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.

Impact of fluoroquinolone treatment on delay of tuberculosis diagnosis: A systematic review and meta-analysis.

BACKGROUND: Fluoroquinolones are among the most commonly used antibiotics for the treatment of respiratory infections. Because fluoroquinolones show bactericidal activity against Mycobacterium tuberculosis, there is concern that their use can delay the diagnosis of tuberculosis. We conducted a systematic review and meta-analysis to assess whether empiric treatment with fluoroquinolones delays the diagnosis and treatment of tuberculosis in patients with respiratory tract infections. OBJECTIVES: The primary objective was to assess the delay in days in the diagnosis and treatment of tuberculosis, among patients who received quinolones, compared to those who received non-fluoroquinolone antibiotics. METHODS: We included studies of adult patients treated with fluoroquinolones prior to a confirmed diagnosis of tuberculosis. We performed a literature search of 7 databases (including PubMed, Embase and Cochrane Library) with no language restrictions. We calculated an unweighted mean of estimate of difference in delay across all studies. For the studies for which the estimate was available as a mean with standard deviation, a weighted average using a random effects meta-analysis model was estimated. RESULTS: A total of 3983 citations were identified from the literature search; of these, 17 articles were selected for full-text review. A total of 10 studies were retained for the synthesis. These included 7 retrospective cohort studies and 3 case-control studies. Only one of these studies was from a high TB burden country, South Africa. The most commonly used fluoroquinolones were levofloxacin, gemifloxacin and moxifloxacin. The unweighted average of difference in delay between the fluoroquinolone group and non-fluoroquinolone group was 12.9 days (95% CI 6.1-19.7). When these differences were pooled using a random effects model, the weighted estimate was 10.9 days (95% CI 4.2-17.6). When stratified by acid-fast smear status, the delay was consistently greater in the smear-negative group. CONCLUSION: Although results are variable, the use of fluoroquinolones in patients with respiratory infections seems to delay the diagnosis of TB by nearly two weeks. Consistent with the International Standards for TB Care, their use should be avoided when tuberculosis is suspected.

A cross-metathesis approach to novel pantothenamide derivatives.

Pantothenamides are known for their in vitro antimicrobial activity. Our group has previously reported a new stereoselective route to access derivatives modified at the geminal dimethyl moiety. This route however fails in the addition of large substituents. Here we report a new synthetic route that exploits the known allyl derivative, allowing for the installation of larger groups via cross-metathesis. The method was applied in the synthesis of a new pantothenamide with improved stability in human blood.

High rates of obesity and greater associated disability among people with rheumatoid arthritis in Canada.

Obesity in rheumatoid arthritis has been associated with increased risk of comorbidities, larger medical costs, decreased quality of life, higher disease activity, and reduced therapeutic responses. We assessed the burden of obesity among rheumatoid arthritis patients and its impact on patient-reported outcomes. Patients receiving care at two Canadian University Centers were included. Height and weight were measured and selected sociodemographic and rheumatoid arthritis (RA) characteristics as well as patient-reported outcomes were obtained. Patients were classified according to WHO criteria and proposed RA cut points, and results were compared with national data. Using WHO criteria, 68 (34 %) RA patients were classified as obese (vs. ~25 % of Canadians). Using RA cut points, 112 (55 %) RA patients were classified as obese. With both classification methods, obese individuals had significantly higher mean HAQ scores and a higher odds of significant disability (HAQ ≥ 1: WHO OR 2.3; 95 % CI 1.2, 4.2 and RA-specific OR 1.8; 95 % CI 1.0, 3.2). Independent of the classification method use, RA patients have significantly higher rates of obesity than national prevalence estimates. Obese RA patients had about twice the odds of reporting moderate to severe disability.