ABSTRACT
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.
ABSTRACT
Tuberculosis (TB) and non-communicable diseases (NCD) share predisposing risk factors. TB-associated NCD might cluster within households affected with TB requiring shared prevention and care strategies. We conducted an individual participant data meta-analysis of national TB prevalence surveys to determine whether NCD cluster in members of households with TB. We identified eligible surveys that reported at least one NCD or NCD risk factor through the archive maintained by the World Health Organization and searching in Medline and Embase from 1 January 2000 to 10 August 2021, which was updated on 23 March 2023. We compared the prevalence of NCD and their risk factors between people who do not have TB living in households with at least one person with TB (members of households with TB), and members of households without TB. We included 16 surveys (n = 740,815) from Asia and Africa. In a multivariable model adjusted for age and gender, the odds of smoking was higher among members of households with TB (adjusted odds ratio (aOR) 1.23; 95% CI: 1.11-1.38), compared with members of households without TB. The analysis did not find a significant difference in the prevalence of alcohol drinking, diabetes, hypertension, or BMI between members of households with and without TB. Studies evaluating household-wide interventions for smoking to reduce its dual impact on TB and NCD may be warranted. Systematically screening for NCD using objective diagnostic methods is needed to understand the actual burden of NCD and inform comprehensive interventions.