ABSTRACT Tuberculosis (TB) is a major cause of pediatric morbidity and mortality (1.2 million new cases and 224,000 deaths in 2021), with most deaths occurring in children < 5 years who do not initiate treatment. These children often present with disseminated or extrapulmonary TB as their immune systems may fail to contain their M. tuberculosis (Mtb) infections in their lungs. Since host responses play critical roles in TB development, analysis of Mtb- and host-derived could improve early diagnosis and TB management. However, TB tests still primarily rely on microbiologic evidence from respiratory samples that are difficult to obtain from, and less informative for, young children. Sputum culture detects only 30-60% of pediatric TB, and similar results are obtained with PCR- based Xpert MTB/RIF and Xpert Ultra assays. There is thus an urgent need for non-sputum-based tests that can effectively diagnose pediatric TB and monitor its response to treatment. Serum biomarker assays are ideal for pediatric TB diagnosis. The assay used in our initial proposal detects low abundance Mtb proteins in serum to effectively diagnose pediatric TB and has since received breakthrough device status by the FDA prior to pending validation studies for its FDA clearance. We have further shown that we can employ a nanoparticle-enhanced immunoassay (NEI) read by an inexpensive portable device to detect Mtb-derived factors on extracellular vesicles (EVs) for TB diagnosis, an approach suitable for use in resource- limited settings. We have also shown that a simple modification can markedly increase sensitivity and reproducibility to improve measurements. Biomarkers on EVs secreted by Mtb-infected phagocytes and by immune cells may reflect Mtb burden and its containment by the immune system, and provide critical information to required improve pediatric TB diagnosis and to allow rapid and accurate assessment of effective anti-TB treatment. However, few studies have evaluated the diagnostic and prognostic performance EV-based Mtb assays, and none have evaluated the ability of host-derived EV biomarkers to improve assay performance. We now propose modify our initial NEI test to incorporate both Mtb- and host-derived EV markers associated with TB disease in order to enhance its performance for TB diagnosis and treatment monitoring. We will employ the following aims to develop and characterize this modified assay: 1) We will optimize its EV biomarker signature using additional Mtb- and host-derived proteins we will identify from serum EVs of TB cases and controls. 2) We will conduct an analytical validation of this refined NEI to establish its characteristics using STARD guidelines. 3) We will establish a prediction model of NEI array’s multiplex performance for TB diagnosis and treatment evaluation using longitudinal serum and data from a retrospective cohort of HIV-exposed infants at increased risk for TB (IMPAACT P1041 cohort) with longitudinal samples. 4) We will validate...