Barnabas Bakamutumaho, Kevin Baker, Matthew L Cavuto, Suleyman Sarp Pinar, Katarzyna M Szostak-Lipowicz, Winters Muttamba, Ingrid Ampeire, Allan Asiimwe, Brian Odaga, Ivana Pennisi, Katerina-Theresa Mantikas, Nicolas Moser, Elliot Quigley, Henry Kyobe Bosa, Misaki Wayengera, Andrew Obuku, Haruna Muwonge, Jane Achan, Bruce Kirenga, Alison H Holmes, Carlos Maluquer de Motes, David Ulaeto, Stella Neema, Jesus Rodriguez-Manzano, Wilber Sabiiti
Dragonfly showed high sensitivity and specificity for mpox detection across a broad range of viral loads in the field, supporting its potential as a point-of-care diagnostic in high-burden, resource-limited settings (eg, low-income and middle-income countries, remote environments, or small clinics). Further research should assess integration with existing diagnostics, cost-effectiveness, and performance across clades and key populations.
BACKGROUND: Mpox continues to spread across east and central Africa, with Uganda among the most affected countries. The diagnostic reference standard, centralised real-time quantitative PCR (qPCR), requires specialised infrastructure and sample-transport logistics, producing extended turnaround times that delay public health responses, particularly in remote or underserved settings. Point-of-care molecular diagnostics could address this, but prospective clinical evaluation data from affected regions remain limited.
METHODS: We conducted a prospective diagnostic accuracy study across six health facilities in Kampala and Wakiso, Uganda, within routine outpatient and inpatient pathways. Individuals of any age were consecutively enrolled if they presented at a participating site during the enrolment period with signs or symptoms meeting the WHO suspected-case definition for mpox and had at least one active cutaneous lesion amenable to swabbing. Cutaneous lesion swabs were tested at the point-of-care with Dragonfly, a sample-to-result molecular platform using a dual-target design to detect orthopoxvirus (OPXV) and monkeypox virus (MPXV), followed by confirmatory qPCR. The primary outcome was the diagnostic accuracy (sensitivity and specificity) of Dragonfly for MPXV and OPXV against qPCR. Usability and acceptability were assessed in a focus group discussion with front-line users. Clinical and epidemiological associations were examined among concordant participants using a prespecified, literature-informed binary feature set with Fisher's exact tests and Benjamini-Hochberg correction.
FINDINGS: Between Sept 17 and Nov 28, 2025, of 300 enrolled participants, 196 (65%) were positive for MPXV by qPCR (median cycle threshold 21·1 [IQR 19·1-23·9]). Among confirmed MPXV cases, 108 (55%) were male and 88 (45%) were female by self-report. For the primary outcome, Dragonfly showed 98·7% (95% CI 96·6-99·5) overall agreement with qPCR, with results available in under 40 min. For MPXV, sensitivity was 98·5% (95·6-99·5) and specificity was 96·2% (90·5-98·5); for OPXV, sensitivity was 100% (98·1-100) and specificity 96·2% (90·6-98·5). Front-line users reported high acceptability, attributing this to avoidance of centralised laboratory logistics, while noting training and supply-chain requirements for routine use.
INTERPRETATION: Dragonfly showed high sensitivity and specificity for mpox detection across a broad range of viral loads in the field, supporting its potential as a point-of-care diagnostic in high-burden, resource-limited settings (eg, low-income and middle-income countries, remote environments, or small clinics). Further research should assess integration with existing diagnostics, cost-effectiveness, and performance across clades and key populations.
FUNDING: UK Biotechnology and Biological Sciences Research Council, UK Medical Research Council, and Wellcome Trust funded Centres for Antimicrobial Optimisation Network programme.