Metabolomic Characteristics of Patients With Non-Small Cell Lung Cancer in Plasma, Tissue and Exhaled Breath Condensate: A Case-Control Study

Abstract

Introduction

Lung cancer is a malignant tumor with high global incidence and mortality. Early screening and diagnosis can substantially extend patient survival, yet achieving this remains a challenging objective.

Methods

We collected tissue and plasma samples from patients with benign pulmonary nodules and non-small cell lung cancer (NSCLC). Exhaled breath condensate (EBC) samples were obtained from these patients using a breath condenser. Differential metabolites were identified in samples from 22 NSCLC patients and 20 benign controls by means of ultra-performance liquid chromatography–high resolution mass spectrometry–based untargeted metabolomics.

Results

Several differential metabolites overlapped between plasma and tissue samples. Further analysis indicated that the differential metabolites from all three sample types participated in central carbon metabolism in cancer, protein digestion and absorption, and aminoacyl-tRNA biosynthesis. ROC curve analysis of plasma and tissue metabolites showed that each of the top 10 upregulated and downregulated metabolites yielded AUC values greater than 0.8. In EBC, only 29 differential metabolites were detected. Among these, lysine, acetildenafil, and 1-(cyclohexylmethyl) proline may be involved in the progression of NSCLC.

Conclusion

This study identified distinctive metabolic markers in plasma, tissue, and EBC from NSCLC patients. However, the purpose of this observational study was merely exploratory, rather than to use these differential metabolites as a diagnostic test for non-small cell lung cancer.

Introduction

Lung cancer remains a leading cause of cancer-related mortality worldwide, with high incidence and fatality rates. ¹ Its development is primarily associated with smoking, occupational exposures, air pollution, genetic predisposition, and chronic lung diseases. ² Pathologically, lung cancer is categorized into non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC), with NSCLC representing approximately 85% of all cases. ³ NSCLC itself comprises several histological subtypes, including adenocarcinoma, squamous cell carcinoma, and large cell carcinoma. ³ Current treatment strategies for NSCLC encompass surgical resection, radiotherapy, chemotherapy, targeted therapy, and immunotherapy. ⁴ While these interventions have substantially improved patient survival, early detection remains paramount as it dramatically increases the potential for a cure.^4,5 Consequently, achieving earlier diagnosis of NSCLC represents a major objective in current global oncology research.

Chest computed tomography (CT) is the preferred modality for lung cancer screening and diagnosis, whereas pathological biopsy remains the diagnostic gold standard.^6,7 Blood-based assays for tumor markers such as carcinoembryonic antigen (CEA), cytokeratin 19 fragment antigen (CYFRA21-1), and neuron-specific enolase (NSE), along with circulating tumor DNA (ctDNA), serve as auxiliary diagnostic indicators.^8,9 Nevertheless, the inherent risks of puncture biopsy, the limited ability of CT to definitively characterize lesions, and the poor specificity of serum tumor markers underscore the need for novel approaches to improve lung cancer detection and diagnosis.^10-13

Exhaled breath condensate (EBC), a non-invasive fluid derived from the airway-lining fluid of the lungs, contains numerous measurable substances, including biomarkers and drugs.^14,15 Current analyses of EBC have detected markers such as reactive nitrogen species, arachidonic acid metabolites, and oxidative stress mediators. ¹⁶ These substances are widely employed in studies of various lung diseases, including asthma, chronic obstructive pulmonary disease (COPD), and lung cancer.^16-19 However, EBC consists of over 99% condensed water vapor and aerosolized droplets from the airway lining fluid, indicating that its constituent compounds are highly diluted. ²⁰ Furthermore, the concentrations of non-volatile compounds in EBC vary significantly not only among individuals but also within the same person.^21,22 Consequently, although EBC represents a valuable non-invasive biological sample, its accuracy requires validation against other sample types to ensure the reliability of the findings.

Several studies have suggested that EBC can be used to detect lung cancer. Kazeminasab et al demonstrated through polymerase chain reaction (PCR) that TP53 and KRAS mutations in the EBC of lung cancer patients fully matched those in tumor tissues. ²³ Comparing let-7 levels in NSCLC patients—across tumor tissues, serum, and EBC—with those in healthy volunteers, Chen et al found significantly reduced let-7 levels in patients’ tumor tissues, serum, and EBC. ²⁴ Proteomic investigations further support the considerable potential of EBC for the early detection of lung cancer.^25-28 Collectively, this evidence indicates that EBC holds promise for diagnosing and screening lung cancer and other diseases. ²⁹

Metabolomics has become a widely used research approach in recent years and is now applied across numerous scientific disciplines. Since alterations in metabolites frequently correlate with disease states, they offer valuable insights into pathological mechanisms. Li et al employed LC-MS/MS-based untargeted metabolomics to identify 25 differential metabolites in the EBC of pneumonia patients. ³⁰ Their subsequent analysis suggested that the hypotaurine/taurine metabolic pathway could contribute to the development of pneumonia, with 3-sulfoalanine emerging as a potential diagnostic biomarker. ³⁰ Similarly, Barberis et al reported 26 differentially metabolites in EBC samples from healthy individuals and patients infected with SARS-CoV-2. ³¹ These findings collectively support the feasibility of using metabolomics to analyze EBC samples in respiratory diseases. Additionally, multiple studies have reported metabolomic profiles of EBC obtained from lung cancer patients.^32-35 Together, these results further substantiate the potential of EBC-based metabolomics for lung cancer screening.

Although EBC analysis offers a non-invasive approach for lung cancer screening, the differential metabolites identified thus far show limited consistency and repeatability. These shortcomings likely stem from analytical batch effects and variation between testing platforms. To more accurately identify differential metabolites associated with lung cancer, we analyzed tissue, plasma, and EBC specimens from NSCLC patients alongside plasma and EBC samples from patients with benign pulmonary nodules. Our metabolomic analysis identified several differential metabolites that could potentially support future lung cancer screening efforts.

Materials and Methods

Cohort Information

This study enrolled 22 NSCLC patients (19 AC and 3 SCC) and 20 patients with benign pulmonary nodules (control group). all patients signed a written informed consent form before sample collection (including but not limited to blood, surgically resected tissue and EBC) to confirm their willingness to provide biological samples for scientific research. The reporting of this study conforms to STROBE guidelines. ³⁶ Moreover, all patient data were de-identified. The NSCLC patients received their diagnosis via bronchoscopic mucosal or core biopsy samples. From the NSCLC group, we obtained preoperative and postoperative plasma, tumor and peritumoral tissues, and EBC. In the benign pulmonary nodule group, plasma and EBC were collected. No patient had undergone chemotherapy or radiotherapy prior to sample collection. Table 1 and S1 summarize the cohort characteristics for both the NSCLC and benign pulmonary nodule groups.

Table 1.

The Demographic of NSCLC Patients and Benign Nodule Controls

Sample		Number of samples	Gender (F vs M)	Number of smokers	Number of quit smoker
Lung tissue	Tumor	10	8 vs 2	4	2
	Peritumoral
Plasma	NSCLC-preoperative	7	5 vs 2	3	1
	NSCLC-postoperative
	BPN	10	7 vs 3	1	0
EBC	NSCLC	10	4 vs 6	1	1
	BPN	10	4 vs 6	2	2

NSCLC: Non-small cell lung cancer; BPN: Benign pulmonary nodules.

EBC, Plasma and Tissue Collection

EBC samples were collected from patients in the NSCLC and benign pulmonary nodule groups using an EBC collection device (Patent No: ZL201921297152.7, Figure 1). During inhalation and exhalation, the four-way control valves prevent external gas from entering the condensate collection system and also prevent gas in the condenser tube from flowing back into the valves. Under the action of the condenser in the condenser tube sleeve, the exhaled gas is pre-condensed and cools down, forming liquid which then flows along the tube wall into the collection cup. Participants breathed through a suction nozzle connected to a condenser tube sleeve for 15 minutes, allowing ambient air to flow into the lungs. The exhaled breath, totaling 1-2 ml, was collected in a temperature-controlled box maintained at -5 °C equipped with a collection cup. This setup caused the exhaled breath to rapidly cool and freeze within the collection cup. Before sampling, all subjects were required to fast for 2 hours.

Figure 1.

Exhaled breath condensate sample collector unit (A) Diagram of the external structure of the EBC collector. (B) Front view of the condensate collection system. (C) Back view of the condensate collection system. 1. Air intake nozzle, 2. Four-way control valve for exhalation and inhalation, 3. Condenser tube sleeve, 4. Insulation cover, 5. Angle adjustment rod, 6. Lifting pole, 7. Flexible pipe, 8. Support pole, 9. Temperature display, 10. Temperature control box, 11. Intake, 12. Condenser, 13. Collect cups, 14. Exhalation outlet

Blood samples were collected from patients with benign pulmonary nodules or preoperative NSCLC during admission examinations, whereas postoperative samples were obtained one week after surgery. The samples were drawn into EDTA-containing tubes, and plasma was separated by centrifugation at 3000×g for 10 minutes.

For the tissue samples, we obtained lung tissues from NSCLC patients, encompassing both tumor tissues and macroscopically normal lung tissue located at least 2 cm from the lesion.

The collected specimens were immediately flash-frozen in liquid nitrogen and stored at -80 °C until analysis. Untargeted metabolomic analysis was subsequently performed by Suzhou Baiqu Biotechnology Co., Ltd.

Metabolites Extraction

For EBC and plasma samples, 100 μL of each sample was mixed with 400 μL of protein precipitant (MeOH: ACN, 1:1 (v/v)) containing deuterated internal standards after thawing at room temperature. The mixture was vortexed for 30 s, sonicated for 10 min in a 4 °C water bath, and incubated for 1 h at -40 °C to precipitate proteins.

For tissue samples, approximately 25 mg of tissue was combined with beads and 500 μL of extraction solution (MeOH: ACN: H₂O, 2:2:1 (v/v)) that also contained deuterated internal standards. This mixture was vortexed for 30 s, homogenized at 35 Hz for 4 min, and sonicated for 5 min in a 4 °C water bath; the homogenization and sonication cycle was repeated three times. The samples were then incubated for 1 h at -40 °C to precipitate proteins.

All EBC, plasma, and tissue samples were subsequently centrifuged at 12,000 rpm for 15 min at 4 °C. The resulting supernatant was transferred to a fresh glass vial for analysis. A quality control (QC) sample was prepared by combining equal aliquots of the supernatant from all samples. Shanghai BIOTREE Biomedical Technology Co., LTD performed all the steps described above.

LC-MS/MS Analysis

Polar metabolites were analyzed by LC-MS/MS using a UHPLC system (Vanquish, Thermo Fisher Scientific) equipped with a Waters ACQUITY UPLC BEH Amide column (2.1 mm × 50 mm, 1.7 μm) and coupled to an Orbitrap Exploris 120 mass spectrometer (Orbitrap MS, Thermo). The mobile phase comprised (A) 25 mmol/L ammonium acetate and 25 mmol/L ammonium hydroxide in water (pH = 9.75) and (B) acetonitrile. The auto-sampler temperature was maintained at 4 °C, with an injection volume of 2 μL. The Orbitrap Exploris 120 mass spectrometer acquired MS/MS spectra via information-dependent acquisition (IDA) under the control of Xcalibur software. This software continuously monitored the full scan MS spectrum during acquisition. Electrospray ionization source conditions were as follows: sheath gas flow rate 50 Arb, auxiliary gas flow rate 15 Arb, capillary temperature 320 °C, full MS resolution 60,000, MS/MS resolution 15,000, stepped normalized collision energy 20/30/40, and spray voltages of 3.8 kV (positive) or –3.4 kV (negative). All analyses were performed by Shanghai BIOTREE Biomedical Technology Co., Ltd.

Data Preprocessing and Annotation

The raw data were converted to mzXML format using ProteoWizard and processed with an in-house program developed in R and based on XCMS for feature detection, extraction, alignment, and integration. Metabolite identification was performed using the R package and BiotreeDB (V3.0). ³⁷

Data Analysis

The statistical analysis proceeded through several steps to identify differences between the lung cancer and benign pulmonary nodule groups. Data were visually analyzed using R and SIMCA. For the untargeted metabolomics datasets, statistical tests were performed exclusively on metabolites consistently detected across the samples.

To ensure the quality of the data, we determined the stability of the detection by comparing the peak heights of the internal standard in the QC samples, and assessed residual substances by analyzing the peak heights of the internal standard in the blank samples. Moreover, the stability of the samples during the extraction and detection was assessed using Principal Component Analysis (PCA). Furthermore, to reduce the influence of detection system errors on the results and make the results better reflect the biological significance, we performed the following preprocessing steps on the raw data: (1) Deviation value filtering: features with a Coefficient of variation (CV) above a predefined threshold in QC samples were removed to reduce noise; (2) Missing value filtering: Features with >50% missing values in any single group or >50% missing values across all groups were excluded; (3) Missing value filling: missing values were replaced with half of the minimum positive value detected for each feature; (4) Data normalization: The data were normalized to the total ion current (TIC) of each sample. When all of these are satisfactory, proceed to the next stage of analysis.

In the initial stage of statistical analysis, sample grouping was assessed via PCA. Significant differences between the sample group and its subgroup were subsequently identified using volcano plots (Table S2-S5). These volcano plots were generated with a fold change threshold of 1.3 and a false discovery rate (FDR)-adjusted p-value threshold of 0.05. The top 10 up-regulated and down-regulated metabolites were visualized using heat maps and box plots. Metabolite p-values across different groups were calculated with the Wilcoxon rank sum test.

Correlation analysis of differential metabolites was performed to clarify their mutual regulatory relationships during biological state transitions. These correlations often reflect synergistic changes among metabolites. The correlation coefficient r quantifies the relationship between two variables, ranging from −1 to 1. For each group of comparisons, we computed correlation coefficients from the quantitative values of the differential metabolites. The Spearman method was used for these calculations, and the results were displayed as heatmaps.

We then classified the differential metabolites using Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. The horizontal axis indicates the proportion of differential metabolites annotated to a given pathway relative to the total number of annotated differential metabolites, while the vertical axis lists the names of the enriched KEGG metabolic pathways. KEGG annotation analysis identified the pathways in which the differential metabolites participate.

Finally, we evaluated the differential metabolites using receiver operating characteristic (ROC) curve analysis.

Results

Plasma Metabolic Changes Between the NSCLC and the Benign Pulmonary Nodule Groups

To investigate plasma metabolic changes as a potential approach for lung cancer screening, we collected plasma from patients with NSCLC and benign pulmonary nodules. For the NSCLC group, plasma was obtained both preoperatively and one week postoperatively. We then analyzed the metabolites in these samples. PCA revealed tight clustering of quality control samples, indicating high data quality (Figure 2A). A total of 28,473 distinct metabolites were detected. Features with zero values in the samples or lacking candidate identities in HMDB were excluded, resulting in a final set of metabolites for subsequent analysis.

Figure 2.

Metabolic characteristics of plasma samples from patients with NSCLC before and after surgery and those with benign pulmonary nodules (A) The PCA dispersion point maps of patients with NSCLC before and after surgery, benign pulmonary nodules and quality control plasma samples. (B) The volcano map showed the differential metabolites of plasma samples from NSCLC patients before and after surgery. (C) The volcano map showed the differential metabolites of plasma samples from patients with NSCLC before surgery and those with benign pulmonary nodules. (D) The heat map showed the top 10 up-regulated and down-regulated metabolites in NSCLC patients before and after surgery. (E) The heat map showed the differences in the top 10 up-regulated and down-regulated metabolites in patients with NSCLC before surgery and benign pulmonary nodules. P1: preoperative NSCLC group, P2: postoperative NSCLC group, P3: benign pulmonary nodule group

Further analysis comparing the preoperative NSCLC group (P1) with the postoperative group (P2) identified 61 up-regulated and 32 down-regulated differential metabolites (Figure 2B). A comparison between P1 and the benign pulmonary nodule group (P3) revealed 175 up-regulated and 148 down-regulated metabolites (Figure 2C). The top 10 up- and down-regulated differential metabolites from the P1 vs. P2 and P1 vs. P3 comparisons were displayed as heat maps (Figure 2D and E). Among these, 2-Hydroxypalmitic acid, 16-Hydroxypalmitic acid, 3-hydroxy-tetradecanoic acid, and 1-22:0-2-18:3-phosphatidylserine were up-regulated in both comparisons, with three of the four belonging to the fatty acyl class. Conversely, 4-hydroxy-7H-furo[3,2-G]chromen-7-one, M321T177, Xylulose, and 3′-Sialyllactose were down-regulated in both P1 vs. P2 and P1 vs. P3; these metabolites are classified as phenylpropanoids and polyketides, organic oxygen compounds, or other compounds. A correlation analysis of the top 10 up- and down-regulated metabolites showed positive correlations between metabolites that were co-up-regulated or co-down-regulated, most of which were statistically significant (sFigure 1).

To elucidate the pathways associated with these differential metabolites, we performed a KEGG-based pathway enrichment analysis. The results indicate that the differential metabolites identified in both P1 vs P2 and P1 vs P3 comparisons were collectively involved in glycine, serine and threonine metabolism, choline metabolism in cancer, central carbon metabolism in cancer, mineral absorption, protein digestion and absorption, carbon metabolism, glycerophospholipid metabolism, and aminoacyl-tRNA biosynthesis (Figure 3A and B). Subsequently, ROC curve analysis revealed that the AUC values for the top 10 upregulated and downregulated metabolites all exceeded 0.8 (sFigure 2 and sFigure 3, Tables 2 and 3, Table S6 and Table S7).

Figure 3.

The pathways involved in differential metabolites in plasma samples of patients with NSCLC before and after surgery and with benign pulmonary nodules (A) KEGG enrichment analysis revealed the pathways involved in differential metabolites in the plasma of NSCLC patients before and after surgery. The horizontal axis represents the percentage of the number of differentially metabolites annotated under a certain pathway to the total number of differentially metabolites annotated, and the vertical axis represents the name of the KEGG metabolic pathway enriched. (B) KEGG enrichment analysis revealed the pathways involved in differential metabolites in the plasma of patients with NSCLC before surgery and those with benign pulmonary nodules

Table 2.

Identification of Candidate Differential Metabolites in the Plasma of Patients With NSCLC before and after Surgery

No.	Experimental m/z	Rt (s)	Elemental formula	CAS	Metabolite annotation	Log2(FC=Preoperative/Postoperative)	AUC
1	241.1430	167.6	C13H22O4	1083193-56-9	2-Oct-7-enylpentanedioic acid	0.84	0.92
2	728.5551	38.3	C41H78NO7P	​	PE(P-18:1(9Z)/18:1(9Z))	0.87	0.88
3	481.2780	69.1	C26H42O8	​	1.α.-Methyl-5.α-androstan-3.α.,17.β-diolglucuronide	0.98	0.94
4	271.2261	25.0	C16H32O3	764-67-0	2-Hydroxypalmitic acid	0.84	0.92
5	271.2261	25.0	C16H32O3	506-13-8	16-Hydroxypalmitic acid	0.84	0.92
6	137.0701	181.2	C7H9N2O	3106-60-3	1-Methylnicotinamide	1.84	1.00
7	243.1949	28.2	C14H28O3	​	3-hydroxy-tetradecanoic_acid	0.96	0.96
8	804.5471	55.6	C46H78NO8P	​	PC(38:7)	0.43	0.94
9	397.1308	165.6	C22H20O7	148719-52-2	Artonin_L	0.57	0.90
10	840.5710	62.3	C46H84NO10P	​	1-22:0-2-18:3-phosphatidylserine	0.69	0.92
11	203.0391	173.1	C11H6O4	486-60-2	4-hydroxy-7H-furo[3,2-g]chromen-7-one	-0.54	0.92
12	321.0659	176.8	C15H14O8	​	M321T177	-0.73	1.00
13	367.1028	176.3	C17H20O9	62929-69-5	3-O-Feruloylquinic acid	-0.35	0.84
14	519.2665	119.6	C27H40N2O8	361432-41-9	3-Hydroxystanozolol glucuronide	-0.56	0.90
15	213.0158	101.9	C6H11ClO6	​	Glucose chloride	-0.59	0.94
16	177.0393	61.9	C6H10O6	90-80-2	Glucono-1,5-lactone	-0.38	0.92
17	149.0445	41.5	C5H10O5	551-84-8	Xylulose	-0.61	0.94
18	143.0340	176.9	C6H8O4	2459-05-4	(2E)-4-Etgoxy-4-oxo-2-butenoic acid	-0.37	0.90
19	632.2011	246.1	C23H39NO19	35890-38-1	3′-Sialyllactose	-1.11	0.96
20	230.0237	75.1	C8H10NO5P	120667-15-4	4-Phosphonophenylglycine	-0.91	0.94

Table 3.

Identification of Candidate Differential Metabolites in the Plasma of Patients With NSCLC and Benign Pulmonary Nodules

No.	Experimental m/z	Rt (s)	Elemental formula	CAS	Metabolite annotation	Log2(FC=Preoperative/Postoperative)	AUC
1	353.0248	252.4	C18H8O8	479-64-1	Thelephoric_acid	-2.82	0.96
2	285.0385	252.4	C15H10O6	491-70-3	Luteolin	-2.80	0.99
3	225.0531	252.3	C14H8O3	784-50-9	9-Oxo-9H-fluorene-2-carboxylic acid	-2.30	0.93
4	334.0505	252.3	C16H12ClNO5	95617-09-7	Fenoxaprop	-2.80	0.96
5	368.9986	253.0	C15H13IO3	119113-94-9	4-Benzyloxy-3-iodo-5-methoxybenzaldehyde	-2.97	0.97
6	267.0285	252.5	C15H8O5	479-13-0	Coumestrol	-2.67	0.97
7	284.0909	252.3	C16H13NO4	115610-36-1	avenanthramide D	-2.47	0.94
8	266.0792	252.3	C14H10F3NO	347-37-5	N-[1,1′-Biphenyl]-4-yl-2,2,2-trifluoroacetamide	-2.79	0.96
9	293.0961	252.4	C10H16N2O8	62-33-9	EDTA	-2.95	0.93
10	247.0908	252.3	C14H14O4	​	Nodakenitin	-2.80	0.93
11	106.0492	231.3	C3H7NO3	56-45-1	Serine	0.98	0.96
12	271.2261	25.0	C16H32O3	764-67-0	2-Hydroxypalmitic acid	0.90	0.96
13	271.2261	25.0	C16H32O3	506-13-8	16-Hydroxypalmitic acid	0.90	0.96
14	276.1540	282.2	C11H21N3O5	​	γ-Glutamyllysine	0.64	0.96
15	271.0909	98.7	C17H12F2O	53369-00-9	1,5-Bis-(4-fluorophenyl)penta-1,4-dien-3-one	1.44	0.97
16	256.0918	248.7	C14H15N3S	71196-81-1	N-(4-Methylphenyl)-N’-(6-methyl-2-pyridinyl)thiourea	2.16	0.99
17	440.3117	130.0	C28H41NO3	199875-69-9	N-Arachidonoyldopamine	1.18	0.96
18	303.2309	20.3	C20H32O2	506-32-1	Arachidonic acid (AA)	1.77	1.00
19	496.3363	123.8	C24H50NO7P	17364-16-8	LPC(16:0)	0.52	0.90
20	482.3572	126.5	C24H52NO6P	52691-62-0	1-O-Hexadecyl-sn-glycero-3-phosphocholine (LPC(O-16:0/0:0))	0.95	0.89

Tissue Metabolic Changes Between Tumor and Peritumoral in the NSCLC Group

Metabolic activity in lung tissue often provides a more accurate reflection of cancer cell characteristics. We therefore collected tumor and peritumoral tissues from NSCLC patients. PCA indicated high data quality (Figure 4A). Further analysis identified 295 up-regulated and 149 down-regulated differential metabolites when comparing tumor tissue (T1) with peritumoral tissue (T2) (Figure 4B). The 10 most significantly up- and down-regulated metabolites were displayed in a heat map (Figure 4C). Down-regulated metabolites fell into 7 categories: steroids and steroid derivatives, azoles, organonitrogen compounds, carboxylic acids and derivatives, glycerophospholipids, fatty acyls, and other compounds. Up-regulated metabolites were classified into 8 categories: steroids and steroid derivatives, hydroxy acids and derivatives, benzene and substituted derivatives, furofurans, glycerophospholipids, azolidines, carboxylic acids and derivatives, and other compounds. Correlation analysis of the top ten up- and down-regulated metabolites revealed positive correlations between metabolites altered in the same direction, most of which were statistically significant (sFigure 4).

Figure 4.

Metabolic and pathway characteristics of NSCLC samples and peritumoral tissues (A) The PCA discrete point maps of tumor, adjacent and quality control specimens in patients with NSCLC. (B) The volcano map showed the differences in metabolites between NSCLC samples and peritumoral tissues. (C) Compared with the peritumoral tissues, the heat map showed the top 10 upregulated and downregulated metabolites in the NSCLC samples. (D) KEGG enrichment analysis revealed the pathways involved by differential metabolites in NSCLC samples and peritumoral tissues. T1: tumor tissue group, T2: peritumoral tissue group

To further elucidate the pathways associated with these differential metabolites, we performed a KEGG pathway enrichment analysis. The results indicated that the differential metabolites in T1 and T2 were involved in alanine, aspartate and glutamate metabolism, arginine biosynthesis, central carbon metabolism in cancer, choline metabolism in cancer, protein digestion and absorption, proximal tubule bicarbonate reclamation, and other metabolic pathways (Figure 4D). Subsequent ROC curve analysis revealed that the AUC values for the all of the top 10 upregulated or downregulated metabolites exceeded 0.8 (sFigure 5, Table 4 and Table S8).

Table 4.

Identification of Candidate Differential Metabolites in Tumor Tissues and Adjacent Tissues of Patients With NSCLC

No.	Experimental m/z	Rt (s)	Elemental formula	CAS	Metabolite annotation	Log2(FC=Preoperative/Postoperative)	AUC
1	460.2776	116.3	C25H37N3O5	118675-50-6	4-Imidazolidineheptanoic acid, 3-[(2-cyclohexyl-2-hydroxyethyl)amino]-2,5-dioxo-1-(phenylmethyl)	1.01	0.90
2	438.2955	116.7	C21H44NO6P	174062-72-7	LPE(P-16:0)	1.15	0.91
3	436.2804	117.1	C21H44NO6P	​	LysoPE(P-16:0/0:0)	1.02	0.90
4	189.1335	303.8	C7H16N4O2	17035-90-4	L-NG-Monomethylarginine	1.80	0.97
5	497.9685	289.1	C9H16N3O15P3	​	5-hydroxy-CTP	1.87	0.96
6	667.2762	148.7	C35H42N2O11	84-36-6	Syrosingopine	1.48	0.90
7	347.0234	56.5	C12H12O12	72691-25-9	Dehydro-L-(+)-ascorbic acid dimer	3.15	0.93
8	584.2560	151.8	C31H37NO10	147444-03-9	Pyripyropene A	0.79	0.89
9	243.0075	28.3	C10H9ClO5	827592-22-3	(5-Chloro-4-formyl-2-methoxyphenoxy)acetic acid	1.08	0.91
10	142.9976	75.4	C3H3F3O3	684-07-1	3,3,3-Trifluorolactic acid	0.94	0.87
11	201.1587	169.6	C10H20N2O2	14040-76-7	N1,N2-Diisobutylethanediamide	-1.58	0.92
12	734.5642	19.4	C40H80NO8P	63-89-8	PC(16:0/16:0)	-0.60	0.92
13	505.2540	56.7	C24H43O9P	​	1-α-linolenoyl-sn-glycero-3-phospho-(1′-sn-glycerol)	-1.30	0.96
14	402.2076	237.0	C21H27N3O5	​	3-(1-(Cyclohexylmethyl)-1H-indazole-3-carboxamido)-2,2-dimethylsuccinic acid	-1.36	0.95
15	935.4879	277.8	C45H74O20	118524-12-2	Ampeloside_Bs1	-0.64	0.93
16	750.5401	19.4	C40H80NO7PS	113881-60-0	1-Palmitoyl-2-thiopalmitoyl phosphatidylcholine	-0.66	0.94
17	540.4961	65.7	C33H65NO4	​	Hexacosanoyl_carnitine	-1.65	1.00
18	399.3243	51.6	C27H44O2	19356-17-3	25-Hydroxyvitamin D3	-0.88	0.87
19	268.0583	101.3	C14H11N3OS	26028-88-6	4-(4-Phenyl-5-sulfanyl-4H-1,2,4-triazol-3-yl)phenol	-0.32	0.89
20	757.4481	285.6	C39H64O14	265092-98-6	(25S)-Spirostane-3b,5b,6a-triol_3-[4″-rhamnosylglucoside]	-0.76	0.91

EBC Metabolic Changes Between the NSCLC and the Benign Pulmonary Nodule Groups

PCA revealed high data quality (Figure 5A). Comparison with the benign pulmonary nodule group (EBC2) identified 3 up-regulated and 26 down-regulated differential metabolites in the NSCLC group (EBC1) (Figure 5B). The 3 upregulated metabolites and the top 10 significantly downregulated metabolites were displayed as box plots (Figure 5C and D). The upregulated metabolites were categorized as carboxylic acids and derivatives or pyridines and derivatives. The downregulated metabolites fell into 6 classes: piperidines, fatty acyls, organooxygen compounds, carboxylic acids and derivatives, benzene and substituted derivatives, and isoindoles and derivatives. A subsequent correlation analysis indicated a positive relationship between metabolites that were co-upregulated or co-downregulated, with most correlations reaching statistical significance (sFigure 6).

Figure 5.

Metabolic and pathway characteristics of EBC samples from patients with NSCLC and benign pulmonary nodules (A) PCA discrete point plots of EBC and quality control specimens in patients with NSCLC and benign pulmonary nodules. (B) The volcano map showed the differences in metabolites in EBC samples from patients with NSCLC and benign pulmonary nodules. (C) The box plot showed the upregulated differential metabolites in EBC samples from patients with NSCLC. (D) The box plot showed the top 10 downregulated differential metabolites in EBC samples from patients with NSCLC. (E) KEGG enrichment analysis revealed the pathways involved in differential metabolites in EBC samples from patients with NSCLC and benign pulmonary nodules. EBC1: NSCLC group, EBC2: benign pulmonary nodule group

To further elucidate the pathways associated with these differential metabolites, we performed a KEGG pathway enrichment analysis. The results indicated that the differential metabolites in EBC1 and EBC2 were enriched in arginine biosynthesis, central carbon metabolism in cancer, protein digestion and absorption, amino acid biosynthesis, and other metabolic pathways (Figure 5E). ROC curve analysis revealed that the models of lysine, acetildenafil, and 1-(Cyclohexylmethyl) proline performed relatively well (sFigure 7, Table 5 and Table S9).

Table 5.

Identification of Candidate Differential Metabolites in EBC of Patients With NSCLC and Benign Pulmonary Nodules

No.	Experimental m/z	Rt (s)	Elemental formula	CAS	Metabolite annotation	Log2(FC=Preoperative/Postoperative)	AUC
1	427.3024	19.8	C20H38N6O4	55123-66-5	Leupeptin	0.70	0.78
2	295.1631	119.8	C15H22N2O4	40829-32-1	Tyr-Ile	1.48	0.82
3	304.2984	37.1	C21H38N	7773-52-6	1-Hexadecylpyridinium cation	0.92	0.77
4	175.1180	282.8	C6H14N4O2	74-79-3	Arginine	-1.22	0.83
5	322.0686	257.9	C16H10F3NO3	728887-93-2	3-Oxo-2-[3-(trifluoromethyl)phenyl]-2,3-dihydro-1H-isoindole-4-carboxylic acid	-0.50	0.77
6	147.1120	286.2	C6H14N2O2	56-87-1	Lysine	-1.12	0.87
7	102.1271	109.8	C6H15N	617-79-8	2-Ethylbutan-1-amine	-1.12	0.79
8	161.1276	304.7	C7H16N2O2	98063-21-9	(R)-Aminocarnitine	-0.48	0.80
9	205.1688	324.6	C13H20N2	55455-92-0	1-(3-Phenylpropyl)piperazine	-0.48	0.80
10	467.2765	16.6	C25H34N6O3	831217-01-7	Acetildenafil	-0.65	0.81
11	226.1790	44.6	C13H23NO2	31582-45-3	2,2,6,6-Tetramethyl-4-piperidinyl 2-methylacrylate	-0.55	0.82
12	212.1634	51.7	C12H21NO2	180845-74-3	1-(Cyclohexylmethyl)proline	-0.31	0.84
13	291.1979	17.0	C18H28O3	125559-74-2	9-Oxo-10E,12Z,15Z-octadecatrienoic acid	-1.16	0.77

Discussion

Early detection of lung cancer is crucial for improving patient survival. Currently, however, reliable markers for early-stage screening remain scarce. EBC, a non-invasive sample type that can be collected rapidly, shows considerable promise for this application. Metabolic profiling captures the aberrant metabolic pathways characteristic of lung cancer, offering insights that could inform dietary interventions and targeted therapies. While several studies support the feasibility of using EBC metabolomics for lung cancer screening, the differential metabolites identified have not been validated in independent sample types, casting doubt on their reliability.^20,33-35 In this study, we collected samples from patients with benign pulmonary nodules and from individuals with NSCLC both before and after surgery. By analyzing differential metabolites across plasma, tissue, and EBC samples, we identified several potentially valuable metabolites in EBC. The differential metabolites found in these different types of samples are potentially more reliable than those from a single sample source previously. ³⁵

Untargeted metabolomics identified numerous up- and down-regulated metabolites in the P1 versus P2 and P1 versus P3 comparisons. To more precisely define the effective differential metabolites, we further analyzed these results. The analysis revealed that 2-Hydroxypalmitic acid, 16-Hydroxypalmitic acid, 3-hydroxy-tetradecanoic acid, and 1-22:0-2-18:3-phosphatidylserine were upregulated in group P1. Most of these differential metabolites are classified as fatty acyls. Through KEGG enrichment analysis, we found that the glycerophospholipid metabolic pathway was annotated. A recent study corroborates that glycerophospholipid metabolism is the most significantly altered lipid metabolism pathway in the blood of lung cancer patients. ³⁸ We also observed that glycine, serine, and threonine metabolism may significantly influence NSCLC progression. These metabolic pathways have similarly been reported as upregulated in other lung cancer studies. ³⁹ Inhibiting their metabolism effectively reduces the viability of NSCLC cells. ⁴⁰ In summary, this study is consistent with previous research, and both confirm that there are abnormal metabolisms of glycerophospholipids, glycine, serine, and threonine in the patients’ plasma.

Tissue biopsy remains the gold standard for diagnosing lung cancer, as cancerous tissues can more directly reflect the physiological environment within the patient’s body. ⁴¹ A comparison with plasma samples revealed rarely overlap in differential metabolites between tissues and plasma. This indicates that metabolites entering the bloodstream from the lungs may undergo multiple metabolic transformations, leading to their decomposition or conversion into new metabolites. KEGG enrichment analysis further supported this interpretation. The results indicate that choline metabolism in cancer, central carbon metabolism in cancer, protein digestion and absorption, carbon metabolism, glycerophospholipid metabolism, and aminoacyl-tRNA biosynthesis may represent important metabolic pathways in NSCLC progression. Previous studies have demonstrated that: (i) abnormal choline metabolism promotes M2-type macrophage polarization and endothelial cell proliferation, establishing an immunosuppressive microenvironment that accelerates NSCLC progression ⁴² ; (ii) markedly suppressed central carbon metabolism and glycerophospholipid metabolism can inhibit NSCLC cell proliferation^43-45; (iii) aminoacyl-tRNA biosynthesis and glycerophospholipid metabolism may be closely associated with cisplatin resistance. ⁴⁶ Consequently, these different metabolites in these pathways may play an important role in the progression of NSCLC.

To gather more patient information to better understand the metabolic process of NSCLC, we identified EBC as a suitable medium. As anticipated, the differential metabolites in EBC differed from those found in plasma and tissues. These EBC metabolites, however, participate in several metabolic pathways associated with NSCLC progression. A comparison across sample types revealed that central carbon metabolism in cancer, protein digestion and absorption, and aminoacyl-tRNA biosynthesis were commonly enriched, underscoring the relevance of these pathways in the progression of NSCLC. Glutamine, as a crucial fuel and supplement in the central carbon metabolism of cancer cells, may play a key role in the tumor progression and aggressiveness. ⁴⁷ Broad-spectrum glutamine blockade enhances the immune therapy response in lung cancer by increasing immune cell infiltration and reducing PD-L1 expression in cancer cells and tumor-associated macrophages (TAMs). ⁴⁷ Abnormalities in protein digestion and absorption can profoundly affect the metabolic behavior, oxidative stress levels of lung cancer cells, and the immune response in the tumor microenvironment by shaping the amino acid nutritional status throughout the body, inducing changes in the gut microbiota ecology and systemic immune-inflammatory imbalance.^48-51 However, a limitation arises from the composition of EBC, which is over 90% water, resulting in the identification of only 3 up-regulated and 26 down-regulated metabolites. Another recent study similarly reported just 18 differential metabolites in the EBC of lung cancer patients. ³⁵ These differential metabolites are also involved in the above three metabolic pathways. Consequently, while informative, the data obtainable from EBC remains limited.

Our research has enriched the understanding of metabolism in NSCLC. Unfortunately, this study also has certain limitations, namely a small sample size, and failure to address the changes caused by potential confounding factors. Moreover, although we identified some shared or matrix-specific metabolic alterations across tissue, plasma, and EBC, we did not conduct sufficient validation. Therefore, the purpose of this observational study was purely exploratory, aiming solely to identify the differential metabolites in the samples of NSCLC and benign pulmonary nodules, rather than to propose these differential metabolites as diagnostic tests for NSCLC.

Conclusions

Untargeted metabolomics was applied to plasma, tissue, and EBC samples from patients with NSCLC and control subjects. The analysis revealed three upregulated metabolites in the EBC of NSCLC patients: two carboxylic acids and their derivatives (tyr-Ile and leupeptin), along with one pyridine and its derivative (1-Hexadecylpyridinium cation). In contrast, 29 metabolites were downregulated in the EBC of NSCLC patients, comprising 5 carboxylic acids and derivatives, 3 organooxygen compounds, 5 benzene and substituted derivatives, and 2 organonitrogen compounds, etc.

Analysis of the top 10 differential metabolites revealed that the models of lysine, acetildenafil, and 1-(Cyclohexylmethyl) proline performed relatively well. Classification of the differential metabolites across sample types revealed that central carbon metabolism in cancer, protein digestion and absorption, and aminoacyl-tRNA biosynthesis represent key pathways associated with these differential metabolites in three sample categories.

Therefore, this study is a discovery phase research, suggesting that the metabolites in these three pathways may play a crucial role in the progression of NSCLC.

ORCID iD

Dewei Li https://orcid.org/0000-0002-9429-3881

Ethical Considerations

All procedures and experiments were approved by the Clinical Research Ethics Committee of Xinjiang Uygur Autonomous Region People’s Hospital (No: KY2023060145, Urumqi, Xinjiang, China) and complied with the Declaration of Helsinki. Moreover, this study has been publicly registered in the ClinicalTrials.gov (registration number: NCT07532447).

Consent to Participate

All patients recruited for this study signed written informed consent forms (et al, including but not limited to blood, surgical tissue resection and EBC) inclusion to confirm their willingness to provide biological samples for scientific research.

Data Availability Statement

All study data are included in the article and/or Supporting information. If you would like to obtain any information, please contact Dewei Li (ldw980608@163.com).