Coexistence of gastrointestinal stromal tumors and gastric adenocarcinomas

RESEARCH ARTICLE

Coexistence of gastrointestinal stromal tumorsand gastric adenocarcinomas

Yan Yan & Ziyu Li & Yiqiang Liu & Lianhai Zhang &

Jiyou Li & Jiafu Ji

Received: 3 October 2012 /Accepted: 11 December 2012 /Published online: 3 January 2013# International Society of Oncology and BioMarkers (ISOBM) 2012

Abstract The purpose of this study is to detect the clinico-pathology of gastrointestinal stromal tumors (GISTs) occur-ring synchronously with gastric adenocarcinomas and tounveil the potential underlying relationship between the syn-chronous GIST and gastric adenocarcinoma. This study in-cluded 15 patients with incidental GISTs found duringoperations for gastric adenocarcinoma and 30 patients whounderwent gastrectomy for gastric cancer without discoveringGIST between January 2005 and December 2010 at theBeijing Cancer Institute. We collected the clinicopathologicaldata and analyzed the KIT/PDGFRA mutational status ofGISTs, corresponding gastric adenocarcinoma specimens,and the normal tissue around the cancer lesions.Additionally, as a control group, the mutational status of thepatients with gastric adenocarcinoma and no other tumors wasassayed. Overall, 18 GISTs were found in 15 gastric adeno-carcinoma patients. Multiple GIST lesions were found in threecases (20 %). The patients’ age ranged from 46 to 85 years,with an average of 67.6 years. The average size of the GISTswas 0.85 cm. All mesenchymal lesions showed low

proliferative activity, were of low or very low risk, and wereidentified as CD117-positive by immunostaining. In GISTlesions, mutations in KIT were detected in 7 out of 13 cases,and of these mutations, 6 were found in exon 11 (46.2 %), and1 was found in exon 9 (7.7 %). A total of five deletions andone point mutation were in exon 11, and one insertion was inexon 9. Mutations were not detected in exon 17 or 13 of KIT.There was no remarkable mutation analyzed in the gastricadenocarcinoma lesions or normal tissues from either the testor control groups. Clinicopathological profiles and molecularanalysis of KIT/PDGFRA showed no obvious relationshipbetween gastric cancer and GISTs in tumor genesis, such assimilar oncogene mutations.

Keywords Gastrointestinal stromal tumors . Synchronousneoplasms .KIT . PDGFRA

Introduction

Carcinogenesis is caused by mutations of the genetic materialof normal cells, which upsets the balance between cell prolif-eration and cell death. In recent years, the mutations of genesthat play vital roles in signal transduction pathways have beenthe focus of much interest and debate. Thus, therapies thattarget these genes attract lots of attention. Gastrointestinalstromal tumors (GISTs) are mesenchymal tumors driven byKIT signaling and are considered a paradigm of molecularlytargeted solid tumors. The type and frequency of KIT/PDGFRA mutations are thought to relate to the biologicalbehavior of GISTs, and the mutational status largely deter-mines the clinical response to imatinib treatment [1–4].

Gastric adenocarcinoma (GA) is still the second most com-mon cancer in China with an incidence of 37.88/100,000 [5],and the prognosis of advanced gastric cancer remains poor.According to a recent report published by the Ministry ofHealth and National Cancer Center of the People’s Republic

Electronic supplementary material The online version of this article(doi:10.1007/s13277-012-0627-5) contains supplementary material,which is available to authorized users.

Yan Yan, Ziyu Li, and Yiqiang Liu contributed equally to this study.

Y. Yan : Z. Li : L. Zhang : J. Ji (*)Key Laboratory of Carcinogenesis and Translational Research(Ministry of Education), Department of Surgery, Beijing CancerHospital and Institute, Peking University School of Oncology,Beijing, Chinae-mail: [email protected]

Z. Lie-mail: [email protected]

Y. Liu : J. LiKey Laboratory of Carcinogenesis and Translational Research(Ministry of Education), Department of Pathology, Beijing CancerHospital and Institute, Peking University School of Oncology,Beijing, China

Tumor Biol. (2013) 34:919–927DOI 10.1007/s13277-012-0627-5

http://dx.doi.org/10.1007/s13277-012-0627-5

of China in 2012, the death rate is 22.64/100,000 in 2008 [5].Several genetic and molecular abnormalities have been foundplaying a role in the carcinogenesis of GAs, including activa-tion of oncogenes, overexpression of growth factors/receptors,inactivation of tumor suppression genes, DNA repair genesand cell adhesion molecules, point mutations of tumor sup-pressor genes, etc. Some have caused intensive attention theseyears, such asCDH1mutation, human epithelial growth factorreceptor 2 gene amplification. Therapies targeting these ge-netic defects are used for selected patients. Also, the drugtargeting vascular endothelial growth factor receptor gets re-sponse in some cases [6]. Because of heterogeneity and com-plicacy of the cancer, the overall efficacy is still not good asresearchers expected. We encountered several cases ofincidental GISTs that were found during operations forgastric adenocarcinoma in our practical experience, andwe tried to reveal more information about GA fromthese uncommon cases.

Historically, the association between specific tumors of-ten leads to the discovery of novel molecular pathways oftumorigenesis or genetic defects. Also, the coexistence ofdifferent tumors raised a new problem on treatment, includ-ing surgery and medical principles, especially in the era ofpersonalized medicine.

The high incidence of simultaneous occurrence of epi-thelial and stromal gastrointestinal tumors raised the ques-tion of whether there is a simple incidental association orwhether the two types of lesions are connected by a causalrelationship. The mutation of KIT and the platelet-derivedgrowth factor receptor-alpha (PDGFRA) genes, which aremutually exclusive of each other, has been discovered inmore than 95 % of clinical GISTs. Moreover, it is widelyaccepted that mutations in these genes stimulate down-stream signaling, which activates cytogenetic changes asso-ciated with tumor progression [7]. Candidate genes areconsidered to be genes whose mutation status plays a clearand simple role in carcinogenesis, so we have selected KITand PDGFRA for further study.

We collected cases with synchronous GISTs and GA. Wereviewed the clinicopathological profiles; analyzed the KIT/PDGFRA mutational status of GISTs, corresponding GA,and peritumoral tissue specimens; and set a random controlgroup with GA only. To our knowledge, this studycontained the largest amount of cases with synchronizedGISTs and gastric adenocarcinomas.

Materials and methods

Patients and specimens

This study consisted of two groups of randomly selectedpatients with primary gastric adenocarcinoma who underwent

surgery between January 2005 to December 2010 at theBeijing Cancer Institute. The first group (group 1) included15 patients with GISTs that were incidentally discoveredduring an operation. To determine whether GA tissue withoutthe presence of GISTs harbored any KIT/PDGFRAmutations,we selected another group of 30 patients (group 2) whounderwent gastrectomy for gastric cancer without discoveringGISTs. All the diagnoses were confirmed by histology at theDepartment of Pathology at the Beijing Cancer Institute. Allthe specimens were collected separately from the cancerlesions and surrounding normal tissues. For patients whohad GISTs, specimens of the GIST lesions were also included.Hospital charts and operative and pathological reports werereviewed for each patient.

In group 1, all patients were adults (ages 46 to 85 years,with an average age of 66.6 years) with no clinicopathologicsigns of NF1 and no familial history for GIST, NF1, orCarney’s syndromes. One of the patients had a familialhistory of GA (case 9). Most of the patients were men(66.7 %, 10 of 15 cases), and most of the primary tumorswere located in the proximal (53 %) and distal stomachregions (33 %).

In the same period, 2,438 patients underwent gastrecto-my for gastric cancer, and 17 GISTs were found incidentallyduring the procedure. Also, we encountered 270 overtGISTs of the gastrointestinal tract without the presence ofother malignancies.

Methods

Gastrectomy was performed to all the patients, and thegastric adenocarcinoma and GISTs were resected at thesame time. All of the specimens were processed usingroutine hematoxylin and eosin stain and immunohistochem-istry after being fixed in 10 % buffered formalin and em-bedded in paraffin. For mesenchymal lesions that weremorphologically suspicious for GIST, mitoses were countedin 50 high-power fields. Malignant potential was stratifiedaccording to the risk categories proposed by Fletcher et al.[8]. Most of the mesenchymal lesions were stained using apanel of immunohistochemical markers, including smoothmuscle actin (1A4, 1:200, Dako), desmin (D33; 1:250,Dako), S100 (polyclonal;1:2500, Dako), CD117 (antihumankit proto-oncogene product, polyclonal; 1:200, Dako), andCD34 (BI-3C5; 1:200, Zytomed). For GA, stage informa-tion was created according to the seventh TNM stagingsystem of the American Joint Committee on Cancer/International Union Against Cancer (AJCC/UICC).

Mutation analysis of KIT and PDGFRA

Isolation of genomic DNA from tumor tissue sections Afterstandard formalin-fixation and paraffin-embedding procedures,

920 Tumor Biol. (2013) 34:919–927

tumor tissues were cut in sections of 5-μm thickness. Thefollowing procedure for genomic DNA extraction was per-formed according to standard manufacturer’s directions givenfor the QIAamp DNA FFPE Tissue Kit.

Polymerase chain reaction Polymerase chain reaction(PCR) primers were designed by Primer Premier 5 andverified by Oligo Software (Applied Biosystems) for bothKIT (exons 9, 11, 13, and 17) and PDGFRA (exons 12 and18) according to previous reports [9]. The primers weresynthesized by Invitrogen (Guangzhou, China). For PCRsystem and PCR parameters, please see Supplement Table 1.

The PCR products were separated by electrophoresis on a2 % agarose gel. After verification, the products were puri-fied for sequencing, and the sequencing results wereBLASTed to identify mutations using Survey softwareaccording to the gene sequence available in Genbank (theNIH genetic sequence database).

Results

Overall, in group 1, 18 GISTs were found in 15 GA patients,and multiple GIST lesions were found in 3 cases (20 %).

The average size of the GISTs was 0.85 cm. Except for oneGIST with a diameter of 2.5 cm, the lesions ranged from0.2–2 cm in size. Male patients constituted 66.7 % (10 of 15cases) of the patients in group 1.

All mesenchymal lesions were grossly recognizable, andall except the lesion from case 6, which was defined as lowrisk, were defined as very low risk [10]. The correspondingadenocarcinoma lesions were distributed in all parts of thestomach; however, the upper-third of the stomach harbored53 % (8/15) of the lesions. Most of the GAs were consideredin late stage, but three cases were in the very early stage(T1bN0M0). There seems to be no obvious inclinationbetween different stages of gastric adenocarcinomas.

On gross examination, the GISTs were whitish, smooth,and firm intramural nodules that were slightly elevatedabove the stomach wall. On microscopic examination, allGISTs were composed of spindle cells, similar to thoseobserved in conventional GISTs. Some cases showed centraldystrophic calcifications. Cellular atypia or mitotic activitywas not observed in any of these lesions. Twelve cases(material was not available for immunostaining in cases 1,3, and 4) were positive for CD117 and CD34. All GISTsshowed low proliferative activity, and all GISTs had lessthan five mitoses per 50 high power fields. Based on tumorsize and mitotic activity, the GISTs were categorized as very

Table 1 Clinical and pathologic features of cases in group 1

Number Gender/age Adenocarcinoma GIST lesions

Histology Stage Location Size(cm)

Mitosis/50HPF

Grade of risk CD34 CD117

1 F/53 Poorly differentiated T4N0M0 U 0.4 <5 Very low risk N/A N/A

2 M/51 Moderately to poorly differentiated T4N3M0 U 0.8 <5 Very low risk P P

3 M/62 Poorly differentiated T4N2M0 U 0.8 <5 Very low risk N/A N/A

4 F/73 Poorly differentiated and SRC T4N3MX L 0.2 <5 Very low risk N/A N/A

0.2 <5 Very low risk N/A N/A

5 M/68 Moderately differentiated T1bN0M0 M 0.8 <5 Very low risk P P

6 F/46 Poorly differentiated and SRC T1bN0M0 L 2.5 <5 Low risk P P

7 M/78 Moderately differentiated T4N1M0 U 1.5 <5 Very low risk P P

8 M/66 Moderately differentiated T4N3M0 U 1.5 <5 Very low risk P P

9 M/85 Moderately to poorly differentiated T4N2M0 U/M 1 <5 Very low risk P P

0.5 <5 Very low risk P P

10 M/68 Poorly differentiated T4N0M0 L 0.8 <5 Very low risk P P

11 M/69 Moderately differentiated T4N0M0 U 2 <5 Very low risk P P

12 M/77 Moderately to poorly differentiatedand SRC

T2N1M0 U 0.2 <5 Very low risk P P

0.4 <5 Very low risk P P

13 M/71 Moderately and mucinousadenocarcinoma

T4N3M0 L 0.6 <5 Very low risk P P

14 F/77 Moderately to poorly differentiated T4N3M0 L 0.5 <5 Very low risk P P

15 F/70 Moderately to poorly differentiated T1bN0M0 L 0.6 <5 Very low risk P P

SRC signet ring cell carcinoma, U upper third of the stomach, M middle third of the stomach, L lower third of the stomach, N/A not available forimmunohistologic stain, P positive in immunohistologic stain

Tumor Biol. (2013) 34:919–927 921

low risk or low risk. The clinicopathologic features of group1 (cases with synchronous GA and GISTs) are summarizedin Table 1 and Fig. 1. Nearly all the types of histologicalcharacteristics of primary GA were observed in group 1.

Molecular findings

For this analysis, our study contained two groups of sam-ples. Group 1 included samples from patients with dualtumors, and group 2 consisted of patients with only adeno-carcinoma of the stomach. DNAwas isolated from 13 casesin group 1 for subsequent molecular analyses (cases 7 and14 were excluded). The results of mutation analysis arelisted in Table 2. The mutation of codon 567 CCA→CCGin exon 12 has been previously indicated to possess apolymorphic mutation [9], so the remarkable amount ofcases carrying this mutation was neither surprising nor

meaningful. Similarly, we observed synonymous mutationof exon 18: codon 824 GTC→GTT [9]. Thus, no remark-able mutation was observed in gastric adenocarcinomalesions or normal tissues. In GIST lesions, mutations inKIT were detected in 7 cases (out of 13), 6 of which werefound in exon 11 (46.2 %), and 1 of which was found inexon 9 (8.0 %). There were five deletions and one pointmutation in exon 11, and one insertion was in exon 9. Nomutations were detected in exons 17 and 13 of KIT.Molecular analysis was also performed for 30 cases in thecontrol group. The results of this analysis are listed inTable 3. No remarkable mutations were found in this group.

Patients with advanced GA were treated with adjuvantfolinic acid, fluorouracil, and oxaliplatin (FOLFOX) che-motherapy perioperatively without imatinib therapy forGISTs. In the surveillance, no GIST-related adverse eventswere observed in these 15 patients of group 1.

Fig. 1 Gross, macroscopic,histopathological, andimmunohistochemicalcharacteristics of representativeGIST lesions with synchronousgastric adenocarcinomas. Grossspecimens of the stomach withcoexisting tumors: the cancerlesion was in the antrum, whichcould be observed clearly fromthe inner side of the stomach(a), and GIST was just beneaththe serosa of the stomach (b).Macroscopic appearance of thegastric stromal tumor. MostGISTs presented as subserosalplaque-like nodules (c).Microscopic visualization of thetypical spindle cell morphologyof the GIST (d). Strongimmunostaining was observedwith CD117 (e) and CD34 (f)

922 Tumor Biol. (2013) 34:919–927

Discussion

GISTs represent the most common mesenchymal neoplasmsoccurring within the abdominal cavity. GISTs likely arisefrom the transformation of interstitial cells of Cajal (ICC)progenitors within the wall of the gastrointestinal tract.Although they can occur in any part of the alimentary tract(even occasionally in the omentum, mesentery, or retroper-itoneum) most GISTs are located in the stomach (60 %). Themajority of GISTs result from activating mutations in theoncogene KIT, which encodes for a corresponding proteinthat is a receptor protein tyrosine kinase (KITLG, also calledCD117). By immunohistochemical staining, most GISTs(90–95 %) are KIT positive, and of those that are KITnegative, some show an increased proportion of KIT muta-tions. Approximately, 5 % of GISTs have mutations in thePDGFRA gene and express little or no KIT. These twomutations cause constitute activation of protein tyrosinekinase signaling pathways, which can be blocked by imati-nib mesylate. The particular sensitivity of this entity totyrosine kinase inhibitors has made this disease a paradigmfor molecularly targeted solid tumors.

The type and frequency of KIT/PDGFRA mutations arethought to relate to the biological behavior of GISTs, and themutational status largely determines the clinical response toimatinib treatment [1–3]. With the exception of GIST, thesegain-of-function mutations are uncommon in other human

cancers [11]. However, some reports have presented othertumors that carry these mutations and indicate that theyrespond to imatinib treatment as well [12–14].

The exact incidence of GISTs is hard to determine, asGISTs have only been properly recognized and uniformlydiagnosed as an entity since the late 1990s. Recentpopulation-based studies estimate their annual incidence asapproximately 10–20 cases/million, with no significant dif-ferences between eastern and western countries [15–19].However, the coexistence of GISTs and other malignanciesis relatively common. Approximately 10 % of the patientsenrolled in the Polish GIST Clinical Registry had a secondneoplasm [20]. In the experience of Kövér et al., 7 of 43patients (16 %) with histologically proven GISTs werefound to have a secondary neoplasm; 3 of these GISTs werecolorectal adenocarcinomas [21]. Au et al. described 74cases of GISTs, 31 of which were diagnosed with secondtumors (41 %), and 38 % of these secondary malignancieswere intestinal tumors [22]. Wronski et al. reported fourother neoplasms (14 %) of the GI tract in 28 GISTs [23],whereas in a literature review, Agaimy et al. reported find-ing 518 cancers in 486 GIST patients among 4,813 caseswith malignancies other than GISTs (approximately 10 %)[24], synchronous or metachronous. As for gastrointestinalmalignancies, Strandberg et al. reported 311 primary GISTsfound in 13,804 cases of gastrointestinal epithelial malig-nant tumors (2.25 %) [25]. It has also been reported that

Table 2 Molecular mutation analysis of DNA samples of group 1

Number GIST PDGFRA Adenocarcinoma Normal tissue

KIT KIT PDGFRA KIT PDGFRA

1 N/M Exon 12a N/M Exon 12a N/M Exon 12a


3 Exon 9 ins (codon 502–503 dup) N/A N/M Exon 12a N/M N/A

4 Exon 11 del (codon 579) Exon 12 N/A; exon 18b N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b

5 Exon 11 del (codon 554–555) N/A N/M Exon 12a N/M Exon 12a


7 N/A N/A N/A N/A N/A N/A




11 Exon 11 point (codon V559DGTT→GAT)

N/A N/M Exon 12a N/M Exon 12a

12 N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b


14 N/A N/A N/A N/A N/A N/A

15 Exon 11 del (556–563) Exon 12a; exon 18b N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b

N/A no amplification, N/M no mutation, ins insertion, del deletiona p.Pro567Pro, CCA→CCG, polymorphic mutationb p.Val824Val, GTC→GTT, synonymous mutation

Tumor Biol. (2013) 34:919–927 923

GIST tumorlets presented in 10 % of patients undergoingsurgery for esophageal carcinoma (18 GISTs in 15 of 150(10 %) patients) [26]. With regard to the association be-tween GISTs and adenocarcinomas, Kawanowa et al. statedthat microscopic GISTs can be found in 35 % of stomach-resected patients with gastric cancer [27].

According to the numbers presented above, the ratevaries from 2.25–41 %, and this variation can be explainedby many reasons (e.g., the sensitivity of GIST diagnosis,involvement of different populations, etc.). After all, thecoexistence of GISTs with other malignancies is relativelycommon. Determining whether tumors coexisting with

GISTs share similar genetic defects, signaling pathways,and responses to targeted therapies is an active area ofoncological research. In these 15 cases of synchronousoccurrence of gastric adenocarcinoma and GISTs in thestomach, we reviewed the clinical profiles and investigatedthe PDGFRA and KIT mutations of the two tumors andnormal tissues. At the same time, for the control group ofpatients with gastric adenocarcinoma alone, we examinedthe mutation status of the tumor lesions and normal tissues.Most lesions were subserosal, with a variable degree ofinvolvement of the muscularis propria. Additionally, eitherthe lesions were not symptomatic or the symptoms caused

Table 3 Molecular mutation analysis of DNA samples of group 2

Number Histology Adenocarcinoma Normal tissue

KIT PDGFRA KIT PDGFRA

1 Moderately to poorly differentiated and SRC N/M Exon 12a N/M Exon 12a

2 Moderately differentiated N/M Exon 12a N/M Exon 12a

3 SRC N/M Exon 12a N/M Exon 12a

4 Moderately to poorly differentiated N/M Exon 12a N/M Exon 12a




8 Moderately to poorly differentiated N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b

9 Poorly differentiated and SRC N/M Exon 12a N/M Exon 12a






15 Poorly differentiated N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b


17 Poorly differentiated and mucinous adenocarcinoma N/M Exon 12a N/M Exon 12a


19 Moderately to poorly differentiated N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b


21 Poorly differentiated and SRC N/M Exon 12a N/M Exon 12a



24 Poorly differentiated and SRC N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b

25 Poorly differentiated N/M Exon 12a N/M Exon 12a

26 Moderately to poorly differentiated NM Exon 12a N/M Exon 12a




30 Poorly differentiated and SRC N/M Exon 12a; exon 18b N/M Exon 12a; exon 18b

SRC signet ring cell carcinoma; N/M no mutationa p.Pro567Pro, CCA→CCG, polymorphic mutationb p.Val824Val, GTC→GTT, synonymous mutation

924 Tumor Biol. (2013) 34:919–927

by GISTs had been non-specific. In our cases, the preoper-ative biopsy fragments showed only adenocarcinoma. Allexcept three of the cases (not available for immunostaining)strongly expressed both CD117 and CD34. All the GISTspresented as a low or very low risk according to the NIH andAFIP risk stratification systems. The coexisting adenocarci-nomas were staged from I to IV by the AJCC Classificationof Gastric Cancer (seventh edition).

Notably, regarding the cases included in our study, theGISTs were all small and showed low levels of proliferation.In accordance with previous reports, few incidental GISTshad a high risk. Upon review of previous reports, we noticedthat GISTs of small size were commonly found duringsurgery for other causes or during autopsy. These mesen-chymal lesions have been variably reported as micro-GISTs,GIST tumorlets, seedling GISTs, ICC hyperplasia, macro-scopic GISTs, sclerosing stromal tumorlets, incidentalGISTs, sporadic ICC hyperplasia, microscopic GISTs, andminimal GISTs [27–32].

Many authors have suggested that the incidence of smallGIST might be much higher than previously estimated. Thecoincidental disease could be benign or malignant, and therehave been cases where tumorlets surround a symptomaticGIST [33, 34]. In the previous reports, these micro-GISTspresented as benign, self-limiting proliferations of intestinalICC. They show an inclination toward spindle-like mor-phology and are immunoreactive for KIT. Consistent withprevious reports, in our series, most cases expressed bothCD117 and CD34, and shared similar KIT/PDGFRA muta-tional status with the overt GISTs. Whereas none of thegastric cancer in our research had any clinically significantmutations of the specific gene, neither did the normal sur-rounding tissues of the control group. In fact, none of thespecimens, aside from GIST lesions, revealed meaningfulmutations in KIT/PDGFRA.

Considering the high frequency of small GISTs, coinci-dence seems more conceivable than a pathophysiologicalrelevant correlation between adenocarcinoma and GIST,especially in a country such as China that exhibits highincidence rates of gastric cancer. There is much debate overwhether GISTs of small size are identical to overt GISTs interms of the biological features and clinical significance.Some researchers suggest that small GISTs are an earlyphase or precursor of clinically manifested GISTs, whileothers think they are biologically indolent. Their histologi-cal features, often typified by spindle-shaped cells set withina hyalinized stroma which also shows calcification, stronglyindicate a tendency to repress [31, 35].

With regard to the molecular aspects of these lesions,controversial results have been observed in different studies.Strandberg et al. and Kawanowa et al. found significantdifferences in the frequency and type of mutations betweenthe tumorlets and the overt GISTs [1, 25, 27], whereas the

mutations encountered by Agaimy et al. were similar tothose observed in clinically apparent, larger GISTs [29].However, we did not find significant differences in thefrequency and type of mutations between our GISTs andpreviously reported overt GISTs.

The prevalence of gastric adenocarcinomas differs greatlybetween western and eastern countries; however, there is noremarkable difference regarding GISTs. The remarkable var-iation in the incidence of microscopic GISTs at different GIsites has been mentioned before; distal esophageal (9 %) andstomach (22 % to 35 %) are the most common locations [26].These observations also favor the hypothesis of a random,incidental association. However, due to the limited number ofcases, we cannot exclude other hypotheses. It is known thatsome populations are susceptible to gastric adenocarcinoma,and there are some etiologic factors responsible for gastricadenocarcinoma. In this epidemic view, there are numerousways in which the two tumors may share something in com-mon that contributes to their synchronous occurrence.Alternatively, it is possible that factors released by one tumorcan affect the other tumor nearby. For example, a stomachharboring a GIST may have a tendency to develop malignantepithelial lesions, or the adenocarcinoma may inhibit theGIST located in the stomach from progressing.

The fact that GISTs in our series presented as ‘benign’caught our attention. It is our hypothesis that mutation ofKIT/PDGFRA in the early stage of tumorigenesis is wide-spread. As extra mutations are needed for GISTs to progress,it is possible that the synchronous tumor may influence theenvironment, releasing factors that inhibit the acquisition offurther genetic changes or simply suppressing the growth ofGISTs. In previous literature, it was mentioned that smallGISTs are more likely to be found in patients with othermalignancies [29, 31, 33].

Some authors also hypothesized that there could be a fieldeffect. The high association of both clinical and microscopicGISTs with GI carcinomas suggests the presence of etiologiccofactors responsible for both lesions. There is some experi-mental evidence supporting this hypothesis [36].

In recent years, synchronous tumors have been reportedmore frequently in publications. Due to the high proportionof GISTs coexisting with other neoplasms, the phenomenonhas received tremendous attention. This study aimed tounveil the underlying connections between dual tumorsfrom a series of cases with synchronous gastric cancer andGISTs. However, clinicopathological profiles and molecularanalysis of the KIT/PDGFRA genes produced no positiveassociations. Considering the high incidence of small GISTsin previous studies, the mere coincidence hypothesis seemsmore likely than a causal association between the tumors.This conclusion is consistent with a previous study [37].

In our series, all the GISTs were diagnosed after R0 resec-tion during a radical surgery for gastric cancer, so additional

Tumor Biol. (2013) 34:919–927 925

surgery for GISTs was not needed. As all of the GISTs werecharacterized by low or very low risk of aggressive behavior,no particular therapy for mesenchymal lesions was recom-mended. The patients with local advanced gastric cancerreceived adjuvant FOLFOX chemotherapy. In surveillance,the patients showed no signs of GIST recurrence.

However, there are some problems needed to be clarified,such as what kind of surgery is sufficient for the GISTs coex-isted with other malignancy, whether the medical therapyshould be modified, even the proper dose of imatinib andconcurrent chemotherapy administrating for both tumors.How to determine a GIST lesion from metastasis or recurrenceof a known malignancy is a challenge under some circum-stances. The answer should be based on a thorough understand-ing of these coexisted GISTs or on more clinical evidences.

Due to the limit of our research, we cannot exclude thefollowing hypotheses: similar susceptible populations, com-mon etiologic factors, interactions between the tumors, mo-lecular mechanisms of tumorigenesis, etc. More researchbased on large populations, animal experiments, and furthermolecular analyses are needed to understand the associationbetween dual tumors.

References

1. Bachet JB, et al. Prognosis and predictive value of KIT exon 11deletion in GISTs. Br J Cancer. 2009;101(1):7–11.

2. Tzen CY, Wang MN, Mau BL. Spectrum and prognostication ofKIT and PDGFRA mutation in gastrointestinal stromal tumors. EurJ Surg Oncol. 2008;34(5):563–8.

3. Lasota J, Miettinen M. Clinical significance of oncogenic KIT andPDGFRA mutations in gastrointestinal stromal tumours.Histopathology. 2008;53(3):245–66.

4. Blay JY, et al. Consensus meeting for the management of gastro-intestinal stromal tumors. Report of the GIST ConsensusConference of 20–21 March 2004, under the auspices of ESMO.Ann Oncol. 2005;16(4):566–78.

5. Gastric Cancer Diagnosis and Treatment Expert Panel of theChinese Ministry of Health. Chinese guidelines for diagnosis andtreatment of gastric cancer (2011 edition). Transl GastrointestCancer. 2012;1:103–14.

6. Hu B, et al. Gastric cancer: classification, histology and applicationof molecular pathology. J Gastrointest Oncol. 2012;3(3):251–61.

7. Zhao X, Yue C. Gastrointestinal stromal tumor. J GastrointestOncol. 2012;3(3):189–208.

8. Fletcher CD, et al. Diagnosis of gastrointestinal stromal tumors: aconsensus approach. Int J Surg Pathol. 2002;10(2):81–9.

9. Gomes A, et al. Molecular analysis of c-Kit and PDGFRA inGISTs diagnosed by EUS. Am J Clin Pathol. 2007;127(1):89–96.

10. Fletcher C. Diagnosis of gastrointestinal stromal tumors: a consen-sus approach. Hum Pathol. 2002;33(5):459–65.

11. Burger H, et al. Activating mutations in c-KIT and PDGFRalphaare exclusively found in gastrointestinal stromal tumors and not inother tumors overexpressing these imatinib mesylate target genes.Cancer Biol Ther. 2005;4(11):1270–4.

12. Terada T. Low incidence ofKIT genemutations and no PDGFRAgenemutations in primary cutaneous melanoma: an immunohistochemical

and molecular genetic study of Japanese cases. Int J Clin Oncol.2010;15(5):453–6.

13. Schneider BJ, et al. Phase II trial of imatinib maintenance therapyafter irinotecan and cisplatin in patients with c-Kit-positive,extensive-stage small-cell lung cancer. Clin Lung Cancer.2010;11(4):223–7.

14. Huh WK, et al. Efficacy and safety of imatinib mesylate (Gleevec)and immunohistochemical expression of c-Kit and PDGFR-beta ina Gynecologic Oncology Group Phase Il Trial in women withrecurrent or persistent carcinosarcomas of the uterus. GynecolOncol. 2010;117(2):248–54.

15. Nilsson B, et al. Gastrointestinal stromal tumors: the incidence,prevalence, clinical course, and prognostication in the preimatinibmesylate era—a population-based study in western Sweden.Cancer. 2005;103(4):821–9.

16. Monges G, et al. The estimated incidence of gastrointestinal stro-mal tumors in France. Results of PROGIST study conductedamong pathologists. Bull Cancer. 2010;97(3):E16–22.

17. Espinosa I, et al. A novel monoclonal antibody against DOG1 is asensitive and specific marker for gastrointestinal stromal tumors.Am J Surg Pathol. 2008;32(2):210–8.

18. Goettsch WG, et al. Incidence of gastrointestinal stromal tumoursis underestimated: results of a nation-wide study. Eur J Cancer.2005;41(18):2868–72.

19. Tzen CY, et al. Incidence of gastrointestinal stromal tumor: aretrospective study based on immunohistochemical and mutationalanalyses. Dig Dis Sci. 2007;52(3):792–7.

20. Ruka W, et al. Other malignant neoplasms in patients with gastro-intestinal stromal tumors (GIST). Med Sci Monit. 2004;10(8):LE13–4.

21. Kövér E, Faluhelyi Z, Bogner B, Kalmár K, Horváth G, TornóczkyT. Kettôs tumorok a gasztrointesztinális traktusban: szinkron ésmetakron stromális (GIST) és epiteliális/neuroendokrin daganatok.Magyar Onkológia. 2004;48(4):315–21.

22. Au WY, Ho KM, Shek TW. Papillary renal cell carcinoma andgastrointestinal stromal tumor: a unique association. Ann Oncol.2004;15(5):843–4.

23. Wronski M, et al. Synchronous occurrence of gastrointestinalstromal tumors and other primary gastrointestinal neoplasms.World J Gastroenterol. 2006;12(33):5360–2.

24. Agaimy A, et al. Occurrence of other malignancies in patients withgastrointestinal stromal tumors. Semin Diagn Pathol. 2006;23(2):120–9.

25. Strandberg L, et al. Interferon-alpha induces up-regulation andnuclear translocation of the Ro52 autoantigen as detected by apanel of novel Ro52-specific monoclonal antibodies. J ClinImmunol. 2008;28(3):220–31.

26. Abraham SC, et al. "Seedling" mesenchymal tumors (gastrointes-tinal stromal tumors and leiomyomas) are common incidentaltumors of the esophagogastric junction. Am J Surg Pathol.2007;31(11):1629–35.

27. Kawanowa K, et al. High incidence of microscopic gastrointestinalstromal tumors in the stomach. Hum Pathol. 2006;37(12):1527–35.

28. Rossi S, et al. Molecular and clinicopathologic characterization ofgastrointestinal stromal tumors (GISTs) of small size. Am J SurgPathol. 2010;34(10):1480–91.

29. Agaimy A, Wunsch PH, Hofstaedter F, Blaszyk H, Rummele P,Gaumann A, Dietmaier W, Hartmann A. Minute gastric sclerosingstromal tumors (GIST tumorlets) are common in adults and fre-quently show c-KIT mutations. Am J Surg Pathol. 2007;31(1):113–20.

30. Agaimy A, et al. Microscopic gastrointestinal stromal tumors inesophageal and intestinal surgical resection specimens: a clinico-pathologic, immunohistochemical, and molecular study of 19lesions. Am J Surg Pathol. 2008;32(6):867–73.

926 Tumor Biol. (2013) 34:919–927

31. Agaimy A, Wunsch PH. Sporadic Cajal cell hyperplasia is com-mon in resection specimens for distal oesophageal carcinoma. Aretrospective review of 77 consecutive surgical resection speci-mens. Virchows Arch. 2006;448(3):288–94.

32. Corless CL, et al. KIT mutations are common in incidental gastro-intestinal stromal tumors one centimeter or less in size. Am JPathol. 2002;160(5):1567–72.

33. Kawanowa K, et al. High incidence of microscopic gastrointestinalstromal tumors in the stomach. Human Pathol. 2006;37(12):1527–35.

34. Chetty R. Small and microscopically detected gastrointestinal stro-mal tumours: an overview. Pathology. 2008;40(1):9–12.

35. Ogasawara N, et al. Frequent c-Kit gene mutations not only ingastrointestinal stromal tumors but also in interstitial cells of Cajalin surrounding normal mucosa. Cancer Lett. 2005;230(2):199–210.

36. Almaca J, et al. TMEM16 proteins produce volume-regulatedchloride currents that are reduced in mice lacking TMEM16A. JBiol Chem. 2009;284(42):28571–8.

37. Agaimy A, Wuensch PH. Gastrointestinal stromal tumours inpatients with other-type cancer: a mere coincidence or an etiolog-ical association? A study of 97 GIST cases. Z Gastroenterol.2005;43(9):1025–30.

Tumor Biol. (2013) 34:919–927 927

Documents

Coexistence of gastrointestinal stromal tumors and gastric adenocarcinomas