Profiling of gene fusion involving targetable genes in Chinese gastric cancer

Abstract

BACKGROUND

Approximately half of all new cases of gastric cancer (GC) and related deaths occur in China. More than 80% of patients with GC are diagnosed at an advanced stage, which results in poor prognosis. Although HER2-directed therapy and immune checkpoint inhibitors have been somewhat successful, new drugs are still needed for the treatment of GC. Notably, several gene fusion-targeted drugs have been approved by the United States Food and Drug Administration for solid tumors, including GC, such as larotrectinib for NTRK fusion-positive cancers and zenocutuzumab for NRG1 fusion-positive cancers. However, gene fusions involving targetable genes have not been well characterized in Chinese patients with GC.

AIM

To identify the profile of fusions involving targetable genes in Chinese patients with GC using clinical specimens and determine the distribution of patients with gene fusion variants among the molecular subtypes of GC.

METHODS

We retrospectively analyzed gene fusion events in tumor tissue samples from 954 Chinese patients with GC. Clinicopathological characteristics were obtained from their medical records. Genetic alterations, such as single nucleotide variants, indels, amplifications, and gene fusions, were identified using a targeted sequencing panel containing 825 genes. Fusions were validated by fluorescence in situ hybridization (FISH) using break-apart probes. The microsatellite instability (MSI) status was evaluated using MSIsensor from the targeted sequencing panel data. Tumor mutational burden (TMB) was calculated using the total number of nonsynonymous mutations divided by the total genomic targeted region. Chi-square analysis was used to determine the enrichment of gene fusions associated with the molecular subtypes of GC.

RESULTS

We found that 1.68% (16/954) of patients harbored 20 fusion events involving targetable genes. RARA fusions (n = 5) were the most common, followed by FGFR2, BRAF, MET, FGFR3, RET, ALK, EGFR, NTRK2, and NRG1 fusions. Two of the RARA fusions, EML4-ALK (E6:E20) and EGFR-SEPTIN14 (E7:E10), have been identified in other tumors but not in GC. Surprisingly, 18 gene fusion events were previously not reported in any cancer types. Twelve of the eighteen novel gene fusions included complete exons encoding functional domains of targetable genes, such as the tyrosine kinase domain of receptor tyrosine kinases and the DNA- and ligand-binding domains of RARA. Consistent with the results of detection using the targeted sequencing fusion panel, the results of FISH (fluorescence in situ hybridization) confirmed the rearrangement of FGFR2 and BRAF in tumors from patients 04 and 09, respectively. Genetic analysis indicated that the fusion genes were significantly enriched in patients with ERBB2 amplification (P = 0.02); however, there were no significant differences between fusion-positive and fusion-negative patients in age, sex, MSI status, and TMB.

CONCLUSION

We characterized the landscape of fusions involving targetable genes in a Chinese GC cohort and found that 1.68% of patients with GC harbor potential targetable gene fusions, which were enriched in patients with ERBB2 amplification. Gene fusion detection may provide a potential treatment strategy for patients with GC with disease progression following standard therapy.

Key Words: Gene fusion, Targetable genes, Gastric cancer, Chinese population, ERBB2 amplification

Core Tip: The proportion of patients with gene fusions in Chinese patients with gastric cancer (GC) has not yet been characterized. In our analysis, we found that 1.68% of such patients harbor fusions involving targetable genes. Moreover, these fusion genes were enriched in patients with ERBB2 amplification. Our study indicates that gene fusion detection may provide a novel approach for GC therapy.

INTRODUCTION

Gastric cancer (GC) is the fifth most frequent cancer and the third leading cause of cancer deaths worldwide, with more than one million new cases and approximately 769000 deaths in 2020[1]. The overall survival rate of patients with early stage disease is around 90% after surgical resection[2]; however, more than 80% of patients with GC are diagnosed at an advanced stage in China, which limits the effectiveness of the treatment[3]. Although chemotherapy has improved the survival of advanced-stage patients with GC, the objective response rate remains less than 40%, and the median overall survival is less than 12 mo[4]. Nevertheless, new targeted therapies are capable of improving the objective response rate and overall survival of patients with GC expressing certain targets[5].

Approximately 13%-22% of GCs exhibit HER2 overexpression or amplification[6-8]. The College of American Pathologists, the American Society for Clinical Pathology, and the American Society of Clinical Oncology recommend that all patients with advanced gastric adenocarcinoma should be tested for HER2 overexpression[9]. Trastuzumab was approved by the United States Food and Drug Administration (FDA) in 2010 as first-line treatment in combination with chemotherapy for patients with HER2-positive GC. Microsatellite instability-high (MSI-H) tumors are considered a molecular subtype of gastric adenocarcinoma by The Cancer Genome Atlas (TCGA)[10]. The incidence of MSI-H GC is 10%-20%[11]. The NCCN guidelines recommend MSI testing as a standard test for all patients with GC. Regarding targeted therapy, the FDA has approved pembrolizumab (PD1 monoclonal antibody) for the treatment of all unresectable or metastatic solid tumors with MSI-H/dMMR (deficient DNA mismatch repair), including GC. Although drug treatments have shown success to some extent, the development of more targeted drugs is required.

With rapid advancements in the field of oncogenomics, gene fusions in cancer have received increasing attention. The FDA has approved larotrectinib (Vitrakvi) and entrectinib (Rozlytrek) for the first- or subsequent-line treatment of solid tumors with NTRK fusions, including GC[12,13]. In 2021, the FDA accelerated the approval of the NRG1 inhibitor, zenocutuzumab (MCLA-128), in patients with pan-cancer harboring an NRG1 fusion. Apart from these fusion genes with approved drugs in pan-cancer, ALK fusions, such as EML4-ALK, TFG-ALK, and STRN-ALK, have been identified in the majority of tumors, including lung adenocarcinoma and colorectal cancer[14-16]. For lung cancer and mesenchymal tumors, patients harboring an ALK fusion are highly responsive to crizotinib and ceritinib[17,18]. Recently, a RAB10-ALK fusion was identified in a patient with GC[19], which indicates the possibility of future applications of ALK-TKIs (tyrosine kinase inhibitors) in these patients. Recent advances in next-generation sequencing (NGS) have contributed to a surge in the discovery of fusion genes, including BRAF; EGFR; FGFR1, 2, and 3; RET; and ROS1[20]. Gene fusion detection can guide the development of targeted therapeutic strategies for patients with GC with disease progression after standard therapy. Notably, there is a lack of comprehensive data characterizing gene fusions involving targetable genes in GC, particularly in the Chinese population.

MATERIALS AND METHODS

Patients

This multicenter retrospective study included 1341 patients with GC admitted to Fujian Provincial Hospital (Fuzhou, China) and Zhejiang Provincial People’s Hospital (Hangzhou, China) between October 2015 and December 2021. The clinicopathological characteristics of the patients were retrieved from their medical records. Additionally, MSI status and tumor mutational burden (TMB) scores were extracted for statistical analysis. This study was approved by the Ethics Committee of the Fujian Provincial Hospital.

Mutational profiling

Mutational profiling of the Onco PanScan panel was performed by Genetron Health (Beijing) Co., Ltd. The coding regions of 825 cancer-related genes were analyzed. Genomic DNA was isolated from formalin-fixed paraffin-embedded (FFPE) tissue specimens with a minimum of 20% viable tumor nuclei. For sequencing, paired tumor and white blood cell DNA libraries were prepared using KAPA HyperPrep Kits (Roche, Germany). Libraries were quantified using Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, United States), and their quality was evaluated using Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, United States). High-throughput sequencing was performed on Novaseq6000 platform (Illumina, United States). Paired-end reads from Illumina sequencing were processed using script bcl2fastq (v. 2.17.1.14) and aligned against the human genome reference build, GRCh37, using Burrows-Wheeler Aligner (BWA, version 0.7.13). Duplicate removal, local realignment, and base quality recalibration were performed using PICARD (http://broadinstitute.github.io/picard/) and the Genome Analysis Toolkit. Variant calling was performed using an in-house developed pipeline. Variants identified as germline variants were excluded, while single nucleotide variants (SNVs) and indels with allelic fractions of more than 5% and supported by more than 4 unique reads, amplification with a fold-change greater than 2.5 in more than 25% of regions covered, and gene fusions supported by more than 3 unique reads were included.

TMB was calculated using the total number of nonsynonymous mutations divided by the total genomic target region (2.13 Mb). MSI status was determined using MSIsensor from paired tumor-normal targeted sequence data, and 309 MSI sites were included in the panel of 825 cancer-related genes. An MSIsensor score below 10 defines microsatellite stability (MSS) status, while that above 50 defines MSI status. The prevalence of gene fusions involving a targetable gene and driver mutations was compared with the OrigiMed2020 and TCGA cohorts[21]. Clinicopathological and genomic data were retrieved from the cBioPortal (https://www.cbioportal.org).

Fluorescence in situ hybridization

FFPE tissue sections (5 μm) were prepared on positively charged slides. After deparaffinizing and rehydrating, the slides were incubated with prewarmed 8% sodium thiocyanate in dH₂O at 80 °C and incubated for 30 min. FGFR2 (10q26) or BRAF (7q34) break-apart probes were placed on the slide, covered with a glass coverslip, and sealed with rubber cement. Hybridization was performed overnight at 37 °C. The slides were washed twice in 50% formamide at 47 °C for 2 min and then twice in 2X standard saline citrate at room temperature for 2 min. Nuclei were stained with DAPI as a counterstain. The slides were scanned using a 90i Nikon fluorescent microscope. For each probe, 200 nuclei were evaluated. The 5′ (red) and 3′ (green) signals separated by ≥ 2 signal diameters were considered split as positive.

Statistical analysis

All statistical analyses were performed using SPSS 24.0 software (IBM, Chicago, IL, United States). χ² or Fisher’s exact test was used to analyze the association between fusion alterations and driver mutations. A P value of < 0.05 was considered statistically significant.

RESULTS

Clinical characteristics of patients

We retrospectively analyzed 1341 Chinese patients with GC who underwent genetic analysis from multiple centers in China. Of these, 387 patients were excluded because gene fusion detection was not performed with the Onco PanScan panel using tumor tissue samples (Figure 1). Gene fusion events were detected in 20 patients; however, 4 patients without any gene fusions involving targetable genes were excluded. Finally, 16 patients with 20 fusion events involving targetable genes were included for further analysis. The clinical characteristics of 954 patients with GC are shown in Table 1. Of these patients, 310 (32.56%) were women and 644 (67.44%) were men, with a median age of 57 and 62, respectively, at diagnosis. There was no significant difference between targetable gene fusion-positive and -negative patients in age (P = 0.293), sex (P = 0.463), MSI status (P = 0.551), or TMB (P = 0.217) (Table 1).

Figure 1 Flowchart of patient selection. GC: Gastric cancer.

Table 1 Clinical characteristics in targetable gene fusion-positive and -negative patients.

Variables	Total, n	Fusion involving targetable genes		P value
		Positive, n (%)	Negative, n (%)
Sex				0.293
Female	310	3 (0.97)	307 (99.03)
Male	644	13 (2.02)	631 (97.98)
Age, yr				0.463
≤ 60	451	6 (1.56)	445 (98.44)
> 60	503	10 (1.98)	493 (98.01)
MSI status				0.551
MSI-H	46	1 (2.17)	45 (97.93)
MSS	908	15 (1.65)	893 (98.35)
TMB				0.217
Median TMB score	2.92	5.63	2.83

Age, sex, microsatellite instability status, and tumor mutational burden between fusion-positive and -negative patients were compared. The one-tailed P value for Fisher’s exact test was calculated. MSI-H: Microsatellite instability-high; MSS: Microsatellite stability; TMB: Tumor mutational burden.

The landscape of gene fusions involving targetable genes in Chinese patients with GC

To gain insight into fusion events in GC, we evaluated 954 patients with GC undergoing gene fusion analysis. In total, 20 patients harbored 24 gene fusions, 2 patients had double fusions (patient 01 and 02), and 1 patient (09) harbored triple fusions. RARA fusions (5/24, 17.8%) and FGFR family gene fusions (5/24, 17.8%) occurred most frequently in the cohort, followed by BRAF (3/24, 10.7%) and MET (2/24, 8.3%) (Figure 2A). ALK, RET, NTRK2, NRG1, and EGFR fusions were identified in one patient each. Remarkably, 20 of 24 (83.3%) fusions involved targetable genes (Table 2). RARA has been frequently reported as a 3′ fusion partner in acute promyelocytic leukemia[22]. RARA was identified as a 5′ fusion partner in 4 patients and as a 3′ fusion partner in 1 patient; however, only the KRPAT9-RARA fusion was detected in patient 01 as the 3′ fusion partner including exons 3-9, which encodes a DNA-binding and a ligand-binding domain required for RARA transcription factor activity[22]. Three BRAF fusions were identified in patient 09 as the 3′ fusion partner containing the complete tyrosine kinase domain, which was coded by exons 11-18. All FGFR2 and FGFR3 fusions were detected as 5′ fusion partners. Four FGFR2 fusions were consistent with other known activating FGFR2 fusions[23], which frequently occur with a breakpoint after exon 17 at the 3′ end of FGFR2 with a 3′ fusion partner. The kinase domain was retained in these fusion genes. In patient 10, the MET fusion involved the 5′ end of MET exon 7, thus retaining an intact MET kinase domain.

Figure 2 Profile of targetable gene fusions in gastric cancer. A: The types and proportion of 24 gene fusions. Others included targetable ALK, RET, NTRK2, NRG1, and EGFR fusions. Four fusions without targetable genes were excluded from the analysis; B: Comparison of gene fusion frequencies in our cohort and the OrigiMed2020 and The Cancer Genome Atlas (TCGA) cohorts. No statistical differences were found among the cohorts.

Table 2 List of gene fusions involving targetable genes in Chinese patients with gastric cancer and drugs under clinical trial or approved by the Food and Drug Administration.

Patients ID	Fusion gene	5' partner gene				3' partner gene				Variant frequency, %	Functional domain is included or not	Targeted drugs
		Gene name	Chromosome	Last observed exon	Breakpoint	Gene name	Chromosome	First observed exon	Breakpoint
Patient 01	RARA-PGAP3	RARA	17	3	38504951	PGAP3	17	8	37828020	24.1	Partially include	Tamibarotene targeting RARA fusion2
Patient 01	KRTAP9-7-RARA	KRTAP9-7	17	downstream	39437039	RARA	17	3	38499547	56.9	Completely include	Tamibarotene targeting RARA fusion2
Patient 02	RARA-KRT13	RARA	17	2	38491648	KRT13	17	8	39657269	29.2	Partially include	Tamibarotene targeting RARA fusion2
Patient 02	RARA-ETV4	RARA	17	2	38499726	ETV4	17	5	41621243	76	Partially include	Tamibarotene targeting RARA fusion2
Patient 03	RARA-IKZF3	RARA	17	2	38504120	IKZF3	17	2	38009555	18.4	Partially include	Tamibarotene targeting RARA fusion2
Patient 04	FGFR2-PDE2A	FGFR2	10	17	123241248	PDE2A	11	7	72307251	1.4	Completely include	Pemigatinib; Erdafitinib targeting FGFR fusion1
Patient 05	FGFR2-intergenic	FGFR2	10	17	123242196	intergenic	10	-	123394107	16.6	Completely include	Pemigatinib; Erdafitinib targeting FGFR fusion1
Patient 06	FGFR2-intergenic	FGFR2	10	17	123240841	intergenic	10	-	122793842	4.2	Completely include	Pemigatinib; Erdafitinib targeting FGFR fusion1
Patient 07	FGFR2-SHTN1	FGFR2	10	17	123242528	SHTN1	10	6	118709305	5.1	Completely include	Pemigatinib; Erdafitinib targeting FGFR fusion1
Patient 08	FGFR3-PHTF2	FGFR3	4	18	1808927	PHTF2	7	11	77567982	3.3	Completely include	Pemigatinib; Erdafitinib targeting FGFR fusion1
Patient 09	STIM2-BRAF	STIM2	4	11	27012641	BRAF	7	9	140487929	12.7	Completely include	Selumetinib targeting BRAF fusion1
Patient 09	STIM2-BRAF	STIM2	4	11	27013243	BRAF	7	10	140486103	1.1	Completely include	Selumetinib targeting BRAF fusion1
Patient 09	TBC1D19-BRAF	TBC1D19	4	4	26629603	BRAF	7	10	140486782	6.5	Completely include	Selumetinib targeting BRAF fusion1
Patient 10	TES-MET	TES	7	1	115867013	MET	7	2	116332227	0.7	Completely include	Crizotinib targeting MET fusion1
Patient 11	MET-TES	MET	7	21	116436166	TES	7	4	115889445	24.6	Not include	Crizotinib targeting MET fusion1
Patient 12	EML4-ALK	EML4	2	6	29447382	ALK	2	20	42498662	3.5	Completely include	Crizotinib; ceritinib targeting ALK fusion1
Patient 13	OPALIN-RET	OPALIN	10	6	98104545	RET	10	11	43610099	5.46	Completely include	Pralsetinib targeting RET fusion1
Patient 14	ARHGAP10-NTRK2	ARHGAP10	4	1	148716754	NTRK2	9	16	87476645	15.3	Completely include	Larotrectinib targeting NTRK2 fusion1
Patient 15	NRG1-FDFT1	NRG1	8	12	32617907	FDFT1	8	8	11685375	8.9	Partially include	MCLA-128 targeting NRG1 fusion2
Patient 16	EGFR-SEPTIN14	EGFR	7	25	55269173	SEPTIN14	7	10	55871179	9.6	Completely include	Afatinib targeting EGFR fusion2

¹FDA-approved drugs targeting gene fusions.

²Drugs targeting gene fusions are under clinical trials.

The frequency of fusion events involving the abovementioned 10 targetable genes in the TCGA GC cohort and another Chinese GC cohort (OrigiMed2020 cohort) were analyzed and compared with our patient data (Figure 2B). Neither our cohort nor the OrigiMed2020 cohort showed significant differences in the incidence of these gene fusions in Chinese patients. In two Chinese cohorts, EGFR fusions occurred less frequently. Fusions in MET, BRAF, RET, ALK, and NTRK2 were only identified in two Chinese cohorts; however, the differences in the incidence of these genes were not statistically significant.

Novel fusions involving targetable genes in GC

In total, 2 of 20 fusions involving the targetable genes, EML4-ALK and EGFR-SEPTIN14, were reported in other cancers, including non-small-cell lung cancer[24-26]. The remaining 18 gene fusions were not reported in any cancer types. In total, 13 of 18 novel gene fusions contained the exon encoding a tyrosine kinase domain, such as exons 11-17 of FGFR2, exons 11-18 of BRAF, exons 12-19 of MET, and exons 16-21 of NTRK2 (Figure 3). All fusions involving FGFR2, BRAF, RET, and NTRK2 retained the kinase domain (Table 2). Furthermore, reads in Integrative Genomics Viewer plots supported these gene fusions. To verify these novel fusions, fluorescence in situ hybridization (FISH) was performed using break-apart probes. Because only two tumor tissue samples were available, only the FGFR2 and BRAF arrangement in patient 04 and 09, respectively, were confirmed by FISH.

Figure 3 Examples of novel gene fusions involving targetable genes in gastric cancer. A-D: Schematic representation and Integrative Genomics Viewer screenshot of FGFR2-PED2A (A), STIM-BRAF (B), OPALIN-RET (C), and NTRK2-ARHGAP10 (D) are shown; A and B: FGFR2 and BRAF fusions were confirmed by fluorescence in situ hybridization using FGFR2 (10q26) or BRAF (7q34) break-apart probes. Red spot: 5′ Probe signal; Green spot: 3′ probe signal; Yellow spot: Target gene without rearrangement. Arrows indicate the cells with separate 5′ (red) and 3′ (green) signals. Bar: 100 μm. TK: Tyrosine kinase domain.

Gene fusions are enriched in patients with ERBB2 amplification but not in those with high MSI and TMB

Because of the low frequency of gene fusions in patients with GC, we determined whether gene fusions are enriched in different molecular subtypes of GC, which may indicate the patients that could benefit from gene fusion detection. Fusions are mutually exclusive with other oncogenic mutations and are enriched in patients without driver mutations[27-29]. In our cohort, the frequency of genetic alterations in oncogenic driver genes of GC, such as TP53, ARID1A, CDH1, and PIK3CA mutations and ERBB2 amplification, were comparable with those in the TCGA cohort (Supplementary Figure 1). There was no significant difference in the frequency of fusions involving targetable genes between patients with any alterations in all five driver genes and those without (Figure 4A). Notably, the fusion alteration frequency was significantly higher in patients with ERBB2 amplification than in those without ERBB2 amplification (Figure 4B, P = 0.01). To determine whether fusion alterations were enriched in other driver genes, TP53, ARID1A, CDH1, and PIK3CA were analyzed. There was no enrichment in fusion alterations for these genes (Supplementary Figure 2). Forty-six patients had the MSI-H phenotype. Of these, one patient with fusion genes exhibited MSI-H. There was no obvious difference in the incidence of gene fusions between patients with MSI-H and MSS (Figure 4C). Similarly, TMB scores were evaluated in targetable gene fusion-positive and -negative patients, but the results were not statistically significant (Figure 4D).

Figure 4 Enrichment of gene fusions in patients with gastric cancer with driver alterations. A: The incidence of gene fusions in patients with and without driver alterations were analyzed, P > 0.05; B: The incidence of gene fusions in patients with and without ERBB2 amplifications were analyzed, P < 0.05; C: The incidence of gene fusions in patients with microsatellite instability-high and microsatellite stability were analyzed, P > 0.05; D: Tumor mutational burden in targetable gene fusion-positive and -negative patients was compared, P > 0.05.

DISCUSSION

Structural gene rearrangements leading to gene fusions are common events that occur in solid tumors. Gene fusions have been considered oncogenic drivers in neoplasia for more than 30 years[30]. Detection and characterization of gene fusions is important for clinical purposes[31]. As the first large-scale study focusing on gene fusion events in Chinese patients with GC, we retrospectively analyzed 954 tumor specimens to identify fusions involving targetable genes and confirmed the occurrence of these fusions in GC.

In this study, 16 of 954 patients harbored 20 fusions involving targetable genes, the majority of which had not been previously reported, including FGFR2-PDE2A, STIM2-BRAF, OPALIN-RET, and ARHGAP10-NTRK2. However, we did not find any significant differences between the Chinese GC cohort (our cohort and OrigiMed2020 cohort) and the TCGA cohort. Fusions in BRAF, RET, ALK, and NTRK2 were detected in two Chinese cohorts but not in the TCGA cohort. This finding may have resulted from the small size of the TCGA cohort, which is prone to bias for gene fusions events because of the low occurrence rate in GC. A comparative study with a larger population is needed to identify differences in fusions involving targetable genes between races.

A major contribution of gene fusions to patients with tumor is the development of drugs that target fusion proteins encoded by these genes. The majority of advances in targeting gene fusions involve kinase domains that constitutively activate downstream signaling pathways[32]. In this study, except RARA and NRG1 fusions, the 14 other fusions involving targetable genes included a receptor tyrosine kinase (RTK) gene, such as FGFR2/3, BRAF, MET, ALK, RET, NTRK2, and EGFR. Furthermore, most of all RTK gene fusions (13/14) completely retained the tyrosine kinase domain, which resulted in functional fusion proteins. We only verified the BRAF rearrangement in patient 09 and the FGFR2 rearrangement in patient 04 using FISH because of insufficient tumor specimens. These fusions were consistent with previously observed fusions[23]; however, only 1 out of 5 RARA fusions contained exons 3-9, which encodes a DNA-binding and ligand-binding domain, which are required for RARA transcription factor activity. These results indicate that most patients with GC with fusions involving targetable genes may benefit from drugs that target fusions. However, patients in this retrospective study had not received targeted drug treatment; thus, we cannot determine whether they would have benefited from fusion-targeted drug therapy.

Interestingly, we also discovered 18 novel fusions with unreported partner genes or with an intergenic space. In other words, screening for known fusions in GC by FISH or polymerase chain reaction will likely miss most of the gene fusions that involve targetable genes. This is not conducive to patients with GC participating in clinical trials of fusion-targeted drugs in pan-cancer. Additionally, we found gene fusions enriched in patients with ERBB2 amplification. We did not confirm all fusions using FISH because of limited tumor tissue samples, nor could we identify gene fusions enriched in distinct molecular subtypes of GC. Moreover, the efficacy of fusion-targeted drugs in GC remains to be further validated in clinical trials. Despite these limitations, for patients who fail standard therapy, NGS-based novel gene fusion detection may provide a new treatment strategy and facilitate participation into clinical trials involving targeted therapy.

CONCLUSION

As the first large-scale study focusing on gene fusion events in Chinese patients with GC, we determined the frequency (16/954) of targetable gene fusions, and the majority of these fusions, including TES-MET, FGFR2-PDE2A, OPALIN-RET, STIM-BRAF, ARHGAP10-NTRK2, and EGFR-SEPTIN14, had not been previously described. These novel fusions completely retain a kinase domain. Additionally, we found gene fusions that were enriched in patients with ERBB2 amplification. Gene fusion detection may aid in the development of novel treatment strategies for patients with GC.

ARTICLE HIGHLIGHTS

Research background

With rapid advancements in oncogenomics, increasing attention has been focused on gene fusions in cancer. The Food and Drug Administration has approved several fusion-targeted drugs for the treatment of solid tumors, such as larotrectinib for NTRK fusion-positive cancers and Zenocutuzumab for NRG1 fusion-positive cancers. However, targetable gene fusions in Chinese patients with gastric cancer (GC) have not been well characterized.

Research motivation

To investigate the incidence of gene fusions involving targetable genes in Chinese patients with GC and explore a potential treatment strategy for patients with GC.

Research objectives

To explore the types and proportion of targetable gene fusions in Chinese patients with GC and determine the distribution of patients with gene fusions among the molecular subtypes of GC.

Research methods

This was a multicenter retrospective study that evaluated patients with GC. A total of 954 tumor tissue samples from patients with GC who underwent gene fusion detection were included. Genetic alterations, including SNVs, indels, amplifications, and gene fusions, were analyzed. The enrichment of gene fusions in the molecular subtypes of GC was explored.

Research results

Twenty fusions involving targetable genes were detected. Among them, 18 novel gene fusion events were previously not reported in other cancers. Owing to a limited number of tumor tissue samples, only BRAF and FGFR2 fusions were identified by fluorescence in situ hybridization. Additionally, we found that gene fusions were enriched in patients with ERBB2 amplification.

Research conclusions

Gene fusions involving targetable genes were characterized in Chinese patients with GC. Testing gene fusions may provide insight for the treatment of GC.

Research perspectives

A large study should be performed to further confirm the targetable gene fusions and identify whether gene fusions are enriched in distinct molecular subtypes of GC.