To illustrate the practical application and inherent challenges of clinical peptidomics, a small-scale, preliminary study was performed. The primary aim was exploratory: To assess the feasibility of generating differential serum peptide profiles between malignant and benign pulmonary nodules. The study cohort consisted of 32 patients from the Affiliated Hospital of Jining Medical University (Jining, China) between January 2025 and June 2025; this included 29 patients with pathologically confirmed lung cancer (the cancer group) and 3 patients with benign pulmonary nodules (the control group). The inclusion criteria were patients aged 18–75 years with a confirmed diagnosis of lung cancer or benign pulmonary nodules. Exclusion criteria included patients with previous treatment (chemotherapy, radiotherapy or surgery), active infections, or any history of other malignancies. Benign nodule status was confirmed either by biopsy or radiographic stability over a period of ≥2 years. Key demographic and clinical characteristics of the participants are summarized in Table I. The present study was approved by the Institutional Review Board of the Institute of Medical Science, Affiliated Hospital of Jining Medical University (approval no. 2024-02-C011).

Peripheral blood samples (5 ml) were collected from all participants. Serum was separated by centrifugation at 4°C for 10 min at 500 × g, then aliquoted and stored at −80°C until analysis. Upon receipt of peptidomic profiling, all samples were transported on dry ice and remained frozen. Visual inspection confirmed the absence of hemolysis or lipemia in all samples.

Serum peptidome analysis, including peptide extraction, matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) analysis and quality control (QC), was performed by Hangzhou Well-healthcare Technologies Co., Ltd., using their proprietary platform. Peptides were extracted from serum using the Well-healthcare Serum Peptide Detection Kit according to the manufacturer's protocol. Briefly, 10 µl serum was mixed with 70 µl ultrapure water and added to a 96-well plate pre-loaded with nanoporous silica microspheres. The mixture was incubated at room temperature with agitation at 1,350 rpm for 15 min. After incubation, the supernatant was removed by centrifugation at 4°C for 10 min at 500 × g. The pellet was washed once with 80 µl ultrapure water and centrifuged again at 4°C for 10 min at 500 × g to remove the supernatant.

The extracted peptides were eluted with a matrix solution containing acetonitrile and trifluoroacetic acid, mixed with an excess of α-cyano-4-hydroxycinnamic acid matrix. A 1-µl aliquot of the mixture was spotted onto a stainless-steel MALDI target plate and allowed to dry under vacuum. MS analysis was performed using a ClinMS-Plat I MALDI-TOF MS instrument (Hangzhou Well-healthcare Technologies Co., Ltd). Data were acquired in positive ion linear mode with the following parameters: Extraction voltage, 2.00 kV; detector voltage, 2.30 kV; delay time, 200 nsec. For each spectrum, 800 laser shots were accumulated across a mass range of 680-18,600 Da.

Rigorous QC measures were implemented. On each target plate, the following control spots were included alongside patient samples: One negative QC (N-QC; from pooled negative serum, prepared by mixing negative serum samples from healthy individuals), one positive QC (P-QC; from pooled positive serum, prepared by mixing serum from patients with lung cancer), and three baseline QCs (B-QC; from pooled negative serum, prepared similarly to N-QC). The positions of N-QC, P-QC, B-QC and calibration standards were randomly distributed on the target plate. QC samples underwent the same pre-treatment process as the test samples. System suitability required B-QC spectral similarity scores to be ≤0.34 (Euclidean and Manhattan distances), N-QC to test negative (score ≤0) and P-QC to test positive (score >0). All samples passed QC criteria, with a minimum of 500 laser shots and a minimum peak detection proportion of 50% per spectrum.

Raw mass spectral data were preprocessed using MDAS PreData software (version 2.23; Well-healthcare Technologies Co., Ltd.). This included baseline correction, smoothing, peak alignment and peak picking. A total of 444 mass spectral peak signals were detected across all samples. These peaks were matched against the PeptideAtlas database (http://www.peptideatlas.org), resulting in the identification of 188 unique peptide sequences.

For statistical comparison between the cancer (n=29) and control (n=3) groups, peptide peak intensity values were used. Peptide signals detected in <80% of the samples within either group were filtered out. Peptides with an absolute log2 fold-change (|log2FC|) >0.2 and P<0.05 were considered differentially expressed. No correction for multiple testing (for example, false discovery rate or Bonferroni) was applied due to the exploratory nature and small sample size of this pilot study.

The proteins corresponding to the identified differential peptides were subjected to functional annotation and enrichment analysis. Gene Ontology (GO) enrichment, Clusters of Orthologous Groups (COG/KOG) classification, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed using InterProScan (https://www.ebi.ac.uk/interpro), KAAS (http://www.genome.jp/kegg/kaas/), UniProt (https://www.uniprot.org), and eggNOG (http://eggnogdb.embl.de). Protein-protein interaction (PPI) networks were constructed using the STRING database (https://string-db.org) and visualized with Cytoscape software (https://cytoscape.org).

Categorical variables were presented as numbers and percentages. For comparisons between the cancer (n=29) and control (n=3) groups Fisher's exact test was used for all 2×2 contingency tables due to the small sample size, particularly in the control group. Differential peptides were identified using a non-parametric Mann-Whitney U test, suitable for comparing two independent groups with non-normally distributed data. P<0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using R software (version 4.1.3; Posit software), which can be accessed at https://www.r-project.org.

The present preliminary feasibility study included serum samples from 32 patients. The cohort consisted of 29 patients with pathologically confirmed lung cancer (the cancer group) and 3 patients with benign pulmonary nodules (the control group), as detailed in Table I. There were no statistically significant differences between the two groups in terms of age distribution (P=0.851) or sex (P=0.548) (Table I). The majority of lung cancer cases were adenocarcinoma (24/29; 82.8%) and were at an early stage (T1a-T1c, 29/29; 100%). Detailed histopathological descriptions are provided in Table II.

The average peptide spectra for the cancer (n=29) and control (n=3) groups are presented in Fig. 1A and B, respectively. Comparative analysis identified 12 peptide signals that met the pre-defined exploratory threshold for differential expression (|log2FC| >0.2 and P<0.05, Mann-Whitney U test; not corrected for multiple testing). As shown in the volcano plot (Fig. 1C), 10 peptides were upregulated and 2 were downregulated in the cancer group. The clustering heatmap (Fig. 1D) shows a general separation trend between the two groups, and the expression intensity distribution of these differential peptides is detailed in the box plots (Fig. 2).

The 12 differential peptides mapped to three parent proteins: Fibrinogen α chain (FGA), inter-α-trypsin inhibitor heavy chain H4 (ITIH4) and coagulation factor XIII A chain (F13A1). GO enrichment analysis (Fig. 3A; Table SI) revealed that these proteins were primarily enriched in biological processes such as ‘platelet alpha granule lumen’ and ‘blood coagulation, fibrin clot formation’. COG/KOG functional classification (Fig. 3B) categorized the main function of these differential proteins under ‘Posttranslational modification, protein turnover, chaperones’ (Category O), Categories K (transcription), O (posttranslational modification, protein turnover, chaperones), and S (function unknown) were also observed. KEGG pathway analysis (Fig. 3C) indicated significant enrichment (P<0.05) in four pathways: ‘Platelet activation’, ‘Complement and coagulation cascades’, ‘Neutrophil extracellular trap formation’ and ‘Coronavirus disease-COVID-19’. PPI network analysis further demonstrated connectivity among these three candidate proteins (Fig. 3D).

Based on the aforementioned findings, the present study proposes a conceptual framework for future pulmonary nodule management (Fig. 4). This framework envisages the integration of validated multi-analyte serum biomarker profiles with clinical and imaging data to achieve more precise risk stratification. The utility of this strategy is contingent upon rigorous future validation of the implicated biomarkers.