Form with function: Rapid phenotypic carbapenemase testing to improve detection of carbapenemase-producing organisms

PROJECT ABSTRACT Infections due to carbapenemase-producing organisms (CPOs) are associated with high mortality rates due to extensive antibiotic resistance, and carbapenemase genes can be transferred between bacteria leading to out- breaks. Therefore, CPO identification has significant implications for antibiotic selection and use of infection control practices in healthcare facilities. No antibiotic susceptibility test profile accurately identifies CPO. Yet, existing methods for CPO detection, including phenotypic, molecular, and lateral flow assays, are not performed by most clinical microbiology laboratories due test complexity, prolonged workflows, subjective results, reagent instability, and high cost. Phenotypic tests have many key features necessary for a general CPO screen, yet logistical barriers still prevent their widespread adoption. The objective of the current proposal is to evaluate the potential impact of rapid CPO detection using a form-factor optimized phenotypic test paired with machine learn- ing algorithms to distinguish carbapenemase gene families at the time of GNB identification. Our central hypoth- esis is that a rapid, low-cost phenotypic test will increase CPO detection, decrease time-to-identification of a CPO infection, and facilitate timely initiation of effective antibiotics and isolation practices. Rapid CPO detection will be achieved with a phenotypic test optimized by our group (the mCNP) and ported into a structurally-pro- grammed sticker microfluidic to perform four discrete reactions in a single channel (the µCNP). These assays provide objective carbapenemase detection in ~20 minutes and will be used to achieve the objectives of this proposal, as laid out in three aims. 1) Distinguish carbapenemase class and gene family using machine learning. This aim is based on the capacity of machine learning to identify subtle changes over time in the colorimetric mCNP test and provide objective, automated results to the carbapenemase class and gene-family level. 2) Eval- uate rapid phenotypic detection of CPO in the µCNP, which will provide reproducible and stable carbapenemase detection in a low-cost, scalable sticker microfluidic that replicates mCNP performance. And 3) Compare rapid carbapenemase detection to standard-of-care testing by leveraging active protocols for collection of clinical cul- tures and ongoing studies of CPO carriage and healthcare environmental contamination. Rapid CPO testing will increase overall CPO detection and decrease time to identification from GNB-positive cultures and surveillance rectal swabs. Expected results include the demonstration of the utility of the µCNP as an accessible phenotypic carbapenemase test for appropriate antibiotic selection and rapid identification of CPO-colonized individuals for timely contact isolation. We envisage an important positive impact as our platform will provide equitable access to carbapenemase testing, improve patient outcomes, and prevent CPO outbreaks in healthcare networks.