RNA-Based Next-Generation Sequencing Approach to Non-Small Cell Lung Cancer: A Single-Center Experience in Turkey

Lung cancer, a heterogeneous disease with histopathological, genetic, and clinical features, continues to be the leading cause of cancer-related death worldwide. It consists mainly of two subtypes, namely, non-small cell lung cancer (NSCLC) and small cell lung cancer [1]. Identification of an increasing number of targetable molecular changes in NSCLC has demonstrated the importance of somatic molecular genetic analysis [2]. According to the literature, it is important to detect specific variants associated with EGFR, ALK, ROS1, BRAF, MET, RET, NTRK1/2/3, KRAS, HER2, and PD-L1 [3]. Next-generation sequencing (NGS) has become a common comprehensive method to analyze multiple genes, thereby guiding targeted treatment in recent years [4]. In this retrospective study, we screened for genetic alterations (GAs) in formalin-fixed paraffin-embedded samples collected from patients diagnosed with NSCLC using a targeted NGS panel consisting of 36 genes. Our analysis provides a comprehensive molecular characterization of NSCLC, including information on fusions, single-nucleotide variations (SNVs), and insertion-deletion mutations (InDels). We believe that this broad molecular perspective on NSCLC provided by our panel offers valuable insights into the disease.

The study utilized tissue samples from paraffin blocks obtained from patients diagnosed with NSCLC at the Division of Medical Genetics, Bursa Yüksek İhtisas Training and Research Hospital, University of Health Sciences, during 2018–2022. Archer^® FusionPlex Comprehensive Thyroid and Lung (CTL) Kit is an NGS assay kit that aims to detect and identify fusions, point variants, and expression levels of 36 genes associated with lung and thyroid cancers (Table 1). This kit uses Archer’s proprietary Anchored Multiplex PCR (AMP^™)-based enrichment to detect fusions of all target genes in a single sequencing test even without prior knowledge of the known fusion partners or breakpoints. The method uses unidirectional gene-specific primers for both the known and unknown variants. Adapters containing both the molecular barcode and patient index are used.

For the amplification of disease-associated genes, gene regions, or fusions, RNA isolation (RNeasy Formalin-Fixed Paraffin-Embedded Kit, Qiagen, Hilden, Germany) was first performed on the primary sample. The quantity and integrity of isolated RNA samples were analyzed using QuBit (Thermo Fisher Scientific). RNA samples with a minimum concentration of 10 ng were included in the study. The RNA from the samples was converted to cDNA using a reverse transcriptase enzyme and the First Strand cDNA Synthesis Kit. Subsequently, 1 µL of the cDNA sample was loaded into a real-time PCR device (Bio-Rad^®) and samples with a QC value between 22 and 28 were selected. Target regions were amplified using the RNA-based Archer^® FusionPlex CTL Kit and sequenced on the Illumina NextSeq platform using the NGS method.

Secondary analyses (data cleaning, alignment, and variant/fusion detection) of data were performed using Archer analysis software. The human reference genome hg19 (GRCh37) was used for alignment. Archer analysis software uses molecular barcodes for deduplication and error correction for accurate variant/fusion and quantitative multiplex data analysis (lists and versions of software used for analysis can be found at “analysis.archerdx.com/about_archer”).

Raw data transferred to the Archer analysis platform were primarily evaluated for quality. Fusion QC values (threshold ≥10) and variant QC values (threshold ≥10) were used as criteria for evaluation. Subsequently, suitable samples were used in the analysis and assessed in terms of the specified disease.

Changes related to the specified tumor type were filtered by considering the patient’s clinical information. These changes were then classified according to their availability in Association for Molecular Pathology/American Society of Clinical Oncology/College of American Pathologists (AMP/ASCO/CAP) guidelines [5] and databases, as well as according to information such as diagnosis, prognosis, and treatment.

The RNA-based NGS panel was selected for this study as it provides a broad molecular perspective on NSCLC and offers comprehensive insights into identifying candidates suitable for targeted therapies. Moreover, the Archer^® FusionPlex CTL Kit is recommended by the manufacturer for research purposes only. Hence, to ensure reliability and accuracy, significant GAs were validated through a DNA-based secondary test specific to the type of alteration. For this purpose, real-time polymerase chain reaction was performed (Rotor-Gene Q Series, Qiagen Ltd) to detect frequent variants in the EGFR gene (Diatech Pharmacogenetics, Jesi, Italy). Specific primers were designed using primer 3 for identifying rare variants in the EGFR gene and other gene regions. Subsequently, the regions sequenced using capillary electrophoresis (Applied Biosystems^® 3500 Genetic Analyzer) were analyzed with the CLC Main Workbench v8.1 software (Qiagen). Distinct from the previously mentioned methods, fluorescence in situ hybridization (FISH) was employed (Allegro Plus, BioView) to confirm the fusions of ALK, ROS1, and RET genes (Abbott Molecular, Des Plaines, IL, USA). Due to limitations, including insufficient samples, and fusions without FISH probe, changes that could not be confirmed and did not provide strong evidence for analysis were not included in the study, thereby excluding possible false-positive GAs.

The NGS data from 462 patients with NSCLC were included in this study. The age range of the patients was quite wide, and more than half of the patients were over 63 years old. There were more male than female patients. The majority of the tested samples were identified as adenocarcinomas based on their distinct histopathological characteristics. However, subgroup information was not available for 149 samples. This high number is due to the referral of numerous patients from various external centers to our laboratory for molecular analysis, often without accompanying subgroup classification data. In some of these cases, molecular tests were prioritized to facilitate timely therapeutic planning, and analyses were conducted while subgroup determination was still ongoing. The others were squamous cell carcinomas, large cell carcinomas, and mixed types (Table 2). Furthermore, SNV/InDel, rearrangements, and expression levels were determined. However, in this study, expression-level results were not used because of some limitations (as described in the Discussion section). GAs were detected in approximately 45% (208/462) of patients. Regarding variant frequency, SNV/indels were the first to be observed, followed by rearrangements of genes (fusions, exon skipping, and isoform formations), and the rare coexistence of changes observed in different gene regions. The most common mutated genes in this study were KRAS and EGFR. Additionally, the most frequently detected SNVs were in codon 12 of the KRAS gene, and the most common among these was G12C. EGFR exon 19 deletion was observed with the second highest frequency among all the GAs. The T790M variant was also detected in 3 patients with EGFR exon 19 deletion and in 2 patients with the L858R variant. These 5 patients had not previously received targeted therapy. The frequencies of variations associated with BRAF, FGFR3, NRAS, AKT1, ERBB2, HRAS, PIK3CA, CTNNB1, IDH2, MET, and RET were low (Table 3). Rearrangements of RET, MET, FGFR3, FGFR1, and SYN2 were also detected. However, ALK and ROS1 were highly frequent (Table 4). Notably, 10 patients had complex GAs (Table 5). All changes detected in patients were tumor tissue-specific somatic variants. In one of our patients, the GNAS (R201G) variant observed with KRAS (G12V) variant was mosaic and somatic, independent of the tumor tissue. A different patient with KRAS (G12V) had a concomitant ALK rearrangement and EGFR isoform formation.

This study presents the results of molecular genetic analysis performed using NGS on samples with known histopathology compatible with NSCLC. Considering that most of the patients were referred from different centers, information about smoking status could not be obtained; therefore, this information was not included. In addition, most of the samples were examined histopathologically in different centers, and subtypes could not be determined in 149 patients with NSCLC.

Previous studies have shown that lung cancer incidence is high among male patients in Turkey, with a median age of 61 years [7]. In our study, most patients were male patients, and the median age was 63 years.

NSCLC constitutes around 80–85% of all pulmonary system cancers, with lung adenocarcinoma being the most frequent subtype due to its complex genomic profile [8]. Although NSCLC samples that were not histopathologically subtyped were included in this study, most of the known subgroups were adenocarcinoma, thus supporting previous reports.

Researchers have previously reported oncogenic genomic alterations in NSCLC, considering the ethnicity, age group, and smoking status of the population studied [9]. Recently, it has been estimated that up to 69% of advanced-stage NSCLC cases have targeted variants in numerous genes [10]. In the present study, the rate was approximately 45%. This proportional difference in our study supports the view that NSCLC-associated GAs can vary significantly among different geographic regions, populations, and individuals with different characteristics or habits [6]. Based on variant frequency, EGFR and KRAS came first. KRAS variant rates are high in non-Asians, smokers, and older populations diagnosed with lung adenocarcinoma [6]. Although not considered for this study, the rate of smoking is significantly high in Turkey (https://www.tuik.gov.tr/). Consistent with literature, the frequency of KRAS variants was high in our study group, which was mostly observed in adenocarcinoma and elderly individuals. Indels within exon 19 and L858R in exon 21 are the most common EGFR variants [11]. The most common KRAS variants are G12C, G12D, and G12V, which are located on codon 12 [12]. Similarly, in the present study, exon 19 deletion and L858R were the most common EGFR variants, and the changes in codon 12 were the most common KRAS variants. Variants detected in other genes (BRAF, FGFR3, NRAS, AKT1, ERBB2, HRAS, PIK3CA, CTNNB1, IDH2, MET, and RET) were less frequent, and all of these have been reported in the Catalogue Of Somatic Mutations In Cancer (COSMIC; https://cancer.sanger.ac.uk/cosmic) database (Table 6).

EGFR (T790M) is often expressed as a resistance variant after targeted therapy. Few studies have been reported on the frequency of this variants in pretreatment cases. Li et al. [13] observed a very low incidence of T790M before treatment. Ye et al. [14] stated that the rate is high when low-level pretreatment T790M variants in NSCLC are considered significant using sensitive methods. In our study, the T790M variant was observed in 5 patients who did not receive targeted therapy. Three of these patients also had an exon 19 deletion, whereas the other 2 patients had the L858R variant.

The NGS assay kit used in the study provided information on known classical variants as well as fusion, rearrangements, and expressions. ALK, ROS1, and RET fusions were confirmed using FISH, and no discordance was observed. EML4-ALK fusions were the most common, which is consistent with previous reports [15]. However, the rare GAs shown in Table 4 could not be confirmed. Amplification of HER2 and MET genes is important for therapy [3]. The Archer analysis program uses a light-to-dark color scale for each gene that constitutes the panel. This scale provides information on gene expression levels; high expression levels are expected for dark-colored genes. HER2 analysis was performed on tissue samples using FISH. However, contrary to expectations, HER2 amplification was observed in patient samples with light-color scales. Notably, it is beneficial to confirm the results with a method such as FISH to ensure that light-colored genes are definitely negative in terms of expression (online suppl. material; for all online suppl. material, see https://doi.org/10.1159/000544697). With the constant development of NGS and kit technology, this process is expected to become well known.

We have observed complex GAs involving at least two different variations in very few patients. Different KRAS subtype variants can show different patterns of co-occurrence [16]. As shown in Table 5, KRAS codon 12 and 13 variants were detected together with variants in different genes in 8 out of 10 patients. Interestingly, in a 66-year-old male patient with visual impairment, hyperthyroidism, and colorectal polyps, the KRAS (G12V) variant coexisted with the GNAS (R201G) variant, which is significant in McCune-Albright syndrome (MIM:174800). The association of the GNAS (R201G) variant, which is rarely found in McCune-Albright cases, with different gene variants in breast cancer cases has been reported [17, 18]. However, to our knowledge, this study is the first to report the association of the GNAS (R201G) variant with the KRAS (G12V) variant in NSCLC.

Concomitant ALK rearrangements with EGFR variants are infrequent in NSCLC [19]. In our study, the association between ALK rearrangements, EGFR isoform formation, and A289V changes was demonstrated, which has not been previously reported. Similarly, we did not find cases in the literature in which AKT1 (E17K) and MAP2K1 (K57N) were observed together.

Recently, studies conducted on smaller cohorts with methodologies similar to ours have been reported in the literature. For instance, in a study by Kulda et al. [20] involving 237 samples, a mutation rate of 57% was identified, while our study demonstrated a slightly lower mutation rate of 45%. Similarly, in a methodologically comparable study by Eser et al. [21] on a cohort of 89 samples, GAs were most frequently detected in the KRAS and EGFR genes, consistent with our findings. KRAS mutation rates were similar between the two studies, with 20.2% reported in their study and 19.48% in ours. However, EGFR mutations were detected in 6.74% of the samples in their study, whereas our study showed a higher proportion of 12.55%.

The success rate of RNA-based NGS for somatic mutation detection has been reported to be relatively low in the literature, with variability depending on sample type, RNA quality, and pre-analytical factors [22]. In this study, we focused on the analysis results and did not conduct a pre-analytical evaluation. For failed RNA-based NGS cases, additional tissue samples were used if available, and in the absence of extra tissue, new biopsies were performed. However, samples that did not yield results despite these efforts were excluded to avoid selection bias. This approach ensured the inclusion of high-quality samples with well-preserved RNA, thereby improving the reliability of molecular profiling.

In this large-patient series from a single center, we aimed to approach NSCLC, a histopathologically and molecularly heterogeneous group of diseases, from a broad perspective. In routine practice, target regions of specific genes are analyzed using different techniques. Often, the aim is to identify changes appropriate for targeted therapy. The assay kit used for this study included 36 genes and was RNA-based; thus, it was informative for the different types of GAs we expected to observe. Therefore, we were able to detect extremely rare fusions and the coexistence of different gene variants. In conclusion, although there were certain limitations, this study is expected to contribute to the understanding of the molecular profile of NSCLC.

The authors thank the patients and their relatives for their participation in the study.

This study protocol consists of a retrospective evaluation of patient data. All participants provided written informed consent. It was reviewed and approved by the Bursa Yüksek İhtisas Training and Research Hospital Ethics Committee (Approval No. 2011-KAEK-25 2019/08-01 and 2022/11-07). The researchers acted ethically by the Declaration of the World Medical Association of Helsinki.

The authors have no conflicts of interest to declare.

No funding was received for this study.

Or.G. was primarily responsible for the conception and design of the study, data collection, and analysis, as well as drafting and critically revising the manuscript. Oz.G. contributed to the interpretation of the data and provided substantial assistance in the preparation of the manuscript. A.T. contributed technical expertise and supported the critical review of the manuscript. All authors have approved the final version of the manuscript and took responsibility for the integrity and accuracy of the work.

All data generated or analyzed during this study are included in this article and its online supplementary material files. Further inquiries can be directed to the corresponding author.