Revolutionizing Point of Care Testing for Bacterial Detection Using Near Infrared Spectroscopy

by Christine Oh

Coauthors: Thank you Dr. Sharief Taraman, Dr. Anthony Chang, and Dr. Timothy Flannery for your guidance in the development of this abstract during the CHOC MI3 Internship.

Medical Devices & Digital Health

Streptococcal pharyngitis—or strep throat—is one of the most common types of sore throat in children. It is a bacterial infection caused by group A Streptococcus bacteria. While there are current ways to test for strep throat, such as the rapid antigen detection test (RADT) and a throat culture, these methods involve a throat swab, which many children find very uncomfortable. As a result, children often resist the examination, making the process even more stressful and traumatizing for them. When children are uncooperative, strep throat tests may be performed inaccurately, affecting the diagnosis. Overall, testing children for strep throat can be a very unpleasant experience for everyone.

StrepScan is a handheld device that uses Fourier-transform near infrared spectroscopy (FT-NIRS) and machine learning to scan patients’ oropharynx area for strep bacteria, avoiding the need for uncomfortable throat swabs. NIRS is a noninvasive technique that uses near infrared light to detect a sample’s spectral signature, which can be used to analyze the composition of a material. When a sample is exposed to near infrared light, its chemical bonds absorb certain amounts of the near infrared light, and as a result the reflected light has less intensity. Analyzing the variations in the reflected spectrum, also known as the spectral signature, helps determine the sample’s chemical composition. While NIRS is mainly used in healthcare for brain imaging and analysis, research has shown that it can also be used in the identification of different strains of bacteria. Thus, StrepScan uses NIRS to swiftly and painlessly detect strep bacteria in the throat: a great alternative to current strep throat tests.
As for the physical design of StrepScan, the device resembles a forehead thermometer, with the near infrared light coming out of the end. Although StrepScan scans mostly the back area of the mouth, the near infrared light is also capable of reaching the throat, which is where strep bacteria is mainly located. The goal is to implement StrepScan in clinical settings, where doctors and nurses can use this device to diagnose their patients.
Machine learning will be used to train the StrepScan device to recognize the spectral signature of group A Streptococcus bacteria. In the early development of the device, StrepScan will scan only cultures of strep bacteria so it can learn to recognize the spectral signature of strep bacteria. Over time, as the computer program becomes more advanced, StrepScan will be capable of distinguishing strep bacteria on its own. Although StrepScan’s main purpose is to detect strep bacteria, it will also be taught to identify many other pathogenic bacteria in the mouth, such as corynebacterium or Neisseria gonorrhoeae. It could also be used to detect bacteria on the skin or on wounds, which would be useful for diagnosing conditions like MRSA, which can take forty-eight hours to diagnose.

StrepScan provides a promising alternative to current uncomfortable strep throat tests; it uses near infrared spectroscopy and machine learning to scan the mouth for harmful bacteria quickly and painlessly. Not only will StrepScan be useful in identifying strep bacteria, but it will also screen for many other potentially harmful bacteria in the mouth or on the skin. With its easy use and noninvasive technology, StrepScan will make doctors’ jobs easier and patient satisfaction higher.