Optimising tongue-swab Xpert MTB/RIF Ultra testing as a diagnostic system

I read with interest the multicountry study by Margaretha de Vos and colleagues, published in The Lancet Microbe, evaluating tongue swabs for tuberculosis diagnosis using the Xpert MTB/RIF Ultra system.1 The study represents an important contribution to non-sputum tuberculosis diagnostics. However, the low sensitivity reported for patients with trace or very-low sputum Ultra levels, together with between-site variability, suggests that future evaluations should not be framed merely in terms of assay or platform optimisation.