Per year, globally an estimated one million children develop tuberculosis (TB) and more than 15 million
children are estimated to be exposed to Mycobacterium tuberculosis (Mtb). The case fatality rate is high in
children < 5 years of age. Current approaches to diagnosis and management of young children that are close
contacts to a TB case are inadequate. Those that are symptomatic may undergo sputum-based diagnostics
that are not well tolerated (eg gastric aspirates), require access to a reference laboratory, and are not
sensitive because TB may be paucibacillary or extrapulmonary. For that reason empirical multidrug anti-TB
treatment predominates in many locales. Management of the asymptomatics is sub-optimal as well. Given
the poor performance of IGRAs and TST in this age group, most are treated with isoniazid preventive therapy
(IPT). In adults, asymptomatic (subclinical TB) is at least as common as active TB and will not be detected
by current symptom-based screening. We do not know how often this is the case in exposed children,
however, IPT, would be inadequate in them. Further, about 19% of children in this age group with latent TB
infection (LTBI) will progress to active TB, usually within the next 3-6 in the absence of IPT (and IPT is only
63% effective). The need therefore is to discover a biomarker or biomarkers that identify those children < 5
years of age with subclinical TB (likely to progress despite IPT); and those without subclinical TB that are
likely to progress. These biomarkers would allow appropriate targeting of IPT and ATT to those likely to
benefit. This consortium of investigators have on-going diagnostic and cohort studies of child (< 5 years of
age) close contacts of TB cases in Uganda that include a rigorous bacteriologic reference standard applied
to asymptomatic as well as symptomatics and evaluation of novel diagnostics and discovery of non-sputum-
based approaches. We propose now to evaluate in children < 5years old that are close contacts of a TB
case a diverse and complementary panel of bacterial, host-based and imaging non-sputum biomarkers that
have shown promise as predictors of progression in adults. Further, we will discover relevant biomarkers in
this population through an unbiased multi-omics approach using proteomics, single-cell omics, T-cell
activation markers, antigen-specific antibody profiling, Mtb exosomal assays, computer-aided detection
(CAD) for chest X-ray interpretation and point-of-care ultrasound (POCUS). Our goal is to characterize a
biomarker or group of biomarkers that meet a minimal target performance profile to identify children with
subclinical TB and/or at high risk of progression. We will apply advanced machine learning and integrative
multiomics to identify combinations of these biomarker signatures alongside TB risk variables to improve
precision of predicting progression. These results will provide novel approaches to risk-stratify children <5
years of age for targeting the administration of preventive therapy and ATT.