Browsing by Author "Nathavitharana R"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Chest X-ray Analysis With Deep Learning-Based Software as a Triage Test for Pulmonary Tuberculosis: An Individual Patient Data Meta-Analysis of Diagnostic Accuracy.
    (2022-Apr-28) Tavaziva G; Harris M; Abidi SK; Geric C; Breuninger M; Dheda K; Esmail A; Muyoyeta M; Reither K; Majidulla A; Khan AJ; Campbell JR; David PM; Denkinger C; Miller C; Nathavitharana R; Pai M; Benedetti A; Ahmad Khan F
    BACKGROUND: Automated radiologic analysis using computer-aided detection software (CAD) could facilitate chest X-ray (CXR) use in tuberculosis diagnosis. There is little to no evidence on the accuracy of commercially available deep learning-based CAD in different populations, including patients with smear-negative tuberculosis and people living with human immunodeficiency virus (HIV, PLWH). METHODS: We collected CXRs and individual patient data (IPD) from studies evaluating CAD in patients self-referring for tuberculosis symptoms with culture or nucleic acid amplification testing as the reference. We reanalyzed CXRs with three CAD programs (CAD4TB version (v) 6, Lunit v3.1.0.0, and qXR v2). We estimated sensitivity and specificity within each study and pooled using IPD meta-analysis. We used multivariable meta-regression to identify characteristics modifying accuracy. RESULTS: We included CXRs and IPD of 3727/3967 participants from 4/7 eligible studies. 17% (621/3727) were PLWH. 17% (645/3727) had microbiologically confirmed tuberculosis. Despite using the same threshold score for classifying CXR in every study, sensitivity and specificity varied from study to study. The software had similar unadjusted accuracy (at 90% pooled sensitivity, pooled specificities were: CAD4TBv6, 56.9% [95% confidence interval {CI}: 51.7-61.9]; Lunit, 54.1% [95% CI: 44.6-63.3]; qXRv2, 60.5% [95% CI: 51.7-68.6]). Adjusted absolute differences in pooled sensitivity between PLWH and HIV-uninfected participants were: CAD4TBv6, -13.4% [-21.1, -6.9]; Lunit, +2.2% [-3.6, +6.3]; qXRv2: -13.4% [-21.5, -6.6]; between smear-negative and smear-positive tuberculosis was: were CAD4TBv6, -12.3% [-19.5, -6.1]; Lunit, -17.2% [-24.6, -10.5]; qXRv2, -16.6% [-24.4, -9.9]. Accuracy was similar to human readers. CONCLUSIONS: For CAD CXR analysis to be implemented as a high-sensitivity tuberculosis rule-out test, users will need threshold scores identified from their own patient populations and stratified by HIV and smear status.
  • Thumbnail Image
    Item
    Diagnostic yield as an important metric for the evaluation of novel tuberculosis tests: rationale and guidance for future research.
    (2024-Jul) Broger T; Marx FM; Theron G; Marais BJ; Nicol MP; Kerkhoff AD; Nathavitharana R; Huerga H; Gupta-Wright A; Kohli M; Nichols BE; Muyoyeta M; Meintjes G; Ruhwald M; Peeling RW; Pai NP; Pollock NR; Pai M; Cattamanchi A; Dowdy DW; Dewan P; Denkinger CM
    Better access to tuberculosis testing is a key priority for fighting tuberculosis, the leading cause of infectious disease deaths in people. Despite the roll-out of molecular WHO-recommended rapid diagnostics to replace sputum smear microscopy over the past decade, a large diagnostic gap remains. Of the estimated 10·6 million people who developed tuberculosis globally in 2022, more than 3·1 million were not diagnosed. An exclusive focus on improving tuberculosis test accuracy alone will not be sufficient to close the diagnostic gap for tuberculosis. Diagnostic yield, which we define as the proportion of people in whom a diagnostic test identifies tuberculosis among all people we attempt to test for tuberculosis, is an important metric not adequately explored. Diagnostic yield is particularly relevant for subpopulations unable to produce sputum such as young children, people living with HIV, and people with subclinical tuberculosis. As more accessible non-sputum specimens (eg, urine, oral swabs, saliva, capillary blood, and breath) are being explored for point-of-care tuberculosis testing, the concept of yield will be of growing importance. Using the example of urine lipoarabinomannan testing, we illustrate how even tests with limited sensitivity can diagnose more people with tuberculosis if they enable increased diagnostic yield. Using tongue swab-based molecular tuberculosis testing as another example, we provide definitions and guidance for the design and conduct of pragmatic studies that assess diagnostic yield. Lastly, we show how diagnostic yield and other important test characteristics, such as cost and implementation feasibility, are essential for increased effective population coverage, which is required for optimal clinical care and transmission impact. We are calling for diagnostic yield to be incorporated into tuberculosis test evaluation processes, including the WHO Grading of Recommendations, Assessment, Development, and Evaluations process, providing a crucial real-life implementation metric that complements traditional accuracy measures.