This article is mentioned in:

Introduction

Gastrointestinal stromal tumors (GIST) are the most common mesenchymal tumors affecting the gastrointestinal tract but still represent less than 1% of all gastrointestinal tract tumors. The estimated annual incidence of clinically relevant GIST is ~1/100,000 annually (1,2). A total of ~85% of GIST harbor mutually exclusive activating mutations of the genes encoding KIT proto-oncogene, receptor tyrosine kinase (KIT) or platelet-derived growth factor receptor alpha (PDGFRA) tyrosine kinase receptors leading to ligand-independent activation of these receptors and making tumor cells sensitive to tyrosine kinase inhibitors (TKI) (3,4). Imatinib mesylate inhibits the KIT/PDGFRA receptor by directly binding to the ATP-binding site within the amino-terminal domain of the kinase and thus competitively inhibits ATP binding (3). In prospective randomized clinical phase III trials, imatinib achieved disease control in 70–85% of patients with advanced KIT/PDGFRA-mutant GIST, and the median progression-free survival was 20–24 months. The median survival in patients with advanced disease who are treated with front-line imatinib is up to 5 years (5–7).

A total of ~10–15% of all GIST lack mutations in either of the aforementioned genes and have been classified as KIT/PDGFRA wild-type (WT) GIST (4,8–10). The molecular mechanism underlying tumorigenesis in this subgroup is not fully understood. ~A total of 20–40% of KIT/PDGFRA WT GIST represent succinate dehydrogenase complex (SDH)-deficient GIST. SDH-deficiency is caused by the loss of function mutations in one of the genes encoding the SDH subunits (SDHA, SDHB, SDHC and SDHD) or SDHC epimutation resulting in transcriptional silencing of the gene (11,12). Another subgroup of KIT/PDGFRA WT GIST with competent SDH complex activity is defined as RAS-pathway (RAS-P)-mutant GIST. This group includes GIST with inactivating mutation of the NF1 gene, an activating mutation of BRAF, or RAS (13,14). SDH-deficient and RAS mutant GIST represent half of the KIT/PDGFRA WT GIST. In the remaining KIT/PDGFRA WT GIST, the oncogenic driver remains unknown. Of late, certain extremely rare genetic alterations have been demonstrated, such as ETV6-NTRK3 fusion, and alterations in FGFR1, or FGF4, MAX and MEN1 genes (15–17). Metastatic KIT/PDGFRA WT GIST demonstrates a poor response to imatinib mesylate treatment, and there are no prospective data on response, disease progression and overall survival in this setting (18).

Standard methods currently used to determine activating mutations in the KIT and PDGFRA genes are reverse transcription-quantitative (RT-q) PCR and direct Sanger sequencing. The disadvantages of these methods are that they detect either predetermined mutations (mostly hot spot mutations) or in the case of Sanger sequencing, mutations in a specific gene segment, but only if allele frequency of mutation is at least 20% in the sample. They also do not detect all possible mutations in the KIT and PDGFRA genes. Alternatively, next-generation sequencing (NGS) enables the identification of numerous changes in a wider range of investigated genes that cannot be easily detected with standard methods.

This study was performed with the aim to identify somatic mutations using NGS in metastatic KIT/PDGFRA WT GIST patients and analyze their treatment outcomes in a real-world setting.

Materials and methods

Patients

Between January 2002 and December 2020, 119 patients with metastatic KIT-positive GIST started their treatment at the Institute of Oncology Ljubljana (Ljubljana, Slovenia). From this population of patients, only patients with KIT/PDGFRA/BRAF WT tumors as assessed by standard methods were included in further analysis. The present study was approved (approval no. 0120-204/2021/3) by the local Institutional Review Board of the Institute of Oncology Ljubljana and by the National Medical Ethics Committee of the Republic of Slovenia. All procedures in the present study were performed following the ethical standards of the responsible committees on human experimentation (institutional and national) and the Helsinki Declaration of 1975, as revised in 2013. Individual patient consent was waived for the present study as it was a retrospective study, the research involved no risk to the subjects, and the institutional informed consent forms for treatment included consent for the use of patient's data, materials, and/or test results for research purposes.

Immunohistochemistry (IHC)

Formalin-fixed, paraffin-embedded (FFPE) tissue samples were examined and evaluated by a dedicated pathologist from the Department of Pathology at the Institute of Oncology Ljubljana (Ljubljana, Slovenia). Diagnosis of GIST was confirmed based on microscopic findings and CD117 antigen expression as previously described by the authors (19). The determination of DOG1 antigen expression was performed after 2010 in selected cases.

In KIT/PDGFRA WT GIST patients, SDH deficiency was assessed using IHC. Specifically, CD117 and SDHB were detected on 2–4 µm FFPE tissue sections and dried at 56°C for 2 h. Until 2017, CD117 was detected according to a previously described protocol by the authors (19). From 2017, both CD117 and SDHB were detected using fully automated IHC system Ventana Benchmark XT (Ventana ROCHE Inc.). Both epitopes were retrieved on board employing HIER using high pH Cell Conditioning Solution 1 (cat. no. 950-124; Ventana ROCHE Inc.) for 88 min at 100°C. CD117 was detected using a commercially available rabbit monoclonal antibody CD117 (clone YR145; cat. no. 117R; Cell Marque; Merck KGaA). Primary antibody was diluted 1:200 using DAKO REAL™ antibody diluent (cat. no. S2022; DAKO Agilent Technologies, Inc.). SDHB was detected using commercially available rabbit monoclonal antibody SDHB (clone EP288; cat. no. 466R; Cell Marque; Merck KGaA). Primary antibody was diluted 1:100 using DAKO REAL™ antibody diluent (cat. no. S2022; DAKO Agilent Technologies, Inc.). Both antibodies were incubated on board for 60 min at 37°C. Bound primary antibodies were visualized using 3-step multimer detection system OptiView DAB IHC Detection kit (cat. no. 760-700; Ventana ROCHE Inc.) according to the manufacturer's protocol.

The determination of DOG1 was introduced in 2010 using a fully automated IHC system Ventana Benchmark XT (Ventana ROCHE Inc.). The epitope was retrieved on board employing HIER using high pH Cell Conditioning Solution 1 (cat. no. 950-124; Ventana ROCHE Inc.) for 88 min at 100°C. Until 2019, DOG1 was detected using a commercially available rabbit monoclonal antibody DOG-1 (clone SP31; cat. no. M331; Spring Bioscience; Abcam). Primary antibody was diluted 1:50 using DAKO REAL™ antibody diluent (cat. no. S2022; DAKO Agilent Technologies, Inc.). Since 2019, DOG1 was detected using a commercially available rabbit monoclonal antibody DOG-1 (clone SP31; cat. no. 244R; Cell Marque; Merck KGaA). Primary antibody was diluted 1:200 using DAKO REAL™ antibody diluent (cat. no. S2022; DAKO Agilent Technologies, Inc.). Both DOG1 antibodies were incubated on board for 60 min at 37°C. Bound primary antibody was visualized using 3-step multimer detection system OptiView DAB IHC Detection kit (cat. no. 760-700; Ventana ROCHE Inc.) according to the manufacturer's protocol.

Tumor genotyping

Genomic DNA and RNA for tumor molecular analysis were extracted from FFPE primary tumor samples. Macro-dissection of FFPE sample slides followed by tumor genomic DNA/RNA extraction was performed as previously described by the authors (19,20).

Detection of somatic mutations in KIT (exons 9, 11, 13 and 17) and PDGFRA (exons 12, 14 and 18) was performed in all GIST samples (n=119) from the present study by RT-qPCR and direct Sanger sequencing. Additionally, patients with KIT/PDGFRA WT GIST were tested for the presence of hot spot mutations in BRAF gene as previously described by the authors (19).

A total of 17 samples defined as KIT/PDGFRA/BRAF WT GIST (including the immunohistochemically SDHB deficient patients) were analyzed by NGS using TruSight Oncology 500 DNA kit (Illumina, Inc.): TSO500-DNA and 8 samples were additionally examined with TruSight Tumor 170 RNA kit (Illumina, Inc.): TST170-RNA. One sample was examined for germline alterations using TruSight Cancer Panel (Illumina, Inc.). Libraries were prepared according to the manufacturer's protocol and sequenced on NextSeq 550 Sequencing System. Read alignment to the hg19 reference genome and variant calling was performed using TruSight Tumor 170 v2 Local App software (Illumina, inc.). Bioinformatical analysis was performed as previously described by the authors (20). Targeted analysis of following genes was performed: KIT (LRG_307t1), PDGFRA (LRG_309t1), BRAF (LRG_299t1), SDHA (LRG_315t1), SDHB (LRG_316t1), SDHC (LRG_317t1), SDHD (LRG_9t1) and NF1 (LRG_214t1). Variants were described according to HGVS v2019.05 nomenclature (21).

Clinical evaluation, follow-up and outcome analysis

Patients were followed-up every 3–4 months with clinical evaluation, laboratory testing (complete blood count, biochemistry), abdominal ultrasound, abdominal contrast-enhanced computerized tomography. RECIST criteria until 2007 and Choi criteria were used thereafter for defining response and disease progression (22,23). In case of no disease progression, treatment was continued. In case of progression on imatinib, sunitinib has been used since 2007 and for those progressing on sunitinib, regorafenib since 2014. Patients with a documented local progression were managed surgically if that was feasible.

Statistical analysis

Overall survival was assessed from the beginning of treatment until the date of death from any cause by Kaplan-Meier test. To test the difference in survival between patients with KIT/PDGFRA/BRAF WT GIST vs. mutated tumors as assessed by RT-qPCR and Sanger sequencing, the log-rank test was used (24). P<0.05 was considered to indicate a statistically significant difference. Swimmer plot was applied to present the clinical outcome of patients with KIT/PDGFRA/BRAF WT tumors and their NGS molecular status (25). Statistical analyses were performed using SPSS Statistics for Windows (version 25; IBM Corp.).

Results

A total of 119 patients with metastatic GIST, all older than 18 years, were registered between 2002 and 2020. All 119 tumor samples stained positive for CD 117 staining (Fig. S1), the staining for DOG1 was performed after 2010 in 53 (44.5%) samples. The median follow-up at the time of analysis in October 2021 was 11.5 years. Clinicopathological characteristics of the whole group are presented in Table SI.

In 2 patients, tumor specimens were inadequate for molecular analysis. In the remaining 117 patients, mutational status and mutation location were assessed using standard methods RT-qPCR and Sanger sequencing. Single KIT or PDGFRA mutations were detected in 85% of all GIST patients: KIT exon 11 mutation in 81 patients (68.1%), KIT exon 9 mutation in 9 patients (7.6%), and KIT exon 13 in 1 patient (0.8%). A total of 2 patients had co-primary mutations: one in KIT exons 9 and 13 and one in KIT exons 11 and 13. In 3 patients, samples were taken and analyzed after the start of treatment with imatinib and revealed compound mutations: twice in KIT exons 11 and 17 and once in KIT exons 13 and 17. A total of 5 patients (4.1%) were PDGFRA mutated (one in PDGFRA exon 14, three in PDGFRA exon 18 and one in PDGFRA exon 12). None of the 17 KIT/PDGFRA WT patients was positive for BRAF. Thus, 17 patients (14.3%) were designated as KIT/PDGFRA/BRAF WT GIST (Table I). All 17 tumor samples stained positive for CD 117 antigen, while staining for DOG1 antigen was performed in 16 (94.1%) cases, one sample did not stain positive, probably due to an inadequate tissue sample.

Table I. Mutational status and mutation location for all 119 patients with GIST, assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing.

Table I.

Mutational status and mutation location for all 119 patients with GIST, assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing.

Mutation genotype	n (%)
KIT/PDGFRA/BRAF-mutant GIST
KIT exon 11	81 (68.1)
KIT exon 9	9 (7.6)
KIT exon 13	1 (0.8)
KIT exons 9 and 13a	1 (0.8)
KIT exons 11 and 17b	2 (1.7)
KIT exons 13 and 17b	1 (0.8)
PDGFRA exon 12	1 (0.8)
PDGFRA exon 14	1 (0.8)
PDGFRA exon 18	3 (2.5)
BRAF	0
KIT/PDGFRA/BRAF wild-type GIST	17 (14.3)
Not determined/Unknown	1 (0.8)

a two primary mutations;

b sample taken and analyzed after beginning of treatment with imatinib; GIST, gastrointestinal stromal tumors.

The majority of 17 patients identified as KIT/PDGFRA/BRAF WT GIST were women (70.6%), their median age was 60 years, and 15 (88.2%) received therapy with imatinib. The majority of tumors were large, with a high mitotic rate and a high risk of progression. The most frequently affected sites were the small intestine in 47.0% and the stomach in 29.4%-similar to the whole study group. The clinicopathological characteristics of KIT/PDGFRA/BRAF WT GIST patients and their tumors are presented in Table II.

Table II. Clinicopathological characteristics of 17 KIT/PDGFRA/BRAF wild-type GIST patients (as assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing) and their tumors.

Table II.

Clinicopathological characteristics of 17 KIT/PDGFRA/BRAF wild-type GIST patients (as assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing) and their tumors.

Clinicopathological characteristic	n (%)
Sex
Male	5 (29.4)
Female	12 (70.6)
Age, years
Median	60
Range	32-76
≤60	9 (53)
>60	8 (47)
Imatinib initial therapy
Yes	15 (88.2)
No	2 (11.8)
Primary tumor location
Esophagus	0 (0)
Stomach	5 (29.4)
Small intestine	8 (47.0)
Rectum	2 (11.8)
Other	2 (11.8)
Mitotic ratea
≤5	6 (35.3)
5.1-10	2 (11.8)
>10	9 (52.9)
Primary tumor size, cm
0-5	1 (5.9)
5.1-10	5 (29.4)
10	11 (64.7)
Risk classificationb
High	11 (64.7)
Intermediate	5 (29.4)
Low	1 (5.9)
Very low	0 (0)

{ label (or @symbol) needed for fn[@id='tfn3-or-48-03-08382'] } GIST, gastrointestinal stromal tumors;

a Mitotic rate counted as the number of mitoses per 5 mm²;

b Risk group for GIST adapted from Miettinen and Lasota (53).

The treatment decision in 15 KIT/PDGFRA/BRAF WT GIST patients as assessed by standard methods receiving initial therapy with imatinib differed regarding the period when they were treated. Before 2016, the decision for imatinib treatment in Slovenia was based on a pathohistological diagnosis of GIST solely (13 patients). After 2016, two patients were offered imatinib treatment based on reports of clinical responses to imatinib also in KIT/PDGFRA/BRAF WT patients (26,27). One of the two (2/17) KIT/PDGFRA/BRAF WT GIST patients not treated with imatinib, had a solitary lesion in the liver and was treated with radiofrequency ablation. The disease did not recur. The second patient was treated with upfront sunitinib. Altogether, five patients were treated with sunitinib occasionally in the course of their disease and none received regorafenib.

The median overall survival of patients with KIT/PDGFRA/BRAF WT GIST vs. mutated tumors as assessed by standard methods was 4.02 years (95% CI 0.17-7.86) and 5.98 years (95% CI 4.92-7.05), respectively (log-rank P=0.232; Fig. S2).

Results of IHC assessment of SDH competence/deficiency in 16 out of 17 KIT/PDGFRA/BRAF WT GIST patients as assessed by standard methods RT-qPCR and Sanger sequencing are provided in Table III. Specifically, loss of IHC staining for SDHB protein was detected in 4 patients.

Table III. Mutational status, mutation location, allelic frequency, pathogenicity, and clinical benefit to imatinib in KIT/PDGFRA/BRAF WT GIST patients as assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing.

Table III.

Mutational status, mutation location, allelic frequency, pathogenicity, and clinical benefit to imatinib in KIT/PDGFRA/BRAF WT GIST patients as assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing.

n	SDH IHC	Clinically significant gene variants in GIST (NGS)	cDNA variant description	Predicted protein	Pathogenicity ACMG/AMPf	VAF (%)	TMB (mut/MB) ns/(ns+s)	Clinical benefit to imatinib	RECIST evaluation (22)
1a	Loss	WT					6/7	NT	NA
2	Comp.	WT					1/2	No	PD
3b	Comp.	WT					1/1	No	PD
4	Comp.	WT					0/1	No	PD
5	Comp.	WT					2/3	No	PD
6b,c	Comp.	NF1 (LRG_214t1)	c.4537C>T	p. (Arg1513*) R1513*	P	96.27	2/2	Yes	SD
7	Loss	SDHA (LRG_315t1)	c.776delinsGC	p. (Tyr259Cysfs*62)	LP	69.66	4/4	No	PD
				Y259Cfs*62
8	Loss	SDHB (LRG_316t1)	c.268C>T	p. (Arg90*) R90*	LP	30.4	NA	No	PD
9b,d	Loss	SDHB (LRG_316t1)	c.688C>T	p. (Arg230Cys) R230C	LP	79.31	4/4	NT	NA
10b,g	Comp.	PDGFRA (LRG_309t1)	c.2525A>T	p. (Asp842Val) D584V	P	41.17	0/0	Yes	CR
11b	Comp.	PDGFRA (LRG_309t1)	c.1682T>A	p. (Val561Asp) V561D	P	84.61	2/3	Yes	CR
12b	Comp.	PDGFRA (LRG_309t1)	c.1936A>G	p. (Lys646Glu) K646E p.	LP LP	40.56	2/3	Yes	CR
			c.1975A>C	(Asn659His) N659H		41.27
13b	Comp.	PDGFRA (LRG_309t1)	c.1679T>A	p. (Val560Asp) V560D	P	43.35	1/1	Yes	CR
14	Comp.	KIT (LRG_307t1)	c.1669_1674del	p. (Trp557_Lys558del)	LP	47.21	0/0	Yes	PR
				W557_K558del
15	Comp.	KIT (LRG_307t1)	c.1671_1676del	p. (Trp557_Val559delinsCys)	LP	41.2	1/3	Yes	PR
				W557_V559delinsC
16b	Comp.	KIT (LRG_307t1)	c.1648-3_1673del	p. (Lys550_Lys558del)	LP	39.9	2/2	Yes	PR
				K550_K558del
17e	NA	NA	NA	NA	NA	NA	NA	No	PD

{ label (or @symbol) needed for fn[@id='tfn6-or-48-03-08382'] } WT, wild type; P, pathogenic; LP, likely pathogenic; VAF, variant allele frequency; TMB, tumor mutational burden; ns, nonsynonymous; s, synonymous; CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; RECIST, response evaluation criteria in solid tumors; NA, not applicable; NT, not treated; comp., competent;

a upfront treated with sunitinib;

b besides DNA, also RNA was sequenced by TruSight Tumor 170-RNA;

c confirmed germline mutation;

d radiofrequency ablation of solitary liver metastasis;

e not enough material for additional testing;

f Pathogenicity determined according to ACMG/AMP guidelines (58);

g resistant mutation PDGFRA D84.2V.

Molecular status of genes frequently altered in GIST was reevaluated using a targeted NGS approach in all 17 KIT/PDGFRA/BRAF WT GIST samples. In one case the quantity of remaining DNA was insufficient for NGS analysis. Mutational status, mutation location, variant allele frequency, and pathogenicity of detected alterations in 17 KIT/PDGFRA/BRAF WT GIST patients are presented in Table III. The deposition of the full sequences of KIT and PDGFRA in 17 GIST patients in the National Center for Biotechnology Information (NCBI) gene bank is in progress. A total of 7 out of 16 samples with sufficient DNA yield for NGS (43.8%) were found to carry a pathogenic alteration, either in KIT (4 patients) or PDGFRA (3 patients). They all responded to imatinib-most surprisingly, including the patient with PDGFRA D842V mutation. Besides pathogenic alterations in the aforementioned genes, NGS also revealed SDH alterations in 3 patients that had already been suspected based on the loss of IHC staining for SDHB protein (Fig. S3). One sample demonstrated a pathogenic variant in NF1, likely in the context of neurofibromatosis type 1. The detected NF1 pathogenic variant in this patient was confirmed to be of germline origin. No fusions in genes included in TruSightTumor 170-RNA kit were detected in examined samples. Only 5 out of 16 cases with sufficient DNA (31%) of KIT/PDGFRA/BRAF WT GIST were confirmed to be KIT/PDGFRA/BRAF/SDH/NF1 WT also by NGS, and none of them responded to TKIs. Their survival was typically poor except for one patient who remains alive with metastatic disease but has been treated surgically.

A total of 7 out of 8 patients that responded to imatinib had activating mutations in KIT or PDGFRA confirmed by NGS. One patient that had clinical benefit with stable disease, had the NF1 mutation (c.4537C>T; p. Arg1513*; Table III). Type of treatment, duration of response, survival, and NGS molecular status of 17 KIT/PDGFRA/BRAF WT GIST patients as assessed by standard methods is presented in Fig. 1.

Figure 1. Clinical outcome in 17 KIT/PDGFRA/BRAF wild-type GIST patients (as assessed by standard methods reverse transcription-quantitative PCR and Sanger sequencing) and their next-generation sequencing molecular status. Patients are listed according to Table III. KIT, KIT proto-oncogene, receptor tyrosine kinase; PDGFRA, platelet-derived growth factor receptor alpha; GIST, gastrointestinal stromal tumors.

The percentage of KIT/PDGFRA/BRAF WT GIST patients as determined by standard methods (14.3%; 17 out of 119 patients) was reduced to 7.6% (9 out of 119 patients) after re-evaluation with NGS. However, only five out of 119 patients (4.2%) were truly WT GIST patients (excluding those without SDH and NF1 variants).

Discussion

In patients with GIST, it has become mandatory to determine the driver mutation in order to predict the response to TKI treatment. The European Society for Medical Oncology (ESMO) guidelines, however, do not specify the screening method for the detection of somatic mutations (28). RT-qPCR, direct Sanger sequencing or NGS may be used, but in the past, the first descriptions of gain of function mutations in KIT/PDGFRA were based on RT-qPCR and direct Sanger sequencing (4,29,30). These methods have remained the cornerstone for detecting oncogenic drivers in GIST until today. Nevertheless, it has been observed in routine clinical practice that there are patients with KIT/PDGFRA WT GIST by conventional RT-qPCR and direct Sanger sequencing that respond to imatinib. With these observations, the reliability of the results of RT-qPCR and direct Sanger sequencing in the identification of genuinely KIT/PDGFRA WT GIST are questioned.

In the present study, mutations detected by NGS as well as clinical outcomes of the KIT/PDGFRA/BRAF WT GIST were reported (n=17). Patients were selected from a larger series of GIST patients (n=119) analyzed for mutational status with conventional methods RT-qPCR and direct Sanger sequencing.

Among initially included 119 patients, the clinicopathological characteristics showed no difference when compared with published data. As previously reported, also in the present study, there was a slight male predominance (50.4%) (5,6,31). The age at diagnosis of our patients ranged from 32 to 89 years, with the median age being 63.5 years. The median age across most studies falls in the mid-60s (1,2,5–7). In our cohort, GIST most commonly arose in small intestine and stomach (37.0%) while in the previous study, the majority of GIST were found in the stomach and less commonly in the small intestine (1,2,8,32). This may be an area of interest for future research. There was also no difference in the proportion of KIT/PDGFRA/BRAF WT GIST as determined with conventional RT-qPCR and direct Sanger sequencing between our study (14.3%) and data published in the literature, where they observed 7.1–15% of WT GIST (2–4).

In the present study, KIT/PDGFRA/BRAF WT GIST by conventional RT-qPCR and direct Sanger sequencing were found to arise predominantly in females (70.6%), at a younger age than the whole study population (median age 60 years), while other clinical characteristics (primary location, mitotic count, risk group) did not differ from the whole study population. KIT/PDGFRA/BRAF WT GIST by conventional RT-qPCR and direct Sanger sequencing were reported to arise predominantly in the stomach if SDH mutant or epimutant and predominantly in the small intestine if SDH competent (33). In the present analysis, they were found more often in the small intestine than in the stomach, regardless of SDH status. Importantly, only adult GIST patients were included in the present analysis. The natural history of KIT/PDGFRA/BRAF WT GIST by conventional RT-qPCR and direct Sanger sequencing has been suggested to be more indolent than that of KIT/PDGFRA-mutated GIST, but the present study demonstrated a shorter (yet not significantly) overall survival compared with the whole study group (median overall survival 4.02 years vs. 5.87 years) (8,34,35). KIT/PDGFRA/BRAF WT GIST by conventional RT-qPCR and direct Sanger sequencing, that are also confirmed to be WT (for KIT/PDGFRA alterations) by NGS, can be associated with hereditary syndromes, such as type 1 neurofibromatosis (NF1), Carney triad (23), Carney Stratakis syndrome and hereditary paraganglioma/pheochromocytoma (HPGL/PCC) syndromes (8,33). In the present study, only one patient had a hereditary syndrome (NF1).

In the present study, 15 of total 17 KIT/PDGFRA/BRAF WT GIST patients by conventional RT-qPCR and direct Sanger sequencing were available for assessing response to imatinib. Patients (n=7) that responded to imatinib were, however, diagnosed to carry pathogenic alterations by NGS; either in KIT or PDGFRA. One patient with long-term, on-going complete response on imatinib was a carrier of PDGFRA D842V mutation. Patients with PDGFRA D842V mutation rarely respond to imatinib (34–37). Farag et al (37) reported partial response in 2 out of 16 PDGFRA D842V-mutated GIST patients and stable disease in 3 of totally 16 patients. The percentage of patients carrying pathogenic alterations either in KIT or PDGFRA in the present study (41.2%) was higher than the one reported by Heinrich et al (7) demonstrating pathogenic alterations either in KIT or PDGFRA in 3 of total 20 patients (15%) (5–7). No new mutations of KIT and PDGFRA were observed in the present study; all mutations discussed in Table III have already been previously reported in NCBI ClinVar and/or Cosmic and discussed (30,38–46).

One patient that had clinical benefit with stable disease was a carrier of NF1 germline pathogenic variant and the reason for benefit could only be hypothesized. Putatively, some of the lesions followed with CT scans could have been other entities (not verified) in the context of confirmed NF1 and not GIST metastases. A possible explanation for response to TKI could also be attributed to the increased KIT activation through phosphorylation. Namely, the efficacy of TKI in patients with increased KIT activation (detected in some of KIT/PDGFRA/BRAF WT tumors), suggested that in some of these patients KIT activation may play a role in the pathophysiology (26). Moreover, similar to our observation for the patient carrying pathogenic variant in NF1 exon 34 (c.4537C>T), Mussi et al (47) reported stable disease in one patient with NF1 exon 18 mutation after therapy with imatinib.

The remaining 7 out of 15 patients available for assessing response to imatinib did not respond. Among them, two samples demonstrated the SDH alteration which had already been suspected by IHC and was later verified with NGS. Patients with SDH alterations (mutation or epigenetic silencing) are known not to respond to imatinib (7,12). A total of 5 patients remained WT also by NGS, among whom 4 were initially treated with imatinib and did not respond, and one was treated with sunitinib and did not respond. This is consistent with the published literature stating that NGS WT GIST does not respond to TKI, regardless of whether SDH deficient or SDH competent by IHC (9). In the remaining not responding patient the quantity of DNA was not sufficient for NGS analysis.

In 16 GIST patients, the tumor mutational burden (TMB) was assessed. The calculated TMB was low in all patients (less than 10 mutations/MB). These results are in accordance with the previous study by Chalmers et al (48) but differ from the results by Feng et al (49). High TMB correlates with the response to immune checkpoint inhibitors in solid tumors and pembrolizumab was approved for adults and children with TMB-high solid tumors (50). In oncogene-driven disease, it would be expected that a lower TMB correlates with a longer response to TKI. At present, to the best of our knowledge, no aware study exists on the correlation between the TMB score and the largest clinical benefit of imatinib in GIST patients.

KIT/PDGFRA WT GIST has been reported to comprise 10–15% of all GIST with standard diagnostic methods RT-qPCR and direct Sanger sequencing. In large randomized clinical trials that establish the recommendations for the treatment of metastatic GIST, NGS has not been performed. In the phase III BFR14 trial, mutational status was determined by PCR and Sanger sequencing and identified 11.4% WT GIST (51). In the B2222 study, mutational status was determined by PCR and high-pressure liquid chromatography (HPLC) (3). Activating mutations were identified in 92.9%. In EORTC 62005 study mutations were screened by denaturing high-performance liquid chromatography (DHPLC) and characterized by bi-directional DNA sequencing (52). Activating mutations of KIT or PDGFRA were found in 84% of GIST. In SWOG S0033/CALGB 150105 study, a combination of PCR amplification, DHPLC screening, and sequencing was used and demonstrated 16.9% KIT/PDGFRA WT GIST (83.1% of samples carried KIT/PDGFRA mutations) (18). In the follow-up analysis of the SSG XVIII/AIO trial, conventional Sanger sequencing was used and identified 7% KIT/PDGFRA WT GIST (53). The results of the present study are absolutely in concordance with the results of the aforementioned studies. Namely, in the present study, 85% of samples were identified to be KIT/PDGFRA mutation-positive with RT-qPCR and Sanger sequencing, while 14.3% of samples were KIT/PDGFRA WT GIST.

However, in recent years, studies of NGS in GIST have been published, demonstrating that this technique has the potential to detect genetic alterations otherwise not detectable with conventional techniques, particularly in KIT/PDGFRA WT tumors (8,13,17,54,55). Indeed, ESMO guidelines recommend tumor molecular testing in GIST patients if TKI are considered as a part of the management, however, the preferential method is not specified (28). On the other hand, the latest National Comprehensive Cancer Network guidelines recommend that all GIST lacking a KIT or PDGFRA mutation should be examined for SDH deficiency and alternative driver mutations using NGS (56).

When comparing NGS with RT-qPCR and direct Sanger sequencing, NGS provides the highest positivity rate (43). This conclusion indicates a potential advantage of NGS for the detection of targetable mutations in real-world clinical routine setting. In the present study, using the NGS in 16 out of 17 KIT/PDGFRA/BRAF WT GIST as determined by conventional RT-qPCR and Sanger sequencing, mutated genotype was identified in 69% (11 patients). NGS identified 4 cases with KIT mutation, 3 cases with PDGFRA mutation, both missed in conventional analysis. It also identified 3 cases with SDH complex mutation (all suspicious on IHC staining) and 1 NF1 germline pathogenic variant. All patients with KIT/PDGFRA mutated genotype by NGS responded to imatinib. Similar results have also been reported by previous studies. In the report from Heinrich et al (18), the presumed causative mutation was identified with a combination of PCR and Sanger sequencing in 83% of GIST (KIT and PDGFRA analysis only). Using the NGS technologies, Heinrich et al (7) estimated that 97.5% of GIST can be assigned a mutated genotype. NGS of 20 cases originally classified as KIT/PDGFRA WT GIST disclosed 17 (85%) carrying a mutation, most commonly (in 60%) in SDH complex. Astolfi et al (43) indicated that ~20% of patients, assessed by Sanger and IHC, with KIT/PDGFRA/SDH/RAS-P WT GIST are carriers of pathogenic KIT mutations, thus expected to be eligible for and responsive to the various therapeutic lines of TKI in use for KIT/PDGFRA mutant GIST. Vanden Bempt et al (57) also demonstrated that the percentage of WT GIST decreases with the use of NGS. Optimizing detection assays of genetic alterations led to the identification of KIT or PDGFRA mutations in 92.0% of GIST samples and reduced the number of WT GIST to 1.2% only.

In conclusion, the results of the present study demonstrated that ~41.2% of patients diagnosed through standard methods (RT-qPCR and Sanger sequencing) with the KIT/PDGFRA/BRAF-pathway WT GIST are actually carriers of pathogenic KIT/PDGFRA mutations when re-evaluated by NGS, and are thus expected to be eligible for and responsive to the various therapeutic lines of TKI in use for KIT/PDGFRA-mutant GIST. The percentage of genuinely KIT/PDGFRA WT GIST in the present study decreased from the initial 14.3% when assessed with the RT-qPCR and Sanger sequencing to 7.6% after NGS. Therefore, it is strongly considered that NGS should become prerequisite for treatment decision at least in KIT/PDGFRA/BRAF-pathway WT GIST as determined by RT-qPCR and Sanger sequencing.

The limitation to the present study is a small number of NGS samples, as well as a long follow-up with different molecular methods used throughout the course of study, diverse methods of evaluation, and variable access to treatment.

Supplementary Material

Supporting Data

Supporting Data

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The data presented in this study are openly available in Sequence Read Archive (NCBI-SRA), accession number PRJNA824718.

Authors' contributions

MU, SN and BJN conceptualized the present study. Formal analysis, MU, AB, SN, VS, VSD, OB, GK and BJN performed formal analysis. MU, AB, SN, VS, VSD, OB, GK and BJN conducted investigation. MU, SN and BJN provided methodology. MU and BJN performed project administration. MU, AB, SN, VS, VSD, OB, GK and BJN provided resources, SN and BJN supervised the present study. MU and SN conducted visualization. MU wrote the original draft. MU, AB, SN, VS, VSD, OB, GK and BJN wrote, reviewed and edited the manuscript. All authors have read and approved the final version of the manuscript. SN and BJN confirm the authenticity of all the raw data.

Ethics approval and consent to participate

The present study was approved (approval no. 0120-204/2021/3) by the local Institutional Review Board of the Institute of Oncology Ljubljana and by the National Medical Ethics Committee of the Republic of Slovenia (Ljubljana, Slovenia). All procedures in the present study were done following the ethical standards of the responsible committees on human experimentation (institutional and national) and the Helsinki Declaration of 1975, as revised in 2013. Individual patient consent was waived for this study as it was a retrospective study, the research involved no risk to the subjects, and the institutional informed consent forms for treatment included consent for the use of patient's data, materials, and/or test results for research purposes.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References