Shouzhi Yang, Yongxia Qiao, Tianxiang Chen, Wenjun Yu, Yuewei Zhao, Zhengying Gu, Xin Xie, Fanyu Meng, Hongwei Meng, Ayizekeranmu Yiming, Chunmeng Ding, Yang Li, Kexin Meng, Wanshan Liu, Jiao Wu, Juxiang Zhang, Yida Huang, Xiaoyu Xu, Rui Kang, Jiayi Wang, Kun Qian, Daolin Tang, Lin Huang
Early diagnosis of lung adenocarcinoma (LUAD), particularly at the adenocarcinoma in situ (AIS) stage, is critical for curative intervention. Profiling blood-based metabolic progression offers a systemic view of tumorigenesis to enable robust, noninvasive early detection. To achieve this, we profile 2,836 serum samples from predominantly female, never-smoking Chinese cohorts spanning normal lung to invasive carcinoma using an ultrafast nanoparticle-enhanced mass spectrometry platform. This integrated retrospective and prospective cohort design enables the construction of a comprehensive metabolic atlas, revealing conserved metabolic perturbations that arise at the pre-invasive stage and progress with histologic severity. Guided by these trajectories, we develop a ten-metabolite diagnostic assay that demonstrates reproducible performance across independent cohorts (area under the curve 0.900 for overall; 0.860 for pre-invasive disease). Notably, assay efficacy remains consistent after adjusting for demographic and lifestyle confounders. These findings identify metabolic rewiring as an early hallmark of LUAD, with potential utility in population-level early detection.