Data resource

The MDR-PTB and XDR-PTB cases were from the NIAID TB Portals Program (TBPP) dataset, DS-PTB cases were obtained from TBPP dataset as well as the Shenzhen Center for Chronic Disease Control, Shenzhen, China between April 2017 and December 2018. We excluded patients with unsatisfactory images or without drug susceptibility test (DST). We also excluded patients with HIV(+) status, as patients with compromised immunofunction have been noted to have different PTB presentation. The patient selection workflow was shown in Fig 1. Finally, 742 eligible patients from 6 countries (488 men and 254 women, mean age, 40.0±14.8 years) were included in our study. The category of being dsptb, mdrptb or xdrptb was based on DST. According to previous anti-tuberculosis treatment or not, we further categorized all the patients into new cases (445 cases) and previously treated cases (297 cases). The cases from Shenzhen Center for Chronic Disease Control were all new cases.

Chest CT reading

All the CT images were read by one radiology trainee (SNT) with over three years’ experience of chest image reading and was checked by a specialist radiologist (YXJW). Consensus was achieved for the final reading results. This study focused on the CT feature of nodular consolidation (NC). For all NCs with diameter ≥ 6 mm, the number was counted for each patient; the diameter for each NC was measured on axial CT images showing the largest size. Some examples of the pulmonary nodules are shown in Fig 2. Location of each NC is also recorded based on the natural anatomy of lung lobes. And other chest imaging characteristics including pleural effusion, tree-bud sign, pneumothorax, bronchiectasis, pleural thickening were also recorded. To address concerns about regional differences, we also compared the incidence of nodules in Eastern Europe (Belarus, Moldova and Romania) and East Asia (China).

Statistical analysis

Data analysis was processed using GraphPad Software (GraphPad Software Inc. San Diego, CA, USA). Categorical and continuous variables were analyzed by the chi-square test and Mann-Whitney U test, respectively. A two-sided P value < 0.05 was considered statistically significant, > 0.1 as not significant, and between 0.05 and 0.1 as with a trend of significance.