Gastrointestinal stromal tumors (GISTs) are the most common stromal or mesenchymal subepithelial neoplasms originating in the gastrointestinal tract. Because of their relatively broad morphological distribution, GISTs were formerly called leiomyomas, leiomyosarcomas, and leiomyoblastomas of the gastrointestinal tract, until they were found to have clinical, histopathological, and molecular features that differentiated them from the other soft tissue tumors. GISTs have a characteristic morphology and are generally positive for the CD117 antigen, an epitope of the KIT receptor tyrosine kinase (RTK). GISTs are primarily caused by oncogenic mutations in genes encoding either of 2 RTKs, KIT or platelet-derived growth factor receptor A (PDGFRA) [1]. KIT and PDGFRA are growth factor receptors that are activated by their respective ligands including stem cell factor and PDGF-AA to trigger cellular pathways that upregulate proliferation, downregulate apoptosis, and control cell differentiation, adhesion, and motility under normal conditions. Mutations in KIT and PDGFRA result in the constitutive activation of these cellular pathways, leading to spontaneous proliferation and uncontrolled tumor growth [2]. Different mutations, including point mutations, and deletions and insertions, have been found in the different exons or in different regions of a single exon in KIT (exons 9, 11, 13, and 17) and PDGFRA (exons 12, 14, and 18) genes in GISTs.

Identifying gene mutations in individual tumors is critical to improve the efficacy of cancer therapy by matching targeted drugs to specific mutations. Next-generation sequencing (NGS) technologies have revolutionized cancer genomics research by providing an unbiased and comprehensive method for detecting somatic cancer genome alterations. Targeted NGS with a gene panel is a powerful and practical approach as it enables the analysis of high-yield genes or genomic regions with relatively rapid turnaround time, low DNA input, and low cost [345].

Several targeted NGS panels are commercially available, although most of them are designed to cover the important alterations in various cancers. The Ion Torrent™ Oncomine Focus Assay (OFA) is a targeted NGS assay that enables the analysis of over 1,000 biomarkers across 52 key solid tumor genes that are well characterized in published literature and associated with current oncology drugs [67]. This assay comprises 2 separate panels (DNA and RNA) that are designed to interrogate hotspot mutations (35 genes), copy number variations (19 genes), and gene fusions (23 genes). Together, these 2 panels can identify current actionable genetic variants and potential future targets for personalized therapy.

In this study, targeted NGS assay with an OFA panel was performed for the genetic characterization of molecular targets in 30 Korean patients with GIST. In addition, we assessed whether our molecular analysis could be considered as surrogate markers when compared to risk assessment criteria.

The cohort comprised of 50% (15/30) male and 50% (15/30) female Korean patients with a median age at 62 years (range, 41–79 years) with GIST. During the study period (median follow-up period 865 days) there was no recurrence or death from any cause. Four out of the 30 patients received adjuvant imatinib treatment. Immunohistochemistry (IHC) results for c-kit or discovered on GIST-1 (DOG1) were available for 28 out of the 30 GISTs: All 28 were positive for c-kit (28/28) and DOG1 (22/22), whereas there was no negative result for IHC of c-kit or DOG1. The clinicopathologic features of the 30 GIST patients are summarized in Supplementary Table 1. After applying stringent parameters for reliable variant calling (allele burden >4%, coverage depth >500×) and after filtering out potential raw base calling errors, 43 hotspot/other likely pathogenic variants (26 hot spots and 17 other variants; 33 SNVs, 8 Indels, and 2 amplifications) in 16 genes were identified in 26 out of the 30 GISTs. Other than hotspot mutations with >allele burden of 20% were confirmed by direct sequencing. Only one patient out of the remaining 4 GISTs without any alterations received the adjuvant imatinib treatment. KIT variants were most frequent (44%, 19/43), followed by 6 variants in PIK3CA, 3 in PDGFRA, 2 each in JAK1 and EGFR, and one 1 in AKT1, ALK, CCND1, CTNNB1, FGFR3, FGFR4, GNA11, GNAQ, JAK3, MET, and SMO. Based on the mutation types, most of the variants were missense mutations (60%, 26/43), followed by 8 frameshifts, 6 nonsense, and 1 stop-loss, and 2 amplifications (Fig. 1). Previous studies showed that functional SNVs may contribute to the development of GISTs [121112]. Fig. 2 illustrates the incidence of SNVs in combination with other mutations in this study. The results showed that 81% (21/26) of samples had at least 1 or more SNVs, 38% (10/26) had at least 2 or more SNVs, while 69% (18/26) of samples incurred no Indels. A detailed list of frameshift, stop-loss, nonsense, and missense mutations, and amplifications identified in this study is provided in Tables 1 and 2.

GISTs are mesenchymal neoplasms that arise primarily from within the muscular wall of the stomach and small intestines; they rarely occur in extra-intestinal locations. The aggressiveness of the tumor is correlated with tumor size, mitotic activity, and anatomical origin [15]. GISTs frequently harbor activating gene mutations in either KIT or PDGFRA genes and thus are highly responsive to several selective tyrosine kinase inhibitors [1216]. Surgical resection is currently the only curative treatment for localized GIST and constitutes over half of potentially curative GIST treatments. However, the clinical course of GIST ranges from that of a tumor cured by surgical resection to that of a locally advanced or even widely metastatic, and ultimately fatal, disease. This clinicopathologic heterogeneity is paralleled by an underlying molecular diversity: the majority of GISTs are associated with spontaneous activating mutations in KIT, PDGFRA, or BRAF genes, while additional subsets are driven by genetic lesions (often inherited) in NF1 or components of the succinate dehydrogenase enzymatic complex [1718]. Routine genotyping has become an integral part of the management of GISTs undergoing tyrosine kinase inhibitor therapy [15].

In this study, we aimed to identify and evaluate genetic mutations in GIST for selective molecular targets using targeted NGS with the OFA panel. This approach identified 19 KIT and 3 PDGFRA mutations as missense or frameshift mutations in 22 different patients with GIST. Most KIT mutations (89%, 17/19) were in exon 11 of the KIT gene; only one mutation each was detected in exon 9 and 17 (Fig. 3). Gain-of-function mutations in KIT result in a growth advantage by constitutive, ligand-independent activation of the RTK [19]. While most GISTs are heterozygous for a given mutation, in around 15% of the tumors, the remaining wild-type KIT allele is lost, and this is associated with malignant behavior, increased mitotic activity, and topoisomerase II expression [20]. The most frequent hotspot mutation in KIT is detected in exon 11; other mutations are less frequently detected in exon 9 and rarely in exon 13 and exon 17 [21]. Importantly, the presence of KIT del-ins557/558 is an important prognostic factor for poor outcome in comparison with other KIT exon 11 mutations, KIT exon 9 and PDGFRA exon 18 mutations, even in tumors classified as non-high-risk ([very]low and intermediate) that originate from the stomach [22]. In this study, 2 KIT del-ins557/558 and 2 missense mutation at the position 557 were identified. Prognostic evaluations were difficult because the patients were alive during the study period.

PDGFRA and KIT mutations are mutually exclusive and activate similar downstream signal transduction pathways. However, PDGFRA-mutant GISTs are almost exclusively of gastric origin (90%–93%), which is prognostically favorable [12]. The most prevalent genotype is the p.D842V substitution involving the second kinase domain (which corresponds to exon 17 of KIT), which is detected in 60%–65% of all PDGFRA mutated tumors [23]. In this study, 2 patients (DN06 and DN23) carried an exon 18 D842V substitution of the PDGFRA only and were classified as very low and intermediate by NIH risk assessment, respectively. A previous study suggested that GISTs with PDGFRA exon 14 mutations represent a subset of clinically favorable gastric tumors (exclusively gastric location) with almost exclusively epithelioid morphology [24]. However, there was no exon 14 alteration of the PDGFRA in this study.

On the other hand, GISTs without activating mutations in KIT and/or PDGFRA genes tend to have 2 or more mutations (75%, 3/4), compared to GIST with KIT and/or PDGFRA mutations (45%, 10//22). Among them, co-occurrence with PIK3CA mutations was dominant: Detection of PIK3CA mutations in large or metastatic KIT-mutant GISTs may suggest that PIK3CA-mutant clones have a proliferative advantage during disease progression. Tyrosine kinase inhibitors have been successfully used in GIST treatment. However, resistance frequently develops due to secondary KIT mutations or activation of downstream signaling pathways, such as the PI3K/AKT/mTOR pathway. Genotyping of PIK3CA in GISTs may help to differentiate between primary and metastatic tumors with the potential to develop resistance to tyrosine kinase inhibitors and guide therapies with PI3K inhibitors [25].

Interestingly, 2 amplifications, one each in the CCND1 and EGFR genes were identified in GISTs with KIT missense mutation. In previous studies, amplifications were reported for CMYC in 3 of 90 (3.3%), for MDM2 in 5 of 94 (5.3%), for EGFR1 in 5 of 94 (5.3%), and for CCND1 in 7 of 79 (8.9%) evaluable cases. Among them, MDM2 and CCND1 amplifications were associated with clinical and histological malignancy [26]. On the other hand, there was no correlation between EGFR gene amplification or EGFR protein overexpression with GIST [27]. In our study, the EGFR amplification was detected in a low risk, whereas the CCND1 amplification was detected in an intermediate by NIH risk assessment case.

Identifying single or combinations of mutations with the aim of delivering individualized treatment with a single or combination of target agents has been an effective strategy for cancer therapy [2]. Approximately 85%–90% of GISTs harboring KIT or PDGFRA mutations benefit from imatinib treatment before or after surgery and in the setting of unresectable/metastatic disease [28], except for tumors with some specific mutations such as PDGFRA exon 18 D842V [29]. The remaining 10%–15% of GISTs without KIT or PDGFRA mutations are classified as wild-type GIST. These tumors do not respond to imatinib. In this group, several mutations have been identified in genes including those encoding succinate dehydrogenase (SDH) complex subunits, neurofibromatosis type 1, BRAF, and other genes [30]. European Organisation for Research and Treatment of Cancer (EORTC) and Scandinavian Sarcoma Group (SSG) trials further suggested that adjuvant imatinib treatment should be carefully applied to high-risk patients and that the tumor genotype should also be taken into consideration. For example, in the advanced/metastatic setting, the PDGFRA exon 18 D842V mutated GISTs do not benefit from imatinib [29] and a higher dose of imatinib (800 mg daily) is recommended by some institutions for KIT exon 9-mutated GISTs [31]. Evaluation of specific mutations can provide information for a specific tailored therapy. Further trials led to the approval of 2 more drugs by the FDA, sunitinib and regorafenib, which expands options for GIST treatment following failure of imatinib.

Unfortunately, there are 2 major limitations in our study. Due to the inherent underrepresentation of genes in the OFA, critical genes could not be investigated. The OFA does not cover the presence of significantly mutated driver genes that were previously identified (TP53, ARID1A, and CDH1) and some new ones (MUC6, CTNNA2, GLI3, RNF43, and others) associated with gastrointestinal tumor/carcinoma [32]. In fact, the OFA used in this study was too small gene panel for excavation of the complicated genetic alterations in GIST. Thus, a combination of the oncomine comprehensive panel and the Ion S5XL platform (ThermoFisher Scientific) would be better suited for application in routine clinical NGS test for solid tumors [33]. Second, it is possible that no calls of mutations in 4 GISTs with wild-type were due to low tumor purity. Both regular prospective and retrospective quality management processes and adequate designation to enrich tumor cellularity by pathologists in the molecular diagnostics laboratory can reduce the risk for a false negative result. The NGS platform with higher analytic sensitivity can detect mutations in specimens with lower tumor cellularity [34]. Finally, in our study, the relapse-free or disease-free survival analysis was not applicable due to the small sample size (n=30) and short median follow-up duration.

In conclusion, our study confirms the utility of the Ion Torrent sequencing platform with an OFA panel to efficiently identify KIT and PDGFRA mutations associated with GISTs, and other gene mutations associated with solid tumors. These findings may provide a genetic basis for developing new GIST therapeutic agents specifically targeting these gene mutations. As more experience and information are gained from the NGS, it is necessary to expand our understanding of the sensitivity of individualized therapies to specific mutations.