Showing papers by "Costas Pitris published in 2013"
TL;DR: This technique can be applied directly to urine samples, and with the enhancement provided by SERS, this method has the potential to be developed into a rapid method for same day UTI diagnosis and antibiogram.
Abstract: The inappropriate use of antibiotics leads to antibiotic resistance, which is a major health care problem. The current method for determination of bacterial susceptibility to antibiotics requires overnight cultures. However most of the infections cannot wait for the results to receive treatment, so physicians administer general spectrum antibiotics. This results in ineffective treatments and aggravates the rising problem of antibiotic resistance. In this work, a rapid method for diagnosis and antibiogram for a bacterial infection was developed using Surface Enhanced Raman Spectroscopy (SERS) with silver nanoparticles. The advantages of this novel method include its rapidness and efficiency which will potentially allow doctors to prescribe the most appropriate antibiotic for an infection. SERS spectra of three species of gram negative bacteria, Escherichia coli, Proteus spp., and Klebsiella spp. were obtained after 0 and 4 hour exposure to the seven different antibiotics. Bacterial strains were diluted in order to reach the concentration of (2x10 5 cfu/ml), cells/ml which is equivalent to the minimum concentration found in urine samples from UTIs. Even though the concentration of bacteria was low, species classification was achieved with 94% accuracy using spectra obtained at 0 hours. Sensitivity or resistance to antibiotics was predicted with 81%-100% accuracy from spectra obtained after 4 hours of exposure to the different antibiotics. This technique can be applied directly to urine samples, and with the enhancement provided by SERS, this method has the potential to be developed into a rapid method for same day UTI diagnosis and antibiogram.
8 citations
TL;DR: A rapid method of diagnosis and antibiogram for a bacterial infection was developed using Surface Enhanced Raman Spectroscopy (SERS) with silver nanoparticles, which can be applied directly to urine or blood samples, bypassing the need for overnight cultures.
Abstract: Antibiotic resistance is a major health care problem mostly caused by the inappropriate use of antibiotics. At the root of
the problem lies the current method for determination of bacterial susceptibility to antibiotics which requires overnight
cultures. Physicians suspecting an infection usually prescribe an antibiotic without waiting for the results. This practice
aggravates the problem of bacterial resistance. In this work, a rapid method of diagnosis and antibiogram for a bacterial
infection was developed using Surface Enhanced Raman Spectroscopy (SERS) with silver nanoparticles. SERS spectra
of three species of gram negative bacteria, Escherichia coli, Proteus spp., and Klebsiella spp. were obtained after 0 and 4
hour exposure to the seven different antibiotics. Even though the concentration of bacteria was low (2x105 cfu/ml),
species classification was achieved with 94% accuracy using spectra obtained at 0 hours. Sensitivity or resistance to
antibiotics was predicted with 81%-100% accuracy from spectra obtained after 4 hours of exposure to the different
antibiotics. With the enhancement provided by SERS, the technique can be applied directly to urine or blood samples,
bypassing the need for overnight cultures. This technology can lead to the development of rapid methods of diagnosis
and antibiogram for a variety of bacterial infections.
4 citations
TL;DR: This paper proposes the use of Rank Order Kernels to classify bacterial samples, and shows that this method is comparable in accuracy to other methods which were used previously for the same data set.
Abstract: The range of applications of Raman-based classification has expanded significantly, including applications in
bacterial identification. In this paper, we propose the use of Rank Order Kernels to classify bacterial samples.
Rank Order Kernels are two-dimensional image functions which operate on two-dimensional images. The first
step in the classification therefore, is to transform the Raman spectra to two-dimensional images. This is achieved
by splitting the spectra into segments and calculating the ratio between the mean value of each and every other
segment. This creates a two-dimensional matrix of ratios for each Raman spectrum. A similarity metric based
on rank order kernels operating on the two-dimensional matrices is then used with a nearest neighbor algorithm
for classification. Our results show that this method is comparable in accuracy to other methods which were
used previously for the same data set.
2 citations
TL;DR: This paper proposes the use of Rank Order Kernels to classify Raman spectra in order to identify bacterial samples and shows that the rank order kernel method is comparable in accuracy to other previously-used methods.
Abstract: Bacterial identification is one of the applications for which classification using Raman spectra has proved to be successful. In this paper, we propose the use of Rank Order Kernels to classify Raman spectra in order to identify bacterial samples. Rank Order Kernels are two-dimensional image functions. The first step in the process transforms each Raman spectrum to a two-dimensional image. This is achieved by splitting the spectra into segments and calculating the ratio between the mean value of each and every other segment. The resulting two-dimensional matrix of ratios for each Raman spectrum is the image processed by the Rank Order Kernels. A similarity metric is used with a nearest neighbor algorithm for classification. The metric is based on rank order kernels. Our results show that the rank order kernel method is comparable in accuracy to other previously-used methods.
1 citations
12 May 2013
TL;DR: In this article, a method for improvement of lateral resolution in optical coherence tomography (OCT) is presented, where signals from all overlapped volumes are combined optimally to improve the resolution using all available cross correlations.
Abstract: A method for improvement of lateral resolution in Optical Coherence Tomography (OCT) is presented. The resolution improvement achieved with this method does not depend on the delivery optics. Moreover the depth of focus is not restricted. The method is based on the lateral oversampling of the image. The laterally oversampled signals are backscattered signals from shifted and overlapped resolution volumes. Signals from successive volumes are correlated due to the region shared by adjacent resolution volumes. By utilizing the cross correlation of signals from such overlapped volumes, resolution can be improved by various degrees depending on which pairs of signals are used. To maximize the resolution improvement for a given oversampling factor, signals from the farthest spaced and overlapped resolution volumes should be processed. The cost of the resolution refinement is the increasing statistical error because the magnitude of the cross-correlation function becomes smaller. In this method signals from all overlapped volumes are combined optimally to improve the resolution using all the available cross correlations. Preliminary results of such an approach on laterally oversampled OCT images have shown that it is possible to achieve a 3.7-fold lateral resolution improvement.
TL;DR: In this article, a new method for lateral resolution improvement of optical coherence tomography (OCT) images is presented, which is based on the lateral oversampling of the image.
Abstract: A new method for lateral resolution improvement of Optical Coherence Tomography (OCT) images is presented. The
improvement is independent of the delivery optics and the depth of focus. It is based on the lateral oversampling of the
image. In OCT, laterally oversampled signals are backscattered signals from shifted and overlapped resolution volumes.
In that way, signals from successive volumes are correlated due to the region shared by adjacent resolution volumes.
Resolution can be improved by utilizing the cross correlation of signals from such overlapped volumes. The resolution
can be improved by various degrees depending on which pairs of signals are used. In this method signals from all
overlapped volumes are combined optimally to improve the resolution using all the available cross correlations.
Preliminary results of such an approach on laterally oversampled OCT images have shown that it is possible to achieve a
3.7-fold lateral resolution improvement.