AIM: To investigate the clinicopathologic features of patients with extra-gastrointestinal stromal tumors (EGISTs) in South Korea.

METHODS: A total of 51 patients with an EGIST were identified. The clinicopathologic features, including sex, age, location, tumor size, histology, mitotic rate, immunohistochemical features, genetic status and survival data, were analyzed.

RESULTS: The median age was 55 years (range: 29-80 years), and male:female ratio was 1:1.04. The most common site was in the mesentery (n = 15) followed by the retroperitoneum (n = 13) and omentum (n = 8). The median tumor size was 9.0 cm (range: 2.6-30.0 cm) and the median mitotic rate was 5.0/50HPF. (1/50 - 185/50). KIT was analyzed in 16, which revealed 10 cases with wild-type KIT and 6 cases with an exon 11 mutation. Among 51 patients, 31 patients had undergone surgery, and 10 had unresectable disease and had taken palliative imatinib, which resulted in 22.7 mo of progression-free survival. Of the patients who had undergone surgery, 18 did not take adjuvant imatinib, and 8 of these were categorized as “high risk” according to the risk criteria. However, the relapse-free survival was not different (P = 0.157) between two groups.

CONCLUSION: Because the biologic behaviors of GISTs differ according to the location of the tumor, a more stratified strategy is required for managing EGISTs including incorporation of molecular features.

A gastrointestinal stromal tumor (GIST) is the most common mesenchymal tumor of the gastrointestinal tract (GIT) with approximately 10 new cases diagnosed per 1 million each year[1-3]. Surgery is the only treatment leading to a potential cure, but more than 40% of cases recur and metastasize[4].

A GIST is thought to originate from the interstitial cells of Cajal (ICC), the pacemaker of the peristaltic movement of the GIT[5]. More than 95% of GISTs express the KIT protein, and recently DOG1 (discovered on GIST-1) has also been suggested as a useful diagnostic marker. These two immunohistochemical markers are considered to be the most specific and sensitive markers for GIST[1,6,7]. As for the genetic aberrations, approximately 80% of GISTs have a KIT mutation, and 8% to 10% have mutations in the gene encoding the platelet-derived growth factor receptor alpha polypeptide (PDGFRa). The gain-of-function mutations of those genes are critical in the carcinogenesis of GIST[8]. Thus, inhibitors of KIT and PDGFRα, such as imatinib[9], sunitinib[10] and regorafenib[11] are reasonable options for treatment.

The most common primary site of a GIST is the stomach (60% to 70%) followed by the ileum to jejunum (25% to 30%), the colorectum (5% to 15%), the duodenum (5%), and the esophagus (< 2%)[12,13]. The prognosis and genetic features are distinguishable according to the anatomical location; a gastric GIST has a better prognosis and a higher incidence of an exon 11 mutation of KIT, which is a favorable predictive marker for imatinib treatment, than that of a small intestinal GIST[13,14]. Some GIST develops outside the GIT, such as in the omentum, mesentery, and retroperitoneum, and this type of tumor is called an extra-gastrointestinal stromal tumor (EGIST)[12,15]. Although the incidence of EGISTs is reported to be approximately 10% of all GIST cases[16,17], the clinicopathologic parameters and clinical implications of an EGIST have yet to be defined because of the rarity of these tumors. Moreover, the role of imatinib, the drug of choice for this disease, is still unclear.

In the current study, we analyzed the clinicopathologic features of patients with an EGIST from multiple institutes in South Korea.

In the current study, we analyzed the clinicopathologic features of 51 patients with EGISTs across 7 institutes. The demographics such as age and the sex ratio were similar to those in patients with stromal tumors that had arisen inside the GIT. And with a GIST, the KIT-positive rate was over 90% (47/51), and DOG1-positive rate was 100% (13/13). With regard to the genetic status, however, rate of KIT wild-type was higher than expected (67.5%). Imatinib achieved median PFS of 22.7 mo which is comparable to that of advanced GIST patients who were treated with the drug[21,22].

Compared to a GIST, the prognosis of EGIST is known to be less favorable[15,17]. This is assumed because an EGIST harbors poor prognostic factors, including high proliferative indices, a large tumor size, lymph node involvement, and distant metastasis. Because development outside GIT may result in a delay of the presentation of clinical symptoms, a considerable portion of EGIST cases are diagnosed at a late stage, which can make it difficult to manage the case surgically and thereby results in worse clinical outcomes. In contrast, there are several reports that tumor size does not impact the prognosis of EGIST patients. Reith et al[15] found that a tumor size larger than 10.0 cm dose not influence the clinical outcomes of 48 patients with EGIST. Furthermore, in Yamamoto’s report, tumor size did not correlate with patient survival[23]. However, in these two studies, proliferation indices, such as the mitotic rate, cellularity and Ki-67 expression, were shown to be prognostic factors for survival.

In the current study, we observed similar results. Although tumor size was not associated with survival, the mitotic rate showed a tendency to be associated with survival. Because a substantial portion of EGISTs are diagnosed with a large tumor size, it is possible that tumor size itself may not reflect the biology of the EGIST. Because the tumor size has different clinical implications to the anatomical sites[13], the prognostic role of tumor size in EGISTs requires further analysis.

Approximately two-thirds of patients with a conventional GIST have a KIT mutation at exon 11[3]. Regarding an EGIST, the incidence of this type of mutation is reported to be approximately 40%-50%, which is somewhat lower[16,23]. In the present study, we found 6 cases with an exon 11 mutation out of 16 patients (37.5%). According to these results, it appears that EGIST patients less frequently harbor an exon 11 mutation. As this mutation is indicative of a good response to imatinib, further analysis with a greater number of cases is required.

Surgery has been the frontline treatment of an EGIST[24-30]. After surgery, the administration of imatinib usually follows according to the NIH criteria, which are determined by the tumor size, mitotic rate and anatomic location. In the current study, physicians did not strictly apply these criteria. Although the median tumor size (11.0 cm vs 6.5 cm) and the median mitotic rate (7.0/50 vs 3.0/50) were higher in patients who were administered imatinib (n = 13) than those of the patients who were only observed after surgery (n = 18), 8 out of 18 patients should have nevertheless been treated with imatinib according to the NIH criteria. However, as previously mentioned, RFS was not different (P = 0.157).

Several hypotheses have been suggested for the carcinogenesis of an EGIST. The tumor is identical to a GIST regarding the histologic, immunohistochemical and genetic features[12,15,23]. Because the presence of interstitial Cajal-like cells has been reported in many organs outside the GIT, it is rational to suppose that an EGIST originates from common precursor cells that differentiate into the ICC-derived neoplasm during their development outside of the GIT. Another hypothesis is that this tumor might come from pluripotential stem cells located outside the GIT. The extramural extension of a stromal tumor within the GIT is another hypothesis.

We reported an analysis of the clinicopathologic features and hospital courses of 51 patients with an EGIST. Considering the distinct features of EGISTs, a more precise strategy is required for managing this tumor.