Field testing of addicad: An innovative diagnostic test for tuberculosis that adds to cad

Tuberculosis (TB) remains the leading cause of infectious disease-related deaths worldwide, with undiagnosed and late-stage cases driving continued transmission and poor treatment outcomes. Despite advancements in TB diagnostics, existing tests are either too costly, too complex, limiting their accessibility in resource-constrained settings. There is an urgent need for a rapid, affordable, non-sputum-based diagnostic tool at point of care.We propose AddiCAD, a novel diagnostic solution that combines AI chest X-ray analysis with a fingerstick-based biomarker test. By integrating CAD4TB scores with 3P-TB, a three-marker host protein biosignature test (SAA-1, CRP, IP-10), AddiCAD enhances diagnostic accuracy beyond current standalone methods.In a study involving 592 participants from two African countries, AddiCAD demonstrated exceptional performance, achieving an AUC of 0.96 (0.93–0.99). Optimized for triage testing, it showed 91.2% sensitivity and 89.5% specificity, marking a 20% improvement in specificity compared to CAD4TB alone, without sacrificing sensitivity. When optimized for confirmatory testing, it reached 93% sensitivity and 88.2% specificity, surpassing all previously reported non-sputum-based TB diagnostics.To bring this innovation to scale, we will:1)Develop and optimize an aptamer-based sensor for the 3P-TB signature, and a software application for integration with CAD4TB, while ensuring it is affordable and scalable.2)Validate AddiCAD in 1,000 adults with presumptive TB across three African countries with high TB and HIV incidence, using microbiological confirmation as the gold standard.3)Engage with key stakeholders, including regulatory bodies, scientific communities, and commercial partners, to support widespread adoption and integration into existing healthcare systems.With its ability to provide rapid, accurate, and accessible TB diagnosis, AddiCAD has the potential to fill a critical gap in high-burden, resource-limited settings.