Abstract
Lyme disease is the most common vector-borne infection in North America and Europe, yet current two-tier serologic testing shows poor sensitivity during early infection when antibiotic treatment is most effective. We developed and validated a single-tier multiplex immunoassay that combines ten Borrelia burgdorferi antigens with machine-learning classification to detect antibody responses across all disease stages. A total of 364 cases were obtained, which included well characterized prospective blood samples obtained during early infection. In a cohort of samples from the Johns Hopkins University SLICE studies, the classifier identified all 30 early Lyme disease cases and their 1–3-month post-treatment follow-up samples. To assess generalizability, an independent EU laboratory synthesized the antigens de novo and a distinct 259-person cohort was evaluated on newly built assays. The classifier achieved an AUC of 0.98 in both cohorts. This single-test platform delivers quantitative serologic classification using standard clinical laboratory instrumentation, addressing critical gaps in Lyme disease diagnosis.

This publication has 51 references indexed in Scilit: