A new generation of diagnostic systems available at the point of care (POC) could save lives and reduce the spread of infectious diseases worldwide through early detection and treatment. Optical microscopy remains the gold standard for the diagnosis of many parasitological diseases; however, its accuracy is dependent on the availability and expertise of the analyst at the POC. This limitation is increased by the dependence on labour-intensive examination processes, lack of standardization, high interobserver variability, insufficient precision in sample quantification and, as a consequence, a high misdiagnosis rate. This project introduces an AI diagnostic system leveraging existing microscopes and mobile technology providing a comprehensive and holistic sample analysis rather than just detecting individual pathogens. MultiplexAI is a scalable, low-cost, autonomous AI diagnostic system for the POC that upgrades any optical microscope into an AI agent able to accurately identify any parasite in a sample. We will collect data, train, deploy and evaluate the integrated system to detect multiple diseases including malaria and parasitic Neglected Tropical Diseases. The project will pursue the following objectives and methodological steps: 1) To design a trustworthy AI system, ensuring technical and social robustness, and adherence to WHO AI ethical principles of safety, transparency, explainability, accountability, equity, and sustainability; 2) To develop AI foundational models for microscopy analysis capable of automating the detection, differentiation and quantification of multiple parasites causing disease and integrate them into an automatic mobile microscopy system; 3) To validate the system in laboratory settings; 4) To undertake a performance evaluation study in clinical workflows of four countries in SSA; 5) To assess usability, acceptability and feasibility with end-users and evaluate the cost-effectiveness of its implementation; 6) To model and evaluate the health impact of introducing our system to improve diagnosis and surveillance at both local and national level; and 7) To execute a regulatory roadmap for compliance in EU and SSA, and determine a path to market. Overall, this project aims to unleash the AI revolution leveraging mobile technologies and upgrading millions of optical microscopes into a network of intelligent POC devices, capable of performing high-throughput sample analysis to provide reliable and ubiquitous diagnostics and medical knowledge for everyone, everywhere.

The main concern of clinicians when addressing sick patients is to promptly identify -i.e., “not miss”- those at higher risk of severe disease, so as to prioritize their care and better target therapeutic interventions. Unfortunately, risk-stratification practices for infections remain suboptimal and prone to misclassification, leading to adverse outcomes and misallocation of resources, particularly in children (and even more so among newborns) from Sub-Saharan Africa (SSA). We recently developed “B-Triage”, a point-of-care rapid triaging test, designed for the quantitative assessment of sTREM-1, a biomarker of sepsis, and a highly performing prognostic marker, irrespective of underlying disease. Levels of sTREM-1 stand out as a quantitative and independent predictor of severity and death in all-cause infections, being superior to other markers and clinical scores, showing promise also for risk-stratification of non-communicable diseases. We propose to specifically validate B-Triage for risk-stratification of all cause sickness in the newborn, the age group concentrating ~50% of all child mortality. ACROBAT-newborns aims to continue and accelerate the valorisation of our device, with clinical studies in Mozambique, Ethiopia, Uganda and Gabon; the industrialization of its prototype; and a go-to market strategy for SSA. The project includes strong components of health economics and impact assessment, as well as socio-behavioural sciences (usability, acceptability, and feasibility studies), with the overarching aim of generating the necessary evidence to support B-Triage’s introduction to the African market. The proactive use of our device for risk-stratification of the sick newborn at first clinical presentation, will determine, objectively and with high precision, those at risk of severe outcome and death, resulting in improved outcomes and survival, and an optimized use of healthcare resources, including antibiotics and high value therapeutics.