A-276 Unlocking an ultra-rapid end-to-end ID and AST workflow from whole blood
Abstract Background The Advanced Microfluidic Device (AMD) is a new ultra-rapid sample preparation technology designed to extract and purify pathogens directly from whole blood to detect bloodstream infections (BSI) and sepsis in minutes. This breakthrough sample prep technology leverages advanced microfluidics and optics to automate processes including identification (ID), efficient buffer exchanges, and monitoring cell concentration to deliver precise amounts of high-purity pathogen populations optimized for downstream analytical techniques. The AMD is compatible with existing diagnostic platforms and analytical techniques. By significantly reducing diagnostic timelines, the AMD is positioned to provide healthcare professionals with clinically actionable information faster, enabling timely, targeted interventions and improving outcomes for high-morbidity bloodstream infections and potentially prevent escalation to sepsis. Methods Representative ESKAPE pathogens were directly spiked into freshly collected whole blood from healthy donors. The contrived blood samples were screened at a wide range of concentrations to demonstrate the robust capture efficiency of the system and determine the limit of detection (LoD). Contrived whole blood samples with clinically relevant ESKAPE pathogen concentrations were processed with mild chemicals while loaded into the AMD. Selectively isolated pathogens were imaged with Quantitative Phase Microscopy (QPM) and using machine learning (ML), we developed preliminary pathogen ID algorithms through key single-cell features such as: cell size, morphology, and dry mass. Cell density was also monitored to determine when to reverse fluid flow and collect the purified pathogens off-chip at clinically relevant concentrations for downstream processing. Recovered pathogens were used for downstream phenotypic antibiotic susceptibility testing (AST) including standard broth microdilution method and an FDA-approved automated AST platform. Results The AMD provides an end-to-end automated platform, from blood collection to ID and AST reports in hours, while keeping existing clinical microbiology laboratory instrumentation intact. Microfluidic chips trapped targeted ESKAPE pathogens input while eliminating all blood cells. Once captured and concentrated on the device, pathogens were fluidically relocated to the imaging module on-chip where QPM was used to characterize the cell size, morphology, and dry mass. The supporting Figure shows captured pathogens imaged with QPM and then segmented to demonstrate the key single-cell features unlocked by super-resolution microscopy. Samples collected from the AMD provided comparable downstream AST results to traditional blood culture methods while drastically improving the time-to-result to be within a same-shift timeline. Conclusion The AMD is a rapid, ultra-sensitive sample prep method to detect BSI and sepsis directly from whole blood. The AMD method enables downstream analytical techniques to operate straight from whole blood, removing the initial bottleneck of waiting on a positive blood culture that requires subsequent purification steps. Initial detection, combined with a preliminary ID, can be accomplished in critical timeframes that will save lives. This technology will be expanded to additional downstream analytical techniques such as PCR, FISH, next-gen sequencing, and beyond with additional pathogens such as Candida spp. and viruses as future expansion areas for the AMD technology. The seamless compatibility of the AMD utilizes the robust existing laboratory infrastructure present in hospitals to allow for maximum adoption by clinicians and meaningfully impact patient outcomes.
Authors
- María Luisa Gutiérrez García (ORCID: https://orcid.org/0000-0002-0035-9211)
- Youlian Goulev (ORCID: https://orcid.org/0000-0003-0370-4567)
- Paulo Garcia (ORCID: https://orcid.org/0000-0002-1041-5205)
- Eduardo L. Latouche
- Johan Paulsson (ORCID: https://orcid.org/0000-0002-8671-4914)
- Tom Feltrup
Institutions
- Harvard University (US)
- Boston Latin Academy (US)
Publication Details
- Journal
- Clinical Chemistry
- Published
- 2025-10-01
- DOI
- https://doi.org/10.1093/clinchem/hvaf086.265
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00