Xuzhen Qin, MD | Chinese Academy of Medical Sciences; Jing Zhao, MD; Ping Wang, MD; Yuan Huang, MS; Di Wang, MD; Yanlian Yang, PhD; Zhili Li, PhD; Yalong Jiang, PhD; Jiayu Xiao, MS
Competition Sponsor: Chinese Academy of Medical Sciences
Awardee Year: 2025
Lung cancer is the leading cause of cancer-related deaths worldwide. The difficulty in its early diagnosis results in a 5-year survival rate of less than 20% for patients. This project addresses the clinical challenge of differentiating benign and malignant pulmonary nodules, innovatively proposing the “multi-omics-AI fusion diagnosis” strategy: by integrating serum proteomics, exosome protein detection, quantitative glycosylation modification sites, and related clinical and imaging information, it breaks through the limitations of traditional single biomarker or imaging diagnosis. Based on the 500-patient prospective cohort established previously (IA stage lung adenocarcinoma confirmed by pathology standards vs. inflammatory/glandular benign nodules, nodules without progression, and no nodules), a standardized process is used to separate patients’ serum, combined with high-throughput multi-omics detection and feature engineering to screen key variables. Multiple algorithms are used for machine learning, analyzing the interaction network of biomarkers, constructing a risk probability model, using interpretable algorithms to provide human-understandable visual explanations for the prediction results of each patient, and verifying in an independent cohort. Expected outcomes: the sensitivity of IA stage lung cancer detection ≥ 85% (AUC > 0.93), the false positive rate reduced by 30%; 1~2 SCI papers and 2 national invention patents will be produced; an auxiliary diagnosis system will be developed, promoting the transformation of lung cancer screening from “empirical follow-up” to “precision intervention”, significantly improving the early diagnosis rate and improving the survival prognosis of patients.