BackgroundCirculating immunometabolic biomarkers show promise for the diagnosis and treatment monitoring of tuberculosis (TB). However, biomarkers that can distinguish TB from nontuberculous mycobacteria (NTM) infections, latent tuberculosis infection (LTBI), and other lung diseases (ODx) have not been elucidated. This study utilized a multi-cohort, multi-omics approach combined with predictive modeling to identify, validate, and prioritize biomarkers for the diagnosis of active TB.MethodsFunctional omics data were collected from two discovery cohorts (76 patients in the TB-NTM cohort and 72 patients in the TB-LTBI-ODx cohort) and one validation cohort (68 TB patients and 30 LTBI patients). An integrative multi-omics analysis was performed to identify the plasma multi-ome biosignatures. Machine learning-based predictive modeling was then applied to assess the performance of these biosignatures and prioritize the most promising candidates.ResultsConventional statistical analyses of immune profiling and metabolomics indicated minor differences between active TB and non-TB groups, whereas the lipidome showed significant alteration. Muti-omics integrative analysis identified three multi-ome biosignatures that could distinguish active TB from non-TB with promising performance, achieving area under the ROC curve (AUC) values of 0.7–0.9 across groups in both the discovery and validation cohorts. The lipid PC(14:0_22:6) emerged as the most important predictor for differentiating active TB from non-TB controls, consistently presenting at lower levels in the active TB group compared with counterparts. Further validation using two independent external datasets demonstrated AUCs of 0.77–1.00, confirming the biomarkers’ efficacy in distinguishing TB from other non-TB groups.ConclusionOur integrative multi-omics reveals significant immunometabolic alteration in TB. Predictive modeling suggests lipids as promising biomarkers for TB-NTM differential diagnosis and TB-LTBI-ODx diagnosis. External validation further indicates PC(14:0_22:6) as a potential diagnostic marker candidate for TB.SummaryOur multi-cohort, multi-omics data integration and predictive modeling identified reliable biomarkers and highlighted the importance of circulating lipids for distinguishing tuberculosis (TB) from complex conditions with similar clinical manifestations, latent infections, and healthy individuals.