Chest X-ray alone is insufficient for predicting drug-resistant pulmonary tuberculosis

To the Editor, Chest X-ray (CXR) in combination with machine learning (ML)based prediction models has been proposed as a promising tool to distinguish drug-susceptible from drug-resistant tuberculosis (DS-/ DR-TB) [1,2].This method builds on the idea that radiological abnormalities differ between patients with pulmonary DS-and DR-TB.Yet, previous studies have yielded conflicting results when comparing CXR findings in clinical practice [3,4].These studies have, however, also been limited by relatively small sizes or by including few types of drug resistance.The aim of this study was to examine whether CXR findings differ across various types of Mycobacterium tuberculosis resistance.From 2019 to 2022, five subgroups with susceptible, monoresistant, multidrug-resistant, pre-extensively, and extensively