AI Diagnostic Tool Could Deliver Results in Two Minutes Using Fingertip Sample

Biofluids such as synovial fluid, blood plasma, and saliva contain proteins that are an important biomarker for the diagnosis of several health conditions. Now, a specially designed platform has been programmed to detect the concentration of these proteins to assist in diagnosis and monitoring disease progression. The research proposes that hospital waiting times could be drastically cut and the option for self-screening and self-monitoring is now possible with the potential for at-home diagnostic kits in the future.

Scientists at Swansea University (Swansea, UK) developing a platform that would use artificial intelligence (AI) to speed up the process of detecting biomarkers in biofluids have shown that the concept could work. It would mean faster test results for health conditions such as cardiovascular disorders, joint quality, and Alzheimer’s disease. The new diagnostic tool could revolutionize the healthcare sector due to the application of a form of AI – machine learning (ML). The implementation of ML has meant it is possible, for the first time, for results to be delivered within minutes.

Image: The μ-rheometer evaluates liquid properties such as viscosity using fingerpick samples (Photo courtesy of Swansea University)

“The key innovation is the fact of providing a result within two minutes, which is a leap forward compared to standard testing that can take several hours,” said Dr. Francesco Del Giudice, project lead. "What this means for the future is that our proof-of-concept study can be further developed in a tool to help clinicians making decisions on clinical data obtained quickly. We also foresee to develop this further for an at-home-point-of-care self-screening diagnostic platform.”

“The ability of artificial intelligence to drive down the time required to complete various tasks has been demonstrated across a number of disciplines,” added Dr. Claire Barnes, co-author on the work. “The advantage of speed offered by the implementation of machine learning allowed us to adjust almost in real-time the experimental parameters to fulfill the requirements of the theoretical model associated with this work.”

Related Links:
Swansea University