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.

Results from early programmatic implementation of Xpert MTB/RIF testing in nine countries.

BACKGROUND: The Xpert MTB/RIF assay has garnered significant interest as a sensitive and rapid diagnostic tool to improve detection of sensitive and drug resistant tuberculosis. However, most existing literature has described the performance of MTB/RIF testing only in study conditions; little information is available on its use in routine case finding. TB REACH is a multi-country initiative focusing on innovative ways to improve case notification. METHODS: We selected a convenience sample of nine TB REACH projects for inclusion to cover a range of implementers, regions and approaches. Standard quarterly reports and machine data from the first 12 months of MTB/RIF implementation in each project were utilized to analyze patient yields, rifampicin resistance, and failed tests. Data was collected from September 2011 to March 2013. A questionnaire was implemented and semi-structured interviews with project staff were conducted to gather information on user experiences and challenges. RESULTS: All projects used MTB/RIF testing for people with suspected TB, as opposed to testing for drug resistance among already diagnosed patients. The projects placed 65 machines (196 modules) in a variety of facilities and employed numerous case-finding strategies and testing algorithms. The projects consumed 47,973 MTB/RIF tests. Of valid tests, 7,195 (16.8%) were positive for MTB. A total of 982 rifampicin resistant results were found (13.6% of positive tests). Of all tests conducted, 10.6% failed. The need for continuous power supply was noted by all projects and most used locally procured solutions. There was considerable heterogeneity in how results were reported and recorded, reflecting the lack of standardized guidance in some countries. CONCLUSIONS: The findings of this study begin to fill the gaps among guidelines, research findings, and real-world implementation of MTB/RIF testing. Testing with Xpert MTB/RIF detected a large number of people with TB that routine services failed to detect. The study demonstrates the versatility and impact of the technology, but also outlines various surmountable barriers to implementation. The study is not representative of all early implementer experiences with MTB/RIF testing but rather provides an overview of the shared issues as well as the many different approaches to programmatic MTB/RIF implementation.