Inflammation associated with rheumatic heart disease (RHD) is influenced by gene polymorphisms and inflammatory cytokines. There are currently no immunologic and genetic markers to discriminate latent versus clinical patients, critical to predict disease evolution. Employing machine-learning, we searched for predictors that could discriminate latent versus clinical RHD, and eventually identify latent patients that may progress to clinical disease..