A total of 64 patients with NSCLC from the Department of Thoracic Surgery at The First People's Hospital of Ping Ding Shan (Pingdingshan, China) were recruited between October 2019 and October 2021. The pathological diagnosis was confirmed by three separate pulmonary pathologists (XZ, HS and YW) using The Fourth Edition of the World Health Organization Classification of Lung Tumors (32). The NGS data were acquired from 64 formalin-fixed and paraffin-embedded (FFPE) tumor specimens. The tumor tissue specimens were fixed with 10% neutral buffered formalin at room temperature (~25°C) for 24–48 h immediately after surgical removal. The present study adhered to the Code of Ethics established by the World Medical Association (Declaration of Helsinki) (33), and was approved by the Medical Ethics Committee of First People's Hospital of Ping Ding Shan (approval no. PYLL20190806; Pingdingshan, China).

The inclusion criteria were: i) Age >18 years; ii) an initial diagnosis of NSCLC by histological examination; iii) an enhanced CT scan of the chest and abdomen, MRI of the brain, and whole-body bone scan (emission CT) results were available; iv) diagnosed primary NSCLC; and v) no history of tumor treatment.

The exclusion criteria were: i) Patients had been subjected to tumor treatments, including chemotherapy, radiotherapy, targeted therapy or immunotherapy; ii) history of exposure to trace elements, toxic elements or heavy metals; iii) patients used antioxidants, vitamins or dietary supplements; iv) patients had undergone surgery within the past year; v) patients had comorbidities, such as autoimmune disease, gout, chronic liver disease, chronic kidney disease, protein-energy malnutrition, thyroid disease or vitamin A/D deficiency; or vi) patients had additional comorbidities deemed inappropriate for the present study by the research team, such as uncontrolled diabetes, psychiatric disorders, and substance abuse.

To be diagnosed with hypertension, patients must have met at least one of the following three criteria: i) Affirmative response from the patient to the question of whether they had a prior diagnosis of hypertension from a medical professional; ii) individuals exhibiting a systolic blood pressure (SBP) of ≥140 mmHg and/or diastolic blood pressure (DBP) of ≥90 mmHg; and iii) individuals currently using blood pressure-lowering medications, including angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, diuretics and beta-blockers. Hypertensive participants were categorized into three hypertension grades based on their blood pressure readings: i) Grade 1, 140≤ SBP <160 mmHg and/or 90≤ DBP <100 mmHg; ii) grade 2, 160≤ SBP <180 mmHg and/or 100≤ DBP <110 mmHg; and iii) grade 3, SBP ≥180 mmHg and/or DBP ≥110 mmHg (34). In addition, the categorized criteria of patients with optimal, normal and high normal blood pressure readings were as follows: Optimal, SBP <120 mmHg and DBP <80 mmHg; normal, 120≤ SBP <130 mmHg and/or 80≤ DBP <85 mmHg; high normal, 130≤ SBP <140 mmHg and/or 85≤ DBP <90 mmHg.

The extraction of genomic DNA (gDNA) from the FFPE tumor specimens was performed using a GeneRead DNA FFPE Kit (Qiagen GmbH). gDNA quantity and purity were assessed using a Qubit^® 3.0 Fluorometer (Invitrogen; Thermo Fisher Scientific, Inc.) and a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Inc.). The Agilent 2100 Bioanalyzer instrument (Agilent Technologies, Inc.) was used to evaluate DNA integrity using High Sensitivity DNA Reagent (Agilent Technologies, Inc.).

Firstly, 300 ng gDNA from each FFPE tumor specimen was mechanically fragmented by an E220 focused ultrasonicator Covaris (Covaris, LLC) at 75W peak incident power, 20% duty factor, 1000 cycles per burst and a 7°C water bath temperature for 95 sec treatment time. The DNA fragments were broken into fragments of 150–200 bp. Secondly, the KAPA library preparation kit (cat. No. KK8500; Kapa Biosystems; Roche Diagnostics) was employed to construct libraries using 10–100 ng gDNA fragments. Finally, the xGen Lockdown Probe pool (Integrated DNA Technologies, Inc.) was used to capture the NGS libraries, followed by amplification of the captured DNA fragments using 1× KAPA HiFi Hot Start Ready Mix (Kapa Biosystems; Roche Diagnostics). The sequences of the forward and reverse primers were 5′-AATGATACGGCGACCACCGAGATCTACAC-3′ (Illumina P5 primer) and 5′-CAAGCAGAAGACGGCATACGAGAT-3′ (Illumina P7 primer). The thermal cycling protocol was as follows: Initial denaturation step at 98°C for 45 sec, followed by 13 cycles of 15 sec at 98°C, 30 sec at 60°C and 30 sec at 72°C. The final extension was conducted at 72°C for 60 sec. The fragment length of library amplification products was analyzed by Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.) and High Sensitivity DNA Kit (cat. No. 5067-4626; Agilent Technologies, Inc.). The library concentrations were determined by Qubit^® 3.0 Fluorometer (Invitrogen; Thermo Fisher Scientific, Inc.) and the final library concentration for sequencing was 1.8 pM. Illumina NextSeq CN500 platform (Illumina Inc.) and NextSeq CN500 Mid Output v2 Kit (150 cycles) (cat. no. R0151; Illumina Inc.) were adopted to perform sequence with a 75 bp pair end mode.

Fastp software (v.0.23.2; OpenGene Bioinformatics) was applied to remove low-quality reads [the proportion of low-quality bases (quality below Q15) in a read exceeds 40%], resulting in clean data. Burrows-Wheeler-Alignment Tool (BWA v.0.7.12; Wellcome Trust Sanger Institute) was used to align all filtered reads using the human genome as the reference (University of California Santa Cruz ID: hg19). Subsequently, Picard tools (v.1.130; Broad Institute) and Genome Analysis Toolkit (v.3.2.2; Broad Institute) were used to remove duplicates, perform indel realignment and recalibrate the base quality scores (35,36). The CollectTargetedPcrMetrics tool of Genome Analysis Toolkit was employed to generate quality statistics. Finally, VarDict (v.1.6.0; GitHub, Inc.) was used to verify single nucleotide variations and insertions/deletions (37).

ANNOVAR (v. 20210202; http://annovar.openbioinformatics.org/en/latest/) software was used to annotate somatic mutations (38). The identification of somatic mutation candidates was performed using the following filtering criteria: i) Removal of the variants coverage depth (VDP) <10; ii) variant sites with a mutant allele frequency (MAF) >0.001 in the 1,000 Genomes databases were removed (1,000 Genomes Project Consortium; http://www.internationalgenome.org/); iii) variant sites with a MAF ≥0.001 and <0.1 in the 1,000 Genomes databases, accompanied by COSMIC (http://cancer.sanger.ac.uk/cosmic) evidence, were retained; iv) variations in the exonic or splicing region (10 bp upstream and downstream of splicing sites) were retained; v) synonymous mutations were excluded; vi) variants with an unknown classification were excluded; and vii) functionally benign variant sites as predicted by PolyPhen 2 (Polymorphism Phenotyping v2; http://genetics.bwh.harvard.edu/pph2/) were excluded. KEGG pathway and GO term enrichment analysis were performed using the R package clusterProfiler (v3.10.1), and P<0.05 was considered to indicate a statistically significant difference (39).

To obtain serum, a minimum of 10 ml whole blood from each patient at initial diagnosis was centrifuged at 1,690 × g for 10 min at 4°C. Serum samples were stored at −80°C. The preparation of standard solutions of CEA, cytokeratin-19 fragments (CYFRA21-1) and neuron-specific enolase (NSE) was performed by serial dilution with a buffer solution (pH 7.5). Chemiluminescence immunoassays were used to measure the concentrations of these three aforementioned conventional serum biomarkers for tumors, as previously described (40,41). An automated chemiluminescence immunoassay analyzer (COBAS 8000 E 801; Roche Diagnostics GmbH) was used for analysis, with Carcinoembryonic Antigen (CEA) Assay Kit (Electrochemiluminescence Method) (cat. no. 07027079190), Cytokeratin 19 Fragment (CYFRA 21-1) Assay Kit (Electrochemiluminescence Method) (cat. no. 07299966190) and Neuron-Specific Enolase (NSE) Assay Kit (Electrochemiluminescence Method) (cat. no. 07299982190), all sourced from Roche Diagnostics GmbH.

The serum levels of 18 heavy metals [As, barium (Ba), Cd, cobalt (Co), Cr, Cu, gallium, Hg, manganese (Mn), Ni, Pb, antimony (Sb), selenium (Se), stannum (Sn), strontium (Sr), thallium (Tl), vanadium and Zn] were determined by ICP-MS, utilizing an Agilent 7800 instrument (Agilent Technologies, Inc.) with an SPS 4 Series autosampler (Agilent Technologies, Inc.) and a peristaltic pump for injecting samples. The Agilent ICP-MS 7800 instrument (Agilent Technologies, Inc.) was chosen for its capability to analyze multiple elements, with specific detection procedures detailed in the manufacturer's instructions (42).

For sample preparation, ≥2 ml whole blood was collected from each patient at initial diagnosis and centrifuged at 1,690 × g for 10 min at room temperature to obtain the serum. Serum samples were stored at −20°C until use. Subsequently, 500 µl serum and 500 µl internal standard solution (10 µg/ml; cat. no. 5191-4570; Agilent Technologies, Inc.) were added to 4,000 µl diluent consisting of 0.1% nitric acid (cat. no. 1.00456; MilliporeSigma) and 0.1% Triton™ X-100 (cat. no. 93443; MilliporeSigma). All components were then mixed and stored at 4°C until analysis.

The analysis of 18 heavy metals was performed using helium gas mode, with the instrument parameters shown in Table I. Sample analysis was preceded by an autotuning system using a tuning solution (1 µg/l cerium, Co, lithium, magnesium, Tl and yttrium; cat. no. 5185-5959; Agilent Technologies, Inc.), with the following major parameters: i) The mass 7 sensitivity should be >3,000 counts/sec; ii) the mass 89 sensitivity should be >10,000 counts/sec; iii) the mass 205 sensitivity should be >6,000 counts/sec; iv) for the aforementioned measurements, the relative standard deviation should be <15%, the oxide ratio <2%, the doubly charged ratio <3% and the peak height at 10% width between 0.65 and 0.8; and v) in helium mode, the Co intensity count should be >2,400. System control and data acquisition and processing were conducted using the MassHunter 4.4 Workstation^® version C.01.04 software (Agilent Technologies, Inc.).

The ‘maftools’ package (version 2.6.05) within the R software environment (version 4.0.3; R Core Team; http://www.r-project.org) was used to visualize genomic landscapes, lollipop plots and spectra of co-occurring and mutually exclusive genomic alterations (43). Statistical differences in the categorical variables between patients with and without hypertension were assessed using χ² test or Fisher's exact test in SPSS (version 22.0; IBM Corp.). Continuous variables are presented as the median and interquartile range (IQR), where comparisons between two groups were conducted using an unpaired Student's t-test or a Mann-Whitney U test using GraphPad Prism (version 7.0; Dotmatics). Spearman's rank correlation analysis by the R software (version 4.0.3; R Core Team; http://www.r-project.org) was used to assess the correlations among age, BMI, three traditional serum biomarkers and 18 heavy metals. P<0.05 was considered to indicate a statistically significant difference.

In the present study, the number of patients with primary NSCLC with stage 0, I, II, III and IV was 1, 33, 7, 17 and 6, respectively. Regarding hypertension status, 16 patients were confirmed to have hypertension (aged 39–82 years, median 64.50; referred to as Hyp group), and 48 patients did not have hypertension (aged 33–80 years, median 61.50; referred to as non-Hyp group). All 16 hypertensive patients were taking antihypertensive drugs, where the numbers of patients with different hypertension grades were 0 in optimal, 4 in normal, 3 in high normal, 5 in grade 1, 4 in grade 2 and 0 in grade 3. Among the 64 patients (Hyp group vs. non-Hyp group, 16 vs. 48), serum samples from 56 patients (Hyp group vs. non-Hyp group, 14 vs. 42) were used for the detection of conventional serum biomarkers, whereas serum samples from 62 patients (Hyp group vs. non-Hyp group, 16 vs. 46) were used for the assessment of 18 heavy metals. For the unaccounted for serum samples, 2 were excluded due to hemolysis during storage, and 6 patients provided an insufficient volume of serum, making it impossible to perform all planned analyses. No significant differences were observed regarding age, height, weight, BMI, sex distribution, the incidence of EGFR or TP53 mutations, pulmonary infection history, smoking history, drinking history, cancer staging, blood vessel invasion, nerve invasion, distant metastasis status, CEA, CYFRA21-1 and NSE levels between the Hyp group and the non-Hyp group, but significant differences were observed for other chronic diseases (including coronary heart disease, diabetes, atrial fibrillation, Parkinson's disease and chronic gastritis) (P<0.001; Table II).

Despite the considerable advantages observed when targeted therapies (such as gefitinib, afatinib and osimertinib) are used for the treatment of patients with NSCLC harboring driver gene mutations (44,45), there remains a knowledge gap regarding the differences in genomic profiles between individuals with hypertension and those without. To explore these differences, somatic mutations in the tumor tissues of 16 hypertensive patients and 48 non-hypertensive patients were examined using an NGS panel comprising 82 tumor-related genes, which were related to the occurrence, progression and treatment of solid tumors, mainly derived from the NCCN clinical practice guidelines in oncology, the cancer knowledgebase database and the oncokb database (Table SI). In total, 47 somatic mutations involving 14 mutant genes were detected in all 16 patients with hypertension (100%). In comparison, 113 somatic mutations affecting 26 mutant genes were identified in 45 of the 48 patients without hypertension (93.75%; Fig. 1; Table SII, Table SIII, Table IV). Among the identified mutant genes, FBXW7, CBR3, CDKN2A, HRAS, SMO and UGT1A1 were exclusively observed in patients with hypertension, whilst 18 mutant genes (including KRAS, MAP2K1, RB1, CTNNB1, STK11, TPMT, NTRK1, CYP2D6, DPYD, KIT, PIK3CA, BRAF, ESR1, FGFR2, GNAS, NOS3, NOTCH1 and PDGFRA) were only observed in patients without hypertension.

In addition, although EGFR, TP53, ERBB2, ALK, APC, FGFR3, RET and SOD2 mutations were detected in both groups, their mutation frequencies and sites markedly differed between the Hyp group and the non-Hyp group (Fig. S1; Table SV). For TP53 mutations, whilst ~50% of the patients in both the Hyp group (56.25%; 9/16) and the non-Hyp group (50%; 24/48; Fig. 1), notable differences were identified in the mutation types and sites (Fig. S1). Specifically, except for the G245D mutation, the other 26 mutant sites differed between the two groups (Hyp group: R273L, A159P, T125K, E221Q, Q192X, D148H, C141W, P152T, 187_193del, and 187_192del; non-hyp group: S241F, R196G, D42fs, S241Y, L344P, E286G, R158L, C275F, 131_132del, T284P, Y236C, E298X, R196X, R306X, R273C, and V203M). Moreover, the mutation rate of EGFR was found to be higher in patients with hypertension (75.00%; 12/16) compared with those without (66.67%; 32/48).

In NSCLC, mutations in EGFR and KRAS tended to occur independently, with instances of patients harboring both mutations exhibiting resistance to EGFR-tyrosine kinase inhibitors (46). In the present cohort, the somatic interactions identified in hypertensive patients differed from those observed in non-hypertensive patients. In the non-Hyp group, mutually exclusive interactions were observed between EGFR and STK11 (P=0.0240; Fig. 2B; Table SVI). However, no significant mutually exclusive interactions were found in the Hyp group (Fig. 2A; Table SVI). By contrast, notable disparities were observed in the sets of co-occurring genes between the two groups. APC and ALK (P=0.0588), FGFR3 and ALK (P=0.0588), HRAS and ALK (P=0.0588), and another 13 pairs of genes were markedly co-occurring in hypertensive patients (Fig. 2A; Table SVI). The significant sets of co-occurring genes in patients without hypertension were as follows: NTRK1 and KIT (P=0.0010), KRAS and STK11 (P=0.0169), RET and ALK (P=0.0217) and another 22 pairs of genes (Fig. 2B; Table SVI).

To obtain a deeper understanding of the functional biological distinctions between the Hyp and the non-Hyp group, KEGG and GO analyses were performed. The top 10 KEGG signaling pathways based on both gene counts and P-values were identified, where the majority of the signaling pathways found were closely associated with cancer (Fig. 3A and B; Table SVII). ‘Non-small cell lung cancer’ was the top KEGG pathway in hypertensive patients, whereas ‘central carbon metabolism in cancer’ was the top pathway in patients without hypertension. In the GO enrichment analysis, the most prevalent functional category in both groups was ‘transmembrane receptor protein tyrosine kinase activity’ (Fig. 3C and D; Table SVII). In addition, the proportions of altered signaling pathways and altered functional terms shared between these two groups were 54.37% (56/103) and 21.62% (8/37), respectively (Fig. 3E). Unexpectedly, all functional terms in the Hyp group were also identified in the non-Hyp group.

To further examine the differences between the groups of patients with NSCLC based on hypertension status, a comparative analysis of three conventional serum biomarkers between individuals with hypertension (n=14) and those without hypertension (n=42) was performed. The serum levels of CEA, CYFRA21-1 and NSE were all found to be similar between the Hyp group and the non-Hyp group (Table SVIII and IX). Subsequently, the serum concentrations of 18 heavy metals in the Hyp group (n=16) and the non-Hyp group (n=46) were compared, revealing a significant difference in Cr, with no significant difference found in the other 17 heavy metals (Fig. 4). The median concentrations (IQR) for Cr were as follows: 2.635 µg/l (1.678–3.438 µg/l) in the Hyp group and 1.97 µg/l (1.525–2.483 µg/l) in the non-Hyp group (Fig. 4E).

In the male group, the proportion of smokers (P=0.0049) and drinkers (P=0.0049) were significantly higher than those in the female group (Fig. S2; Table SVIII). Compared with the non-smoking group, the proportion of drinkers in the smoking group was significantly higher (P=0.0009; Fig. S2; Table SVIII). In addition, compared with that in patients without EGFR mutations, a significant decrease in Sn levels was observed in patients with EGFR mutations (P=0.0110) according to the two-tailed Mann-Whitney U test (Fig. S3; Table SVIII). However, none of the three traditional serum biomarkers, CEA, CYFRA21-1 and NSE, or 18 heavy metals exhibited significant differences between patients with and without TP53 mutations, between patients with and without other chronic diseases, between male and female sex, between non-smokers and smokers or between non-drinkers and drinkers (Fig. S4, Fig. S5, Fig. S6, Fig. S7, Fig. S8). In addition, Spearman's rank correlation analysis was next performed to examine the correlations among age, BMI, three traditional serum biomarkers and 18 heavy metals in the 14 hypertensive patients. Significant correlations were observed between traditional serum biomarkers and heavy metals, including between CEA and Cr (r=−0.67; P<0.01), between CYFRA21-1 and Zn (r=−0.64; P<0.01), between NSE and As (r=−0.66; P<0.01), between NSE and Ba (r=−0.76; P<0.01), between NSE and Pb (r=0.66; P<0.01) and between NSE and Se (r=0.58; P<0.05). In addition, significant correlations were observed between some of the 18 heavy metals, such as between As and Sr (r=0.54; P<0.05), between Ba and Hg (r=−0.59; P<0.05) and between Cd and Cu (r=0.58; P<0.05; Fig. 5).

Compared with patients with EGFR mutations, the proportion of KRAS mutations (P=0.0192) and smokers (P=0.0265) were significantly increased in patients without EGFR mutations (Fig. S9; Table SIX). The proportion of women in patients with TP53 mutations was significantly higher than in patients without TP53 mutations (P=0.0300) (Fig. S9; Table SIX). In addition, a significant decrease of Zn was observed in patients with EGFR mutations in comparison with that in patients without EGFR mutations (P=0.0280) according to the two-tailed unpaired t-test (Fig. S10; Table SIX). Compared with patients without TP53 mutations, patients with TP53 mutations exhibited a significantly lower level of As and significantly higher levels of Cu and CEA (Fig. S11). The serum levels of Co, Cr and Ni in patients with KRAS mutations were significantly higher than those in patients without KRAS mutations (Fig. S12). Compared with female patients, male patients exhibited significantly lower levels of Cu and Mn as well as significantly higher levels of Pb and Sn (Fig. S13). Compared with non-smokers, smokers exhibited a significantly lower level of Mn, and significantly higher levels of As, Cd, Pb and Sr (Fig. S14). Compared with non-drinkers, drinkers exhibited significantly lower levels of Ga and Mn, and a significantly higher level of Pb (Fig. S15). Spearman's rank correlation analysis was next performed to examine the correlations among age, BMI, three traditional serum biomarkers and 18 heavy metals in 42 patients without hypertension. BMI was found to be weakly negatively correlated with the serum concentrations of Cu (r=−0.3227; P=0.0371), but no significant correlations were observed between age and other factors (Fig. 6; Table SIX). Furthermore, CYFRA21-1 was significantly positively correlated with NSE (r=0.60; P<0.001), weakly but positively correlated with As (r=0.35; P<0.05) and Pb (r=0.39; P<0.01), whilst being negatively correlated with Cr (r=−0.39; P<0.01), differing from only a significant negative correlation between CYFRA21-1 and Zn observed in patients with hypertension. Among the 18 heavy metals, significant correlations were found in Fig. 6, such as between As and Sr (r=0.51; P<0.001), between Ba and Sr (r=0.52; P<0.001), and between Hg and Sn (r=0.64; P<0.001; Fig. 6).