Streamlined Point-of-Care Ultrasound Examination for Pulmonary Tuberculosis

Background: Point-of-care ultrasound (POCUS) is a low-cost, radiation-free triage tool but lacks standardized acquisition protocols for tuberculosis (TB). We developed and evaluated a streamlined symmetric POCUS protocol for pulmonary TB in a high-burden setting. Methods: Ugandan participants aged ≥13 years with confirmed pulmonary TB, household TB exposure, or a non-TB respiratory condition underwent an 18-position POCUS examination (12 lung zones and 6 pre-established extrapulmonary positions). Random Forest classifiers using prespecified pathologic features were developed in a fixed stratified 75/25 training/holdout split. All symmetric 2-, 4-, and 6-zone combinations were evaluated in training data. The best protocol of each size was locked and tested in the holdout set. Results: Among 318 analyzed participants, 118 (37.1%) had pulmonary TB and 200 were TB-negative, including 102 with other respiratory conditions (ORC). The model-selected 6-zone protocol comprised bilateral anterior-superior, lateral-superior, and posterior-superior lung zones. For TB versus all TB-negative participants, the AUCs for the 6-zone, 12-zone, and 18-position protocols were 0.82 (95% CI, 0.73-0.92), 0.81 (0.70—0.92), and 0.81 (0.71—0.91), respectively (all P > 0.76). For TB versus ORC, AUCs were 0.81 (0.70—0.91), 0.85 (0.76—0.93), and 0.85 (0.76—0.94), respectively (all P > 0.45). The 2-zone protocol had a lower AUC than the 6-zone protocol for TB versus all TB-negative participants (P = 0.004). Shred sign, consolidation, and B-lines with irregular pleura were the strongest TB-associated findings. Conclusions: A fixed 6-zone lung POCUS examination showed similar discrimination for TB to more comprehensive 12-lung-zone and 18-position protocols. This supports a simpler, scalable approach to TB triage