Clinical Validation of Plasma Metabolite Markers for Early Lung Cancer Detection

Abstract

Early detection of lung cancer significantly improves survival, yet current screening methods have limitations. This study aimed to identify a robust panel of plasma metabolites for early-stage non-small cell lung cancer (NSCLC) diagnosis using a large, clinically diverse patient cohort. A total of 680 archived plasma samples from biopsy-confirmed NSCLC patients and controls (including healthy individuals and patients with non-cancerous lung diseases) were analyzed using targeted, quantitative mass spectrometry-based metabolomics and used as the discovery cohort. An independent set of 216 plasma samples served as the validation cohort. Logistic regression (LR) models developed from the discovery set using ten metabolites achieved area under the receiver-operating characteristic curve (AUROC) values of 93.63%, 93.74%, and 93.91% for distinguishing all-stage, stage I-II, and stage I NSCLC patients from controls, respectively. Incorporating smoking history further improved model performance. The validation cohort confirmed the model's robustness, demonstrating high sensitivity and specificity for early-stage detection. These results support the potential of metabolomic biomarkers as a minimally invasive, accurate tool for early NSCLC diagnosis. This approach may complement current screening methods, enabling earlier intervention and improved patient outcomes. Further studies are warranted to validate these findings in more diverse populations and real-world clinical settings.

Keywords: early detection; metabolite markers; non-small cell lung cancer.

Figure 1

Metabolomic profiles of lung cancer patients and controls are significantly different. (A) PLS-DA 2D-scores plot of all-stages lung cancer patients vs. controls. (B) A hierarchical clustering heat map of the metabolites measured in the plasma of lung cancer patients and healthy controls. Only the top 25 metabolites were shown. (C,D) PLS-DA 2D-scores plot of early stages (stages I + II) lung cancer patients (C), and Stage I lung cancer patients (D) vs. controls.

Figure 2

Logistic regression modeling can effectively discriminate lung cancer patients from controls: (A–C) ROC curves generated by the logistic regression models using metabolite features only for lung cancer patients at all stages (A), early stages (stage I + stage II) (B), and stage I (C). ROC curves and their 95% CI on the discovery set are shown in magenta. ROC curves obtained from the validation set are colored in cyan. (D–F) ROC curves generated by the logistic regression models using both metabolite features and smoking factor for lung cancer patients at all stages (A), early stages (stage I + stage II) (C), and stage I (E). ROC curves and their 95% CI on the discovery set are shown in red. ROC curves obtained from the validation set are colored in blue. (G–I) Precision–recall curves of the logistic regression models for lung cancer patients at all stages (B), early stages (stage I + stage II) (D), and stage I (F). Curves of the discovery set and the validation set are colored in red and blue, respectively.