RESUMO
BACKGROUND: Computer-aided detection (CAD) can help identify people with active tuberculosis left undetected. However, few studies have compared the performance of commercially available CAD products for screening in high tuberculosis and high HIV settings, and there is poor understanding of threshold selection across products in different populations. We aimed to compare CAD products' performance, with further analyses on subgroup performance and threshold selection. METHODS: We evaluated 12 CAD products on a case-control sample of participants from a South African tuberculosis prevalence survey. Only those with microbiological test results were eligible. The primary outcome was comparing products' accuracy using the area under the receiver operating characteristic curve (AUC) against microbiological evidence. Threshold analyses were performed based on pre-defined criteria and across all thresholds. We conducted subgroup analyses including age, gender, HIV status, previous tuberculosis history, symptoms presence, and current smoking status. FINDINGS: Of the 774 people included, 516 were bacteriologically negative and 258 were bacteriologically positive. Diverse accuracy was noted: Lunit and Nexus had AUCs near 0·9, followed by qXR, JF CXR-2, InferRead, Xvision, and ChestEye (AUCs 0·8-0·9). XrayAME, RADIFY, and TiSepX-TB had AUC under 0·8. Thresholds varied notably across these products and different versions of the same products. Certain products (Lunit, Nexus, JF CXR-2, and qXR) maintained high sensitivity (>90%) across a wide threshold range while reducing the number of individuals requiring confirmatory diagnostic testing. All products generally performed worst in older individuals, people with previous tuberculosis, and people with HIV. Variations in thresholds, sensitivity, and specificity existed across groups and settings. INTERPRETATION: Several previously unevaluated products performed similarly to those evaluated by WHO. Thresholds differed across products and demographic subgroups. The rapid emergence of products and versions necessitates a global strategy to validate new versions and software to support CAD product and threshold selections. FUNDING: Government of Canada.
RESUMO
BACKGROUND: Tuberculosis remains an important clinical and public health issue in South Africa, which has one of the highest tuberculosis burdens in the world. We aimed to estimate the burden of bacteriologically confirmed pulmonary tuberculosis among people aged 15 years or older in South Africa. METHODS: This multistage, cluster-based, cross-sectional survey included eligible residents (age ≥15 years, who had slept in a house for ≥10 nights in the preceding 2 weeks) in 110 clusters nationally (cluster size of 500 people; selected by probability proportional-to-population size sampling). Participants completed face-to-face symptom questionnaires (for cough, weight loss, fever, and night sweats) and manually read digital chest X-ray screening. Screening was recorded as positive if participants had at least one symptom or an abnormal chest X-ray suggestive of tuberculosis, or a combination thereof. Sputum samples from participants who were screen-positive were tested by the Xpert MTB/RIF Ultra assay (first sample) and Mycobacteria Growth Indicator Tube culture (second sample), with optional HIV testing. Participants with a positive Mycobacterium tuberculosis complex culture were considered positive for bacteriologically confirmed pulmonary tuberculosis; when culture was not positive, participants with a positive Xpert MTB/RIF Ultra result with an abnormal chest X-ray suggestive of active tuberculosis and without current or previous tuberculosis were considered positive for bacteriologically confirmed pulmonary tuberculosis. FINDINGS: Between Aug 15, 2017, and July 28, 2019, 68â771 people were enumerated from 110 clusters, with 53â250 eligible to participate in the survey, of whom 35â191 (66·1%) participated. 9066 (25·8%) of 35â191 participants were screen-positive and 234 (0·7%) were identified as having bacteriologically confirmed pulmonary tuberculosis. Overall, the estimated prevalence of bacteriologically confirmed pulmonary tuberculosis was 852 cases (95% CI 679-1026) per 100â000 population; the prevalence was highest in people aged 35-44 years (1107 cases [95% CI 703-1511] per 100â000 population) and those aged 65 years or older (1104 cases [680-1528] per 100â000 population). The estimated prevalence was approximately 1·6 times higher in men than in women (1094 cases [95% CI 835-1352] per 100â000 population vs 675 cases [494-855] per 100â000 population). 135 (57·7%) of 234 participants with tuberculosis screened positive by chest X-ray only, 16 (6·8%) by symptoms only, and 82 (35·9%) by both. 55 (28·8%) of 191 participants with tuberculosis with known HIV status were HIV-positive. INTERPRETATION: Pulmonary tuberculosis prevalence in this survey was high, especially in men. Despite the ongoing burden of HIV, many participants with tuberculosis in this survey did not have HIV. As more than half of the participants with tuberculosis had an abnormal chest X-ray without symptoms, prioritising chest X-ray screening could substantially increase case finding. FUNDING: Global Fund, Bill & Melinda Gates Foundation, USAID.
