Researchers from Chengdu University used saliva microbiomics and a machine-learning optimized BOXGB model to detect pulmonary nodules. With an AUC of 0.8831, the study highlights how microbial signatures like Defluviitaleaceae_UCG-011 can guide early diagnosis, offering a promising tool complementary to imaging techniques.