Molecular Profiling of KIT/PDGFRA-Mutant and Wild-Type Gastrointestinal Stromal Tumors (GISTs) with Clinicopathological Correlation: An 18-Year Experience at a Tertiary Center in Kuwait

Simple Summary Although gastrointestinal stromal tumor (GIST) is a relatively rare mesenchymal neoplasm of the digestive tract, molecular advancements in GISTs over the past two decades have caused a paradigm shift in the field of precision oncology. GISTs are mainly driven by activating mutations in the KIT or PDGFRA oncogenes, rendering them sensitive to targeted therapies. However, there is a significant lack of data from Kuwait that needs to be addressed. We carried out this retrospective analysis of a cohort of 200 GIST patients at the Kuwait Cancer Center to provide much-needed insights into the genetic makeup and clinicopathological characteristics of our population. We detailed the mutational spectrum in KIT and PDGFRA, identified a small subgroup of wild-type tumors, and shed some light on the clinical implications. This study opens the doors for potential larger-scale, multi-institutional outcome studies in the Arabian Gulf region. Abstract In gastrointestinal stromal tumors (GISTs), identifying prototypical mutations in the KIT/PDGFRA oncogenes, or in rare alternate genes, is essential for prognostication and predicting response to tyrosine kinase inhibitors. Conversely, wild-type GISTs (WT-GIST), which lack known mutations, have limited treatment options. Data on the mutational landscape of GISTs and their impact on disease progression are very limited in Kuwait. Using a targeted next-generation sequencing panel, we investigated the spectrum and frequency of KIT, PDGFRA, and RAS-pathway-related mutations in 95 out of 200 GISTs diagnosed at Kuwait Cancer Center from 2005 to 2023 and assessed their correlation with clinicopathological parameters. Among the 200 tumors (median age 55 years; 15–91), 54% originated in the stomach, 33% in the small bowel, 7% in the colorectum, 1.5% in the peritoneum, and 4.5% had an unknown primary site. Of the 95 molecularly profiled cases, 88% had a mutation: KIT (61%), PDGFRA (25%), NF1 (2%), and one NTRK1 rearrangement. Ten WT-GISTs were identified (stomach = 6, small bowel = 2, and colorectum = 2). WT-GISTs tended to be smaller (median 4.0 cm; 0.5–8.0) (p = 0.018), with mitosis ≤5/5 mm2, and were of lower risk (p = 0.019). KIT mutations were an adverse indicator of disease progression (p = 0.049), while wild-type status did not significantly impact progression (p = 0.934). The genetic landscape in this cohort mirrors that of global studies, but regional collaborations are needed to correlate outcomes with genetic variants.


Introduction
Gastrointestinal stromal tumor (GIST) is a distinctive oncologic-molecular paradigm and a leading example of the utility of tyrosine kinase inhibitor (TKI)-targeted therapies.Although accounting for <1% of gastrointestinal neoplasms, GISTs are the most common mesenchymal tumors of the digestive tract [1].They originate from the interstitial cells of Cajal or related stem cells in the gut wall [2], primarily arising in the stomach (50-60%), followed by the small bowel (30-35%) [3], and less frequently in the colorectum (5%), esophagus (<1%), and very rarely in the appendix and extra-visceral locations [4][5][6][7].Their biological behavior is variable, ranging from indolent to highly aggressive, with risk stratification dependent on tumor size, location, and mitotic rate, which are the main prognostic parameters in localized disease [8].
Data on GIST in the Kuwaiti population is scarce.In this study, we aimed to analyze and molecularly profile 95 cases from a cohort of 200 patients treated at the Kuwait Cancer Center over an 18-year period to gain insights into the genetic landscape, clinicopathological characteristics, and the impact of genotype on disease progression.

Clinicopathological Data
This study included GIST pathology specimens referred to the Kuwait Cancer Center from 2005 to 2023.A retrospective review of medical and histopathological records was conducted to collect clinicopathological data.Pathology slides were reviewed by two consultant pathologists.The pathological parameters included (as per the College of American Pathologists cancer reporting protocols) tumor location, size, histological phenotype, mitotic count, TNM stage, risk assessment category, and immunohistochemical expression of KIT, DOG1, and CD34.Additionally, evidence of metastasis was documented.The clinical data reviewed comprised stage, progression details, and information on TKI therapy.

Molecular Analysis Data
Next-generation sequencing (NGS) was performed on formalin-fixed paraffin-embedded (FFPE) tissues using the Oncomine Comprehensive Assay v3 (OCA v3), a targeted panel that covers 161 cancer-relevant genes (Thermo Fisher Scientific, Waltham, MA, USA) [31] (see Supplementary Table S1 for the gene list).DNA and RNA were extracted using the RecoverAll Total Nucleic Acid Isolation Kit (Thermo Fisher Scientific, Waltham, MA, USA), and the nucleic acid concentration was measured with a Qubit 3.0 Fluorometer.Library preparation was conducted manually following the manufacturer's instructions with the Ion AmpliSeq Library Kit Plus (Thermo Fisher Scientific, Waltham, MA, USA).Sequencing was carried out on the Ion Torrent S5 XL platform (Thermo Fisher Scientific, Waltham, MA, USA).Data analysis was performed using the Ion Reporter™ Software (v.5.10).Sequence reads were aligned to the human genome assembly GRCh37/hg19, achieving a minimum of 10 million total mapped reads.Variant calling was conducted with a minor allele frequency (MAF) cutoff of 5% and an average mean depth of at least 200×.To ensure data quality, additional parameters were applied: base quality score (Q30 or higher), mapping quality (MQ ≥ 60), duplicate read removal (threshold of <5% duplicates), and uniformity (≥90%).Variants were filtered based on predefined quality thresholds, including read depth (minimum of 20 reads supporting the variant), allele balance (≥20%), and strand bias (≤10%).The final variant call set was annotated with clinical and functional information using databases such as ClinVar, COSMIC, and dbSNP.Potential pathogenicity was assessed using tools like SIFT and PolyPhen-2 for comprehensive variant interpretation.Detected somatic variants included single-nucleotide variants (SNV), small deletions (del), insertions (ins), insertions/deletions (indel), and selective copy number variants.

Statistical Analysis
Descriptive statistics were calculated for continuous variables, including mean, median, range, and standard deviation, and for categorical variables, which were summarized using frequencies and graphical representations.Clinicopathological parameters were evaluated using univariate analysis: Pearson's Chi-squared test for categorical variables and a two-sample independent t-test for continuous variables.A p-value of less than 0.05 was considered statistically significant.The statistical analysis was conducted using JAMOVI Version 2.5.7.0.
Although spindle-cell morphology was distributed everywhere, it was predominant in the small bowel and colorectum, while epithelioid tumors were clustered in the stomach (18/23; 78%).High-risk tumors accounted for 100% of EGISTs, 70% of colorectal GISTs, 39% of small bowel GISTs, and 15% of gastric GISTs.No-risk to low-risk tumors were predominantly found in the stomach, comprising 70% of these cases.Regarding focality, four cases had synchronous tumors, including one associated with neurofibromatosis type 1 (NF1) (case #83).Immunohistochemical analysis revealed the expression of KIT, DOG1, and CD34 in 96.7%, 97.9%, and 80.2% of cases, respectively.

Overall Frequencies
Out of 200 tumors, 95 (47.5%) were sequenced, primarily from imatinib-naive primary tumor tissues (87%) and less frequently from metastatic tumor tissues (13%).NGS was conducted on fifty-five gastric tumors, twenty-seven small bowel tumors, seven colorectal tumors, two EGISTs, and four tumors of the unknown primary site.Overall, KIT/PDGFRA mutations were detected in 82 cases (86%), with KIT mutations present in 61% and PDGFRA mutations in 25%.Thirteen cases (14%) were wild-type for KIT/PDGFRA based on our NGS panel (Table 2).Annotated genetic variants and their allelic frequencies in the tumor DNA samples are detailed in Supplementary Table S2.
Within the stomach, KIT and PDGFRA mutations accounted for 51% and 38% of cases, respectively, with all KIT mutations involving exon 11.In the small bowel, KIT mutations accounted for 78% of cases, while PDGFRA was found in only one case.In this location, KIT mutations involved various exons: exon 11 (n = 13), exon 9 (n = 6), exon 13 (n = 1), and exon 8 (n = 1) (Figure 3).Exon 9 mutations were predominantly of the A502_Y503dup variant (5 of 6 cases), were exclusively found in the small bowel (with one exception in a tumor of unknown primary site), and were categorized as moderate-to high-risk.In the colorectum, only exon 11 KIT mutants were detected (5/7 cases).
tations accounted for 78% of cases, while PDGFRA was found in only one case.In this location, KIT mutations involved various exons: exon 11 (n = 13), exon 9 (n = 6), exon 13 (n = 1), and exon 8 (n = 1) (Figure 3).Exon 9 mutations were predominantly of the A502_Y503dup variant (5 of 6 cases), were exclusively found in the small bowel (with one exception in a tumor of unknown primary site), and were categorized as moderate-to high-risk.In the colorectum, only exon 11 KIT mutants were detected (5/7 cases).KIT mutations were predominantly found in spindle-cell tumors (72%), while PDG-FRA (both D842V and non-D842V) were primarily observed in purely epithelioid or mixed tumors (96%) (p < 0.001).A single spindle-cell PDGFRA mutant (S566_E571delinsR) was identified in the small bowel.No statistically significant differences were found concerning age, mean tumor size, mitotic index, TNM stage, pathological risk category, or KIT/DOG1 immunohistochemical expression.A summary of all cases with clinicopathological parameters is provided in Figure 4. KIT mutations were predominantly found in spindle-cell tumors (72%), while PDGFRA (both D842V and non-D842V) were primarily observed in purely epithelioid or mixed tumors (96%) (p < 0.001).A single spindle-cell PDGFRA mutant (S566_E571delinsR) was identified in the small bowel.No statistically significant differences were found concerning age, mean tumor size, mitotic index, TNM stage, pathological risk category, or KIT/DOG1 immunohistochemical expression.A summary of all cases with clinicopathological parameters is provided in Figure 4.

KIT/PDGFRA Wild-Type GISTs
Among the thirteen KIT/PDGFRA wild-type GISTs, three harbored alternative molecular alternations: two had NF1 mutations (including one case associated with known neurofibromatosis type 1 syndrome), and one had an LMNA::NTRK1 gene fusion, leading to its reclassification as an "NTRK-rearranged spindle cell neoplasm" (previously reported [32]).
Ten tumors were negative for mutations in KIT, PDGFRA, and RAS-pathway genes (BRAF, RAS isoforms, and NF1), and for all other 161 cancer-relevant genes that were included in our NGS panel.The median age of patients was 55 years, with a sex distribution of seven females and three males.The median tumor size was 4.0 cm, with four tumors ≤2 cm (range 0.5-8 cm).WT-GISTs were smaller in size compared to mutant GISTs (p = 0.018).The primary locations were the stomach (six cases), small bowel (two cases), and colorectum (two cases).Histological patterns included spindle cell (six cases), epithelioid (two cases), and mixed (two cases) (Figure 5).Mitotic counts were predominantly ≤5/5 mm 2 (in eight cases).WT-GISTs appeared to be of lower risk (p = 0.019), with two of ten classified as high-risk.Immunohistochemically, all 10 cases expressed either KIT, DOG1, or both.SDH deficiency testing was not available.The primary locations were the stomach (six cases), small bowel (two cases), and colorectum (two cases).Histological patterns included spindle cell (six cases), epithelioid (two cases), and mixed (two cases) (Figure 5).Mitotic counts were predominantly ≤5/5 mm 2 (in eight cases).WT-GISTs appeared to be of lower risk (p = 0.019), with two of ten classified as high-risk.Immunohistochemically, all 10 cases expressed either KIT, DOG1, or both.SDH deficiency testing was not available.

Discussion
A substantial body of literature discusses the prognostic significance of mutational alterations in GISTs; however, the genetic landscape of GISTs and its impact on disease progression have not been investigated in our population.In this study, we retrospectively evaluated a cohort of 200 GIST cases diagnosed at the Kuwait Cancer Center over an 18-year period, offering a detailed description of the mutational spectrum in 95 cases.
The overall clinicopathological characteristics of this cohort were largely consistent with those reported in the literature.Gastric tumors were the most common, representing over half of all GISTs (54%) [33].As expected, pure epithelioid GISTs clustered in the stomach and were very rare elsewhere, while spindle cell histology was observed at all sites but predominated in the small bowel (75%) and colorectum (86%).Colorectal GISTs, which made up 7% of this study's cases, included eight rectal and six colonic tumors (three in the left colon, one in the transverse colon, and two unspecified).Although colorectal GISTs may represent two distinct subgroups, they are often combined due to their rarity [5].Colonic GISTs, a rare subgroup, have a slight predilection for the left colon, with a risk of progression associated with large tumor size and high mitotic rate, similar to noncolonic counterparts [6].Consistent with some reports [34,35], most of our colorectal cases (70%) were high-risk and appeared to have a worse prognosis compared to gastric cases.In descending order, high-risk tumors in this study comprised 100% of EGISTs, 70% of colorectal GISTs, 39% of small bowel GISTs, and 15% of gastric GISTs.Conversely, 70% of gastric GISTs were classified as no-risk to low-risk [36].
The overall frequency of gene mutations in this study was 88%, with 61% in KIT, 25% in PDGFRA, and 2% in NF1.The spectrum of KIT/PDGFRA mutations mirrored findings in the literature.KIT exon 11 mutations (the most common) included heterogeneous deletions/insertions/SNVs clustering around codons 550-561 and 568-591 [9,33].The W557_K558del deletion and D842V substitution were the most prevalent among KIT exon 11 mutations.The D842V substitution accounted for 75% of PDGFRA cases [37].KIT mutants were distributed throughout the entire GI tract, while the vast majority of PDGFRA mutations were localized to the stomach (p < 0.001).KIT mutations were predominantly found in spindle-cell tumors [38], whereas PDGFRA mutations were more common in epithelioid and mixed tumors (p < 0.001) [39].Among tumors with mixed histological patterns, mutations were evenly divided between KIT and PDGFRA.
Two EGIST cases with epithelioid and mixed histology were identified as D842Vmutants, similar to gastric GISTs [40].EGISTs located in the omentum, mesentery, and retroperitoneum occur in <10% of cases [7,41,42] and may, in some instances, represent extramural GISTs that have lost their connection to the gut wall.Omental and mesenteric EGISTs are likely related to gastric and small intestinal origins, respectively [43].The rectal tumors in this study harbored KIT exon 11 mutations, while both analyzed colonic cases were wild-type for KIT/PDGFRA/RAS.
Another recurrent but less common mutation was KIT A502_Y503dup in exon 9 (observed in six of seven cases).KIT exon 9 mutants were localized exclusively in the small bowel (except for one case of an unknown primary site) and were associated with spindle-cell morphology.Less common exon 9 variants have been reported in the literature, including mutations occurring in gastric and rectal locations [44].All KIT exon 9 mutants in this study were categorized as moderate-to high-risk, with disease progression detected in five of seven cases.The exon 13 K642E substitution [20,45,46], a primary mutation identified in two cases, has been reported in familial cases [47][48][49][50].In contrast, the exon 13 V654A substitution, which was superimposed on an exon 11 mutation following neoadjuvant imatinib in one case, is typically a secondary mutation associated with acquired imatinib resistance [19,20,51,52].
Interestingly, a rectal GIST from a 61-year-old male (case #35) had a somatic KIT variant (c.1679T>A, p.V560D) superimposed on a familial BRCA1 variant (c.4721delT, p.Leu1574fs).The tumor, classified as stage pT3/moderate-risk, was incidentally discovered during screening for prostate cancer.To date, no clear association exists between BRCA1/2 mutations and GIST, with only rare reports available, including a patient harboring simultaneous BRCA2 and KIT germline mutations manifesting as breast cancer and multiple GISTs, respectively [53,54].Another curious finding in the current study was a KIT Exon 11 mutant in the ileum (case #4) showing a hypermutated genomic signature and microsatellite instability (MSI), although high MSI status is not typically involved in GIST tumorigenesis [55,56].Only a single known hereditary syndromic case was included in this study: an NF1 patient with synchronous small bowl GISTs and advanced disease.Hereditary predisposition to GIST is rare and usually due to germline variants in KIT, PDGFRA, NF1, or SDH [48,57].
Quadruple WT-GISTs lack abnormalities in KIT, PDGFRA, SDH, and RAS signaling pathways [62].Our WT-GISTs lacked KIT/PDGFRA/RAS mutations, but their SDH status was unknown.These included six gastric tumors, four of which had an epithelioid/mixed histology, raising the possibility of SDH deficiency.WT-GISTs appeared to be smaller in size compared to mutant tumors (p = 0.018) and were classified in a lower-risk category (p = 0.019).Three of ten WT-GISTs showed disease progression.Accurate characterization of WT-GISTs is needed as they respond poorly to standard TKIs and may benefit from alternative treatment options [23].The single NTRK-rearranged spindle cell neoplasm in this study was labeled as such, as it seems to represent a distinct entity [63].
Tumor size, mitotic count, and small bowel location were the most powerful predictors of disease progression in this study.Tumor genotype seemed to be an important prognostic variable, with KIT-mutants showing a higher tendency to develop disease progression on univariate analysis [36], yet the difference did not reach statistical significance in the Cox model.Studies have shown that KIT exon 11 deletions involving codons 557 and/or 558 and exon 9 mutations are adverse prognostic indicators [64][65][66][67][68], while KIT exon 11 substitutions and insertions seem to have a more favorable prognosis [69].PDGFRA mutations are generally associated with a more favorable outcome, occurring in gastric tumors with low or no malignant potential [36].In the current study, no significant differences existed among KIT exon 11 and 9, deletion variants and SNVs, or WT-GIST and non-WT-GISTs.
Occasional extra-abdominal metastases were noted, with two tumors metastasizing to bone-specifically, rib and sternum.The rib metastasis, accompanied by a soft tissue mass resembling a primary soft tissue sarcoma, originated from a KIT exon 11-mutated colorectal GIST that had a protracted course of nearly two decades and was preceded via liver metastasis.Bone and soft tissue metastasis were rare in two large pioneering studies by Miettinen et al. [33,70].

Conclusions
This study provides a detailed description of the genetic landscape of GISTs in our population with insights into the clinicopathological characteristics, which is essential for refining GIST precision oncology in Kuwait.There are several limitations in this study, mainly related to its retrospective nature, the relative rarity of GISTs, and missing clinical data.The Cox Proportional Hazard model did not reach a level of significance for progression-free survival assessment.Overall, survival was difficult to use as an endpoint, given the relatively long survival of GIST patients and loss of follow-up.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cancers16162907/s1,Table S1: Genes included in the Oncomine Comprehensive Assay panel; Table S2: Annotated genetic variants and their allelic frequency in the tumor DNA samples.

Figure 3 .
Figure 3. Distribution of molecular alterations based on gastrointestinal locations (n = 95).* NTRKfused spindle cell neoplasms are currently classified as a separate entity.

Figure 3 .
Figure 3. Distribution of molecular alterations based on gastrointestinal locations (n = 95).* NTRKfused spindle cell neoplasms are currently classified as a separate entity.

Figure 5 .
Figure 5. Histopathology of KIT/PDGFRA wild-type GISTs with corresponding KIT immunostaining.(A,D) NF1-mutant in the duodenum with spindle cell morphology and diffuse KIT expression in a known neurofibromatosis type 1 patient.(B,E) KIT/PDGFRA/RAS wild-type in the colon with spindle cell morphology and diffuse KIT expression.(C,F) KIT/PDGFRA/RAS wild-type in the stomach with epithelioid morphology and faint KIT expression.All are at 20× magnification.

Figure 5 .
Figure 5. Histopathology of KIT/PDGFRA wild-type GISTs with corresponding KIT immunostaining.(A,D) NF1-mutant in the duodenum with spindle cell morphology and diffuse KIT expression in a known neurofibromatosis type 1 patient.(B,E) KIT/PDGFRA/RAS wild-type in the colon with spindle cell morphology and diffuse KIT expression.(C,F) KIT/PDGFRA/RAS wild-type in the stomach with epithelioid morphology and faint KIT expression.All are at 20× magnification.

Author Contributions:
Conceptualization, R.H.A.; Data Curation, R.H.A., A.K.M., and A.A. (Abdulsalam Alnajjar); Methodology, R.H.A. and A.R.A.; Formal Analysis, A.R.A., R.H.A., A.A.A., and S.B.; Investigation, E.M.A.M., M.A. (Mona Alateeqi), H.J., A.A. (Ammar Almarzooq), N.B., and Z.A.; Resources, E.M.A.M. and M.A. (Mona Alateeqi); Writing-Original Draft Preparation, R.H.A.; Writing-Review and Editing, R.H.A., A.R.A., and M.A. (Mohammad Alkandari); Visualization, R.H.A.; Supervision, R.H.A. and M.A. (Mohammad Alkandari); Project Administration, R.H.A.; Funding Acquisition, R.H.A.All authors have read and agreed to the published version of the manuscript.Funding: This project was partially funded by a research grant from the Research Sector at Kuwait University, grant no.MG02/15.Institutional Review Board Statement: This study was approved by the Ethics Committee of Medical Research at Ministry of Health, Kuwait (approval number 2018/819 and 2024/2586 on 10 July 2024).Informed Consent Statement: Patient consent was waived due to approval from the Ethics Committee of Medical Research at Ministry of Health in Kuwait, exempting this study for its retrospective nature and anonymous archival samples.
* RAS pathway genes include BRAF, RAS isoforms, and NF1.** NTRK-fused spindle cell neoplasms are currently classified as a separate entity.

Table 3 .
Univariate analysis of clinicopathological features and disease progression.
a Linear Model ANOVA.b Pearson's Chi-squared test.* Indicates statistical significance.