Robin Sweeney's Abstracts
Robin Sweeney
Biomedical Engineering
Ph.D. Student
Conference Summary
SPIE Defense + Commercial Sensing: Commercial + Scientific Sensing and Imaging
Anaheim, California
Abstract
The Mie‐scatter based device designed through this work is able to use Mie scatter spectra to distinguish the presence or lack of a skin infection (despite the presence of commensal bacteria on skin) and, further, determine the species of bacteria responsible for the infection, allowing for rapid diagnosis and decreased time to treatment.
A 650 nm LED light source is used to illuminate the tissue sample from angles of 180°, 190°, 200°, 210°, and 225°. A custom 3D printed photodiode array is used to collect light scatter from the surface of a tissue sample at angles from 100° to 170° at 10° increments. Illumination of the tissue and collection of light scatter across a range of angles results in Mie‐scatter spectra. An Arduino microcontroller is used to collect light scatter intensity at each angle. Light scatter intensity trends across the collection angles determine the presence of an infection versus the presence of only commensal bacteria. Principle component analysis (PCA) of Mie scatter spectra is able to distinguish the species of bacteria responsible for an infection.
Escherichia coli and Staphylococcus aureus have been tested on porcine and human cadaver skin with this system, resulting in distinct differences in Mie scatter spectra between these species. The device has been tested with common skin contaminants (e.g. lotion), and does not distinguish a difference between lotion and control groups. Light scatter data used to determine the presence and species of an infection is collected without contact with tissue and within 3 seconds.
The Mie‐scatter based device designed has been shown to distinguish the presence of and bacterial species responsible for a skin infection despite the presence of commensal bacteria. The device uses an angular photodiode array and a 650 nm light source at varied angles to determine differences in Mie scatter spectra based on differences in bacterial species. Testing on a porcine skin model and a human cadaver skin model has shown that light scatter trends across detection angles can diagnose the presence of a skin infection and principle component analysis is able to distinguish the species of bacterial infection.
Lay Abstract
Through this work, a device has been designed and tested to diagnose a bacterial infection on the skin. The device uses light scatter to determine if an infection in present on the skin or not, meaning that the device is able to make this determination without touching a patient or skin sample and within just seconds. In addition to diagnosing the presence of a skin infection, the device is able to determine what kind of bacteria is on the surface of the skin, which is helpful for determining what treatment is best for an individual. Currently, a diagnosis of this kind can take days, as bacteria need to be grown to be identified, but with the technology developed here a diagnosis could be made in seconds without expensive laboratory equipment. Further, bacteria are always present on skin, which can cause difficulty in determining if bacteria found on the skin are healthy, normal bacteria, or the cause of an infection. The device designed here is able to detect the presence of an infection despite healthy bacteria always present on the skin by comparing a location on the skin that is possibly infected to one that is not likely infected. Overall, the device designed aims to greatly decrease the time and cost of diagnosing a bacterial infection on the skin.
The device designed through this work uses a single red LED that is pointed at the tissue sample (area of skin) from five different locations. Shining the red LED on the skin results in any bacteria present scattering the light away from the surface of the skin. Each kind of bacteria scatters this light in a unique manner. The scattered light is then collected by photodiodes, which detect light and turn it into voltage (electricity). In this device, photodiodes are located in a quarter circle around the skin sample, so they detect the scattered light from eight different locations. The amount of light detected at each of these eight locations gives unique pattern that is then used to identify the kind of bacteria on the skin surface. The light detected is analyzed through principle component analysis (PCA), which is a statistical method to find patterns in a large data set, to determine the kind of bacteria likely present to produce the scattered light patterns detected by photodiodes.
The device designed here has been tested on multiple skin models of both pig skin and human cadaver skin with two types of bacteria, E. coli and S. aureus (commonly referred to as a Staph infection). The device has also been tested with other contaminants on the skin models (body lotion in this case), as these contaminants can sometimes interfere with diagnostic testing. The device is able to determine if an infection is present and, if so, what bacteria is causing the infection, even with other contaminants on the skin, in a matter of seconds. The ability to diagnose a bacterial infection on the skin so quickly, without concern for what lotions a patient may have applied, and without even having to physically touch the potentially infected area allows for faster and cheaper diagnosis compared to current techniques and allows clinicians to determine an appropriate course of treatment nearly immediately.