Plasma metabolomics enhances risk prediction for lung cancer: a population-based validation study - PubMed
4 hours ago
- #metabolomics
- #lung-cancer
- #risk-prediction
- Identified five overlapping metabolites (decanoylcarnitine, histidine, tryptophan, L-carnitine, octanoylcarnitine) for lung cancer screening and early diagnosis.
- Developed a metabolite-based risk prediction model with an AUC of 0.878, outperforming models using only epidemiological risk factors.
- Model validated internally (AUC 0.858) and externally (AUC 0.770), showing robust performance across different datasets.
- Study design combined hospital-based case-control and population-based nested case-control studies for comprehensive validation.
- Logistic regression outperformed machine-learning approaches in predictive ability for lung cancer risk.
- Incorporating metabolomic profiles enhances high-risk individual identification, aiding precise screening and early detection.