Researchers Develop Tear-based Covid-19 Biosensor Using AI

2024-10-02
(0)
Researchers Develop Tear-based Covid-19 Biosensor Using AI

A team of Korean researchers has developed a novel biosensor platform that can detect Covid-19 through human tears, combining deep learning with surface-enhanced Raman scattering (SERS).

SERS is a technique that enhances the detection of trace amounts of substances, including chemical and biological molecules, through surface plasmon resonance on metal surfaces like silver and gold. The process amplifies molecular signals by over 100 million times, making it vital for precise diagnostics and characterization of diverse substances.

The researchers' approach, published in the June issue of ACS Publications, uses 3-D gold nanoparticles (AuNPs) to deliver quick, non-invasive testing.

A common method for fabricating gold nanoparticle structures involves a two-step process: first, synthesizing AuNPs using polyvinylpyrrolidone, and second, assembling them into a film through the Langmuir-Blodgett technique. While effective, this approach requires additional time, effort, and specialized equipment for both the nano-synthesis and self-assembly processes.

However, Han Ji-sang, professor of ophthalmology at Kangbuk Samsung Hospital; Moon Sang-woong and Choi Sam-jin, professors of ophthalmology at Kyung Hee University College of Medicine; and Dr. Jung Ho-sang, a specialist in nano-bio convergence at the Korea Institute of Materials Science, have developed a faster, more efficient method. They proposed a single-step fabrication process for creating large-area, three-dimensional AuNP structures, eliminating the need for additional steps.

In their study, the team used molybdenum disulfide as a catalyst, enabling gold nanoparticles to spontaneously form into three-dimensional structures. This occurs as electrons generated on the semiconductor surface reduce gold ions, balancing the Fermi levels between molybdenum disulfide and the gold solution. When polyvinylpyrrolidone is added, it helps form a wide, two-dimensional layer of molybdenum disulfide, which in turn facilitates the creation of densely packed, layered AuNP structures, much like electroless plating.

The entire process is guided by a deep learning model. Leveraging a convolutional neural network, the model achieves “outstanding surface signaling performance at subterascale levels,” all while operating under low-power irradiation and millisecond-level acquisition times, the researchers said. This method delivers rapid and precise detection of Covid-19 susceptibility by analyzing minute chemical changes in tear samples, achieving a 98.5 percent sensitivity in under two minutes, according to the study.

“This technology has shown the potential to quickly and accurately assess the presence of infection in real time with a non-invasive method using human tears,” said Professor Han. He added that the findings indicate its applicability for not only infectious diseases but also for various conditions requiring rapid and precise diagnosis.

 

Read the original article on Korea Biomedical Review (KBR).

 

COMMENTS

RELOAD
CAPTCHA