Speaker
Description
Bacterial meningitis is a severe and potentially lethal infection of the meninges, afflicting more than a million people per year globally. Delays in treatment have been shown to increase both mortality rates and the strain on the healthcare system. In an effort to develop a near-instantaneous and clinically simple diagnostic test for bacterial meningitis, our group has shown that laser-induced breakdown spectroscopy (LIBS) has the capability to detect bacterial pathogens in various media to a high-degree of accuracy, including Escherichia coli and Mycobacterium smegmatis in artificial cerebrospinal fluid (aCSF) to simulate a meningitis infection. Ongoing work seeks to establish the current detectable limit of bacterial presence within aCSF and subsequently reduce the lower bound, as needed, to create a viable diagnostic tool.
In this work, the lower limit of detection of LIBS was determined by assaying several concentrations of E. coli in aCSF on a nitrocellulose medium. Optical densitometry measurements were performed to measure and fix concentrations of eight test suspensions. Dilutions ranging from 26 500 down to 180 colony forming units per laser shot were created. A 1064 nm Nd:YAG laser with 8 mJ pulse energy was focused to 100 µm spot size. Each laser shot created a microplasma and the resulting atomic emission was dispersed by a high-resolution Échelle spectrometer to produce a spectrum spanning 200 nm to 800 nm. An artificial neural network with principal component analysis preprocessing was constructed to identify the presence of bacterial cells in the spectra. In addition, a partial least squares discriminant analysis was performed on a model comprised of 15 line intensities of observed elements from the full spectrum. Both models were used to ascertain the lower limit of detection.
The sensitivities and specificities of each test will be presented. The lower limit of detection from these data will be analyzed and discussed. Future studies contributing to the reduction of this lower bound will be discussed, including dismembrating samples, introducing a double-centrifugation system, and investigating the dependence of this lower limit on the ablation laser wavelength.
| Keyword-1 | Laser spectroscopy |
|---|---|
| Keyword-2 | Bacteria |
| Keyword-3 | Machine learning |