The performance and yield of tuberculosis testing algorithms using microscopy, chest x-ray, and Xpert MTB/RIF.

SETTING: The introduction of Xpert MTB/RIF (Xpert) and renewed interest in chest x-ray (CXR) for tuberculosis testing has provided additional choices to the smear-based diagnostic algorithms used by TB programs previously. More programmatic data is needed to better understand the implications of possible approaches. OBJECTIVE: We sought to evaluate how different testing algorithms using microscopy, Xpert and CXR impacted the number of people detected with TB in a district hospital in Nepal. DESIGN: Consecutively recruited patients with TB-related symptoms were offered smear microscopy, CXR and Xpert. We tested six hypothetical algorithms and compared yield, bacteriologically positive (Bac+) cases missed, and tests conducted. RESULTS: Among 929 patients, Bac+ prevalence was 17.3% (n = 161). Smear microscopy detected 121 (75.2% of Bac+). Depending on the radiologists' interpretation of CXR, Xpert testing could be reduced by (31%-60%). Smear microscopy reduced Xpert cartridge need slightly, but increased the overall diagnostic tests performed. CONCLUSION: Xpert detected a large proportion of Bac+ TB cases missed by microscopy. CXR was useful in greatly reducing the number of diagnostic tests needed even among presumptive TB patients. Loose CXR readings should be used to identify more people for TB testing. More analysis of costs and standardized CXR reading should be considered.

Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems.

Deep learning (DL) neural networks have only recently been employed to interpret chest radiography (CXR) to screen and triage people for pulmonary tuberculosis (TB). No published studies have compared multiple DL systems and populations. We conducted a retrospective evaluation of three DL systems (CAD4TB, Lunit INSIGHT, and qXR) for detecting TB-associated abnormalities in chest radiographs from outpatients in Nepal and Cameroon. All 1196 individuals received a Xpert MTB/RIF assay and a CXR read by two groups of radiologists and the DL systems. Xpert was used as the reference standard. The area under the curve of the three systems was similar: Lunit (0.94, 95% CI: 0.93-0.96), qXR (0.94, 95% CI: 0.92-0.97) and CAD4TB (0.92, 95% CI: 0.90-0.95). When matching the sensitivity of the radiologists, the specificities of the DL systems were significantly higher except for one. Using DL systems to read CXRs could reduce the number of Xpert MTB/RIF tests needed by 66% while maintaining sensitivity at 95% or better. Using a universal cutoff score resulted different performance in each site, highlighting the need to select scores based on the population screened. These DL systems should be considered by TB programs where human resources are constrained, and automated technology is available.