Browsing by Author "Dheda K"

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 of urine lipoarabinomannan and sputum tuberculosis tests in people living with HIV: a systematic review and meta-analysis of individual participant data.
    (2023-Jun) Broger T; Koeppel L; Huerga H; Miller P; Gupta-Wright A; Blanc FX; Esmail A; Reeve BWP; Floridia M; Kerkhoff AD; Ciccacci F; Kasaro MP; Thit SS; Bastard M; Ferlazzo G; Yoon C; Van Hoving DJ; Sossen B; García JI; Cummings MJ; Wake RM; Hanson J; Cattamanchi A; Meintjes G; Maartens G; Wood R; Theron G; Dheda K; Olaru ID; Denkinger CM
    BACKGROUND: Sputum is the most widely used sample to diagnose active tuberculosis, but many people living with HIV are unable to produce sputum. Urine, in contrast, is readily available. We hypothesised that sample availability influences the diagnostic yield of various tuberculosis tests. METHODS: In this systematic review and meta-analysis of individual participant data, we compared the diagnostic yield of point-of-care urine-based lipoarabinomannan tests with that of sputum-based nucleic acid amplification tests (NAATs) and sputum smear microscopy (SSM). We used microbiologically confirmed tuberculosis based on positive culture or NAAT from any body site as the denominator and accounted for sample provision. We searched PubMed, Web of Science, Embase, African Journals Online, and clinicaltrials.gov from database inception to Feb 24, 2022 for randomised controlled trials, cross-sectional studies, and cohort studies that assessed urine lipoarabinomannan point-of-care tests and sputum NAATs for active tuberculosis detection in participants irrespective of tuberculosis symptoms, HIV status, CD4 cell count, or study setting. We excluded studies in which recruitment was not consecutive, systematic, or random; provision of sputum or urine was an inclusion criterion; less than 30 participants were diagnosed with tuberculosis; early research assays without clearly defined cutoffs were tested; and humans were not studied. We extracted study-level data, and authors of eligible studies were invited to contribute deidentified individual participant data. The main outcomes were the tuberculosis diagnostic yields of urine lipoarabinomannan tests, sputum NAATs, and SSM. Diagnostic yields were predicted using Bayesian random-effects and mixed-effects meta-analyses. This study is registered with PROSPERO, CRD42021230337. FINDINGS: We identified 844 records, from which 20 datasets and 10 202 participants (4561 [45%] male participants and 5641 [55%] female participants) were included in the meta-analysis. All studies assessed sputum Xpert (MTB/RIF or Ultra, Cepheid, Sunnyvale, CA, USA) and urine Alere Determine TB LAM (AlereLAM, Abbott, Chicago, IL, USA) in people living with HIV aged 15 years or older. Nearly all (9957 [98%] of 10 202) participants provided urine, and 82% (8360 of 10 202) provided sputum within 2 days. In studies that enrolled unselected inpatients irrespective of tuberculosis symptoms, only 54% (1084 of 1993) of participants provided sputum, whereas 99% (1966 of 1993) provided urine. Diagnostic yield was 41% (95% credible interval [CrI] 15-66) for AlereLAM, 61% (95% Crl 25-88) for Xpert, and 32% (95% Crl 10-55) for SSM. Heterogeneity existed across studies in the diagnostic yield, influenced by CD4 cell count, tuberculosis symptoms, and clinical setting. In predefined subgroup analyses, all tests had higher yields in symptomatic participants, and AlereLAM yield was higher in those with low CD4 counts and inpatients. AlereLAM and Xpert yields were similar among inpatients in studies enrolling unselected participants who were not assessed for tuberculosis symptoms (51% vs 47%). AlereLAM and Xpert together had a yield of 71% in unselected inpatients, supporting the implementation of combined testing strategies. INTERPRETATION: AlereLAM, with its rapid turnaround time and simplicity, should be prioritised to inform tuberculosis therapy among inpatients who are HIV-positive, regardless of symptoms or CD4 cell count. The yield of sputum-based tuberculosis tests is undermined by people living with HIV who cannot produce sputum, whereas nearly all participants are able to provide urine. The strengths of this meta-analysis are its large size, the carefully harmonised denominator, and the use of Bayesian random-effects and mixed-effects models to predict yields; however, data were geographically restricted, clinically diagnosed tuberculosis was not considered in the denominator, and little information exists on strategies for obtaining sputum samples. FUNDING: FIND, the Global Alliance for Diagnostics.