to sepsis. Overall curative resection was performed in 80/92 (87%) patients.

Adjuvant treatment and follow-up

Among 65 patients with high and intermediate malignant potential, 48 received imatinib therapy. All patients who underwent resection (n = 80) were available for follow-up, which ranged from 12 to 120 months. Seventeen cases showed recurrence during a median follow-up of 6 years, 5 of which recurred locally and 11 had distant metastasis (liver, peritoneum, and brain), whereas 1 had both local and distant recurrence. Six patients expired from disease during follow-up. Two patients with local recurrence underwent re-resection (omental excision n = 1, transverse colectomy n = 1), whereas one patient improved on imatinib therapy. Among patients with liver metastasis, three died between 2 and 2.5 years after surgery, while other four are alive 6 years after resection on adjuvant therapy including the one with peritoneal disease. One patient with liver metastasis underwent radiofrequency ablation. Median and 5-year overall survival (OS) was 36 months (12–120) and 75%, respectively, and 5-year RFS was 81.8%.

Factors predicting recurrence: On univariate analysis, size, mitotic activity, malignant potential, and nuclear pleomorphism affected the recurrence significantly (P = 0.05). However, on multivariate analysis, only nuclear pleomorphism is significantly associated with recurrence [Table 4]. Overall median survival in the study group was 36 months (12–120). Overall survival did not show any significant correlation with any of the factors reviewed [Table 5].{Table 4}{Table 5}


GISTs are the most common mesenchymal gastrointestinal tumors. Their incidence rate is mostly based on symptomatic patients, but the actual reported incidence is much higher when including cases reported in autopsy series.[2] More than 90% of GISTs are sporadic as seen in our series also. GIST arising from mesentery and retroperitoneum generally referred to as “extragastrointestinal stromal tumors (EGISTs) constituted 10%–17.8% of cases in various studies.[8],[9] In our study, EGISTs were encountered in 17.4%, which is similar to the usually reported incidence.

Pathology/Immunohistochemical markers

More than 90% of GISTs harbor activating KIT mutations.[10] Some GISTs which harbour KIT mutations do not show protein expression in the form of positive c-kit immunohistochemical staining, whereas some tumors show positivity for the c-kit marker despite not having the required mutation.[10],[11] Since DOG1 is specifically expressed by GISTs as compared with ubiquitous positivity of c-kit in many normal cells and other tumors, it is a good supplementary marker under such circumstances. In this study, a rate of positivity of 82.4% for c-kit and 75.0% for DOG1 was obtained, which were comparable to other similar studies.[5],[12],[13] In 16 cases, although c-kit was negative, a positive DOG1 helped to confirm the diagnosis. This explains that DOG1 can serve as a good complementary marker to c-kit. However, comparative lower sensitivity of DOG1 necessitates the combined use of both the markers to increase the pick-up rate of diagnosis.

Approximately 17.6% of tumors lacking KIT mutations have activating mutations in a gene encoding a related receptor tyrosine kinase, the platelet-derived growth factor receptor-alpha (PDGFR-α).[14] PDGFR-α mutation leads to the activation of downstream pathways identical to those in KIT-mutant GISTs.[15] Although there is significant functional overlap, KIT and PDGFR-α mutations are mutually exclusive in GISTs. However, PDGFR-α can be expressed in other soft tissue tumors such as synovial sarcoma and leiomyosarcoma and it has a variable expression among GISTs.[16],[17] We found a concordant positivity for c-kit and PDGFR-α in 79.0% of the cases which suggests that even if KIT and PDGFR-α mutations are mutually exclusive, their immunohistochemical expression can show a marked overlap. In our study, 10 c-kit negative cases (62.5%) were positive for PDGFR-α, results which are similar to those obtained by Sui XL et al.[13]

Extent of resection

Surgery with negative microscopic margins is the mainstay treatment for resectable nonmetastatic GISTs.[6],[18] Small tumors are usually managed by WLE. The most important factor to prevent recurrence is to prevent tumor rupture or hemorrhage intraoperatively.[19],[20] We also achieved R0 resection in 87% with organ sparing resection being accomplished in 70% of patients.

The resectability rate of gastric and small intestine was 93.2% and 87.1%, respectively. Most of stomach and small bowel tumor did not require radical resection with good long-term outcomes. Duodenal GIST needs to be considered separately from rest of the small bowel as multiple surgical options were available depending upon tumor extension. PD (35.7%), segmental resection (28.6%), WLE (14.3%), and bypass or palliative resection (7.1%) are the choices available. Jejunoileal GISTs are usually managed by segmental resection with or without adjacent organ resection. In half of the patients with retroperitoneal tumors, adjacent organ resection was required suggesting local infiltrating nature of disease. This may be due to larger size and late presentation. Hence, R0 resection is the foremost attempt while performing surgery with or without adjacent organ resection.

Adjuvant therapy

Earlier, surgery was the main modality of treatment with an uncertain role of chemotherapy and radiotherapy.[21] However, even after complete resection, the 5-year OS remains 50%–65% in various studies.[22],[23] Similarly, recurrence was noticed in up to 50% of patients after curative resection within a median span of 2 years.[22],[23],[24],[25] Identification of KIT mutations and the development of specific targeted therapy, Imatinib mesylate a novel tyrosine kinase inhibitor (TKI) for the treatment of GIST, represents a major breakthrough in the outcome. Imatinib therapy is recommended for the tumor larger than 3 cm, especially the tumor adjacent to vital organs and at high risk of recurrence.[26],[27] In our study, with adjuvant imatinib treatment, nearly 82% of patients responded in comparison to no-treatment group, where 28% had recurrence, incidences which are similar to available literature.[26],[27]


The prediction of prognosis for GISTs is difficult on histopathology ground alone. The proposed prognostic factor for survival includes tumor size, mitotic index, tumor location, tumor rupture, and kinase mutational status.[19],[20],[22],[28] In our analysis, we found increasing size, high mitotic rate, malignant potential, and nuclear pleomorphism were associated with recurrence-free survival (RFS).

Decrease in RFS has been directly linked with increasing tumor size similar to our results.[22],[29] Similarly, mitotic rate >5 per 5 mm2 portends a high risk behavior.[5] However, variation in interpretation of mitosis may be lead to lack of reproducibility for mitotic count.[30] Hence, its role in isolation has been doubted.

Malignant potential that included tumor size and mitotic count showed significant correlation with RFS. This suggests that considering both factors simultaneously is a more reliable risk stratification strategy. National Institute of Health (NIH) consensus classification also found tumor size and mitotic count together as a better prognostic factor for resectable GISTs.[7]

Nuclear atypia has also reported to be linked with poor prognosis in patients with GIST.[26],[31],[32] We also encountered nuclear pleomorphism as the factor that was significantly associated with RFS on multivariate analysis. The presence of nuclear pleomorphism and nuclear inclusions were more prominent in malignant, metastatic, and recurrent tumors as reported by M Vij in their experience.[33] Similarly, our study also linked nuclear pleomorphism to RFS, which can act as additional prognostic marker. However, more patients are needed to conclusively link nuclear pleomorphism with recurrence.

Overall survival was not found to be dependent on any single factor evaluated, such as size of tumor, mitotic rate, IHC markers, and whether imatinib therapy was received or not. It may be explained by the fact that most GISTs are a low-grade malignancy with a high 5-year survival rate.

The retrospective nature of study, lack of mutational analysis for c-kit, and PDGFR-α and genetic work up due to constraint of resources are the limitations of the study. However, adequate sample size, sufficient disease spectrum, dedicated IHC study, and long follow-up makes this study comparable.


GISTs are indolent tumors with good outcome with organ sparing resection. A combined panel utilizing IHC for c-kit, DOG1, and PDGFR-α helps in diagnosing majority of patients but does not predict recurrence. Nuclear atypia and pleomorphism are important variables to predict recurrence in addition to malignant potential. They can be considered as parameters for starting tyrosine kinase inhibitor therapy in addition to high malignant potential. RFS and prognosis should be incorporated in the diagnosis, wherever feasible.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Nilsson B, Bümming P, Meis-Kindblom JM, Odén A, Dortok A, Gustavsson B, et al. Gastrointestinal stromal tumors: The incidence, prevalence, clinical course, and prognostication in the preimatinib mesylate era--a population-based study in western Sweden. Cancer 2005;103:821-9.
2Agaimy A, Wünsch PH, Hofstaedter F, Blaszyk H, Rümmele P, Gaumann A, et al. Minute gastric sclerosing stromal tumors (GIST tumorlets) are common in adults and frequently show c-KIT mutations. Am J Surg Pathol 2007;31:113-20.
3Corless CL, Fletcher JA, Heinrich MC. Biology of gastrointestinal stromal tumors. J Clin Oncol 2004;22:3813-25.
4Hirota S, Isozaki K, Moriyama Y, Hashimoto K, Nishida T, Ishiguro S, et al. Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science 1998;279:577-80.
5Miettinen M, Wang ZF, Lasota J. DOG1 antibody in the differential diagnosis of gastrointestinal stromal tumors: A study of 1840 cases. Am J Surg Pathol 2009;33:1401-8.
6Lai EC, Lau SH, Lau WY. Current management of gastrointestinal stromal tumors—a comprehensive review. Int J Surg 2012;10:334-40.
7Fletcher CD, Berman JJ, Corless C, Gorstein F, Lasota J, Longley BJ, et al. Diagnosis of gastrointestinal stromal tumors: A consensus approach. Int J Surg Pathol 2002;10:81-9.
8Iqbal N, Sharma A, Shukla N, Mohanti BK, Deo SV, Sahni P, et al. Advanced gastrointestinal stromal tumors: 10-years experience from a tertiary care centre. Trop Gastroenterol 2015;36:168-73
9Reith JD, Goldblum JR, Lyles RH, Weiss SW. Extragastrointestinal (soft tissue) stromal tumors: An analysis of 48 cases with emphasis on histologic predictors of outcome. Mod Pathol 2000;13:577-85.
10Rubin BP, Singer S, Tsao C, Duensing A, Lux ML, Ruiz R, et al. KIT activation is a ubiquitous feature of gastrointestinal stromal tumors. Cancer Res 2001;61:8118-21.
11Emile JF, Théou N, Tabone S, Cortez A, Terrier P, Chaumette MT, et al. Clinicopathologic, phenotypic, and genotypic characteristics of gastrointestinal mesenchymal tumors. Clin Gastroenterol Hepatol 2004;2:597-605.
12Sözütek D, Yanık S, Akkoca AN, Sözütek A, Özdemir ZT, Avşar CU, et al. Diagnostic and prognostic roles of DOG1 and Ki-67, in GIST patients with localized or advanced/metastatic disease. Int J Clin Exp Med 2014;7:1914-22.
13Sui XL, Wang H, Sun XW. Expression of DOG1, CD117 and PDGFRA in gastrointestinal stromal tumors and correlations with clinicopathology. Asian Pac J Cancer Prev 2012;13:1389-93.
14Corless CL. Gastrointestinal stromal tumors: What do we know now? Mod Pathol 2014;27:S1-16.
15Kang HJ, Nam SW, Kim H, Rhee H, Kim NG, Kim H, et al. Correlation of KIT and platelet-derived growth factor receptor alpha mutations with gene activation and expression profiles in gastrointestinal stromal tumors. Oncogene 2005;24:1066-74.
16Pauls K, Merkelbach-Bruse S, Thal D, Büttner R, Wardelmann E. PDGFR alpha- and c-kit-mutated gastrointestinal stromal tumours (GISTs) are characterized by distinctive histological and immunohistochemical features. Histopathology 2005;46:166-75.
17Peterson MR, Piao Z, Weidner N, Yi ES. Strong PDGFRA positivity is seen in GISTs but not in other intra-abdominal mesenchymal tumors: Immunohistochemical and mutational analyses. Appl Immunohistochem Mol Morphol 2006;14:390-6.
18Zhi X, Jiang B, Yu J, Røe OD, Qin J, Ni Q, et al. Prognostic role of microscopically positive margins for primary gastrointestinal stromal tumors: A systematic review and meta-analysis. Sci Rep 2016;6:21541.
19Rutkowski P, Bylina E, Wozniak A, Nowecki ZI, Osuch C, Matlok M, et al. Validation of the Joensuu risk criteria for primary resectable gastrointestinal stromal tumour-the impact of tumour rupture on patient outcomes. Eur J Surg Oncol 2011;37:890-6.
20Hohenberger P, Ronellenfitsch U, Oladeji O, Pink D, Ströbel P, Wardelmann E, et al. Pattern of recurrence in patients with ruptured primary gastrointestinal stromal tumour. Br J Surg 2010;97:1854-9.
21Joensuu H, Fletcher C, Dimitrijevic S, Silberman S, Roberts P, Demetri G. Management of malignant gastrointestinal stromal tumours. Lancet Oncol 2002;3:655-64.
22De Matteo RP, Lewis JJ, Leung D, Mudan SS, Woodruff JM, Brennan MF. Two hundred gastrointestinal stromal tumors: Recurrence patterns and prognostic factors for survival. Ann Surg 2000;231:51-8.
23Krajinovic K, Germer CT, Agaimy A, Wünsch PH, Isbert C. Outcome after resection of one hundred gastrointestinal stromal tumors. Dig Surg 2010;27:313-9.
24Jiang WZ, Guan GX, Lu HS, Yang YH, Kang DY, Huang HG. Adjuvant imatinib treatment after R0 resection for patients with high-risk gastrointestinal stromal tumors: A median follow-up of 44 months. J Surg Oncol 2011;104:760-4.
25Rutkowski P, Nowecki ZI, Michej W, Debiec-Rychter M, Woźniak A, Limon J, et al. Risk criteria and prognostic factors for predicting recurrences after resection of primary gastrointestinal stromal tumor. Ann Surg Oncol 2007;14:2018-27.
26DeMatteo RP, Ballman KV, Antonescu CR, Corless C, Kolesnikova V, von Mehren M, et al. Long-term results of adjuvant imatinib mesylate in localized, high-risk, primary gastrointestinal stromal tumor: ACOSOG Z9000 (Alliance) intergroup phase 2 trial. Ann Surg 2013;258:422-9.
27Dematteo RP, Ballman KV, Antonescu CR, Maki RG, Pisters PW, Demetri GD, et al. Adjuvant imatinib mesylate after resection of localised, primary gastrointestinal stromal tumour: A randomised, double-blind, placebo-controlled trial. Lancet 2009;373:1097-104.
28Ueyama T, Guo KJ, Hashimoto H, Daimaru Y, Enjoji M. A clinicopathologic and immunohistochemical study of gastrointestinal stromal tumors. Cancer 1992;69:947-55.
29Singer S, Rubin BP, Lux ML, Chen CJ, Demetri GD, Fletcher CD, et al. Prognostic value of KIT mutation type, mitotic activity, and histologic subtype in gastrointestinal stromal tumors. J Clin Oncol 2002;20:3898-905.
30Baak JPA. Mitosis counting in tumors. Hum Pathol 1990;21:683-5.
31Newman PL, Wadden C, Fletcher CDM. Gastrointestinal stromal tumors: Correlation of immunophenotype with clinicopathological features. J Pathol 1991;164:107-17.
32Ng EH, Pollock RE, Munsell MF, Atkinson EN, Romsdahl MM. Prognostic factors influencing survival in gastrointestinal leiomyosarcomas. Implications for surgical management and staging. Ann Surg 1992;215:68-77.
33Vij M, Agrawal V, Kumar A, Pandey R. Cytomorphology of gastrointestinal stromal tumors and extra-gastrointestinal stromal tumors: A comprehensive morphologic study. J Cytol 2013;30:8-12.