Back

Multicentre accuracy trial of FUJIFILM SILVAMP TB LAM test in people with HIV reveals lot variability

Szekely, R.; Sossen, B.; Mukoka, M.; Muyoyeta, M.; Nakabugo, E.; Hella, J.; Nguyen, H. V.; Ubolyam, S.; Chikamatsu, K.; Mace, A.; Vermeulen, M.; Centner, C. M.; Nyangu, S.; Sanjase, N.; Sasamalo, M.; Dinh, H. T.; Ngo, T. A.; Manosuthi, W.; Jirajariyavej, S.; Mitarai, S.; Nguyen, N. V.; Avihingsanon, A.; Reither, K.; Nakiyingi, L.; Kerkhoff, A. D.; MacPherson, P.; Meintjes, G.; Denkinger, C. M.; Ruhwald, M.; FujiLAM Study Consortium,

2022-09-09 infectious diseases
10.1101/2022.09.07.22278961 medRxiv
Show abstract

RationaleThere is an urgent need for rapid, non-sputum point-of-care diagnostics to detect tuberculosis. ObjectivesThis prospective trial in seven high tuberculosis burden countries set out to evaluate the diagnostic accuracy of the point-of-care urine-based lipoarabinomannan assay FUJIFILM SILVAMP TB LAM (FujiLAM) among inpatient and outpatient people living with HIV. MethodsDiagnostic performance of FujiLAM at point of care was assessed among adult people with HIV against a mycobacterial reference standard (sputum culture, blood culture, and Xpert Ultra from urine and sputum at enrollment, and additional sputum culture [≤]7 days from enrollment), an extended mycobacterial reference standard including available non-study test results, and a composite reference standard including clinical evaluation. Measurements and Main ResultsOf 1624 participants enrolled, 294 (18{middle dot}0%) were classified as TB positive by eMRS. Median age was 40 years, median CD4 cell count was 372 cells/ul, 52% were female and 78% were taking antiretroviral therapy at enrollment. Overall FujiLAM sensitivity was 54{middle dot}8% (95% CI: 49{middle dot}1-60{middle dot}4), and overall specificity was 85{middle dot}1% (83{middle dot}1-86{middle dot}9), against the extended mycobacterial reference standard. Sensitivity and specificity estimates varied between sites, ranging from 26{middle dot}5% (95% CI: 17{middle dot}4%-38{middle dot}0%) to 83{middle dot}3% (43{middle dot}6%-97{middle dot}0%), and 75{middle dot}0 (65{middle dot}0%-82{middle dot}9%) to 96{middle dot}5 (92{middle dot}1%-98{middle dot}5%), respectively. Post-hoc exploratory analysis identified significant variability in the performance of the six FujiLAM lots used in this study. ConclusionsLot variability limited interpretation of FujiLAM test performance. Although the results with the current version of FujiLAM are too variable for clinical decision-making, the lipoarabinomannan biomarker still holds promise for tuberculosis diagnostics.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.