Dr. Véronique Suttels’ PhD research focused on the diagnostic performance and scalability of point-of-care lung ultrasound (LUS)—including expert- and AI-assisted interpretation—as a triage tool for tuberculosis (TB) in endemic, resource-limited settings.

In a prospective cohort study of 504 symptomatic adults in Benin, AI-interpreted LUS demonstrated excellent diagnostic performance, achieving a sensitivity of 91% and specificity of 85%, thereby fulfilling WHO-defined criteria for a non-sputum-based triage test. In comparison, human expert interpretation showed lower specificity. These findings represent the first robust evidence supporting AI-driven LUS as a frontline diagnostic triage tool for TB.

Her work also evaluated the scalability and implementation potential of LUS, identifying both barriers and facilitators in low-resource settings. Key enablers included clinical enthusiasm and openness to AI integration, while challenges involved technical skill development and device sustainability. Her study demonstrated that web-based training and skill evaluation for LUS is feasible even in low-resource environments.

The research culminated in the successful development of CAD LUS4TB, an EDCTP3-funded multicountry initiative aimed at validating and implementing AI-assisted LUS for TB triage across Benin, Mali, and South Africa. This project integrates validation, health economics, open science, and capacity building to ensure wide-scale impact. This project will cover the salary of Dr. Suttels for the next 5 years. She is the scientific co-lead of the project.

Overall, Dr. Suttels’ work significantly advances non-invasive, scalable TB diagnostics and provides a promising path forward for equitable, AI-supported solutions in global TB control.