South Korean Researchers Develop Nano-AI Platform That Detects Drug Use From a Drop of Urine in 5 Minutes

A device has been developed that can precisely analyze urine without any additional chemical treatment and determine within 5 minutes whether drugs have been taken. Once commercialized as a portable compact device, it is expected to be used for on-site drug enforcement by investigative agencies. The Korea Research Institute of Bioscience and Biotechnology (KRIBB) announced that Principal Researcher Eun-Kyung Lim of the BioNano Research Center, together with Professor Tae-Hwan Kim of Daegu Catholic University, Professor Jun-Hee Lee of Seoul National University Bundang Hospital, and the Stone Lab AI Research Center, has developed a “nano-AI fusion diagnostic platform” that precisely detects drug use even in urine mixed with other medications and waste products.

As drug-related crimes have surged worldwide in recent years, the need for rapid and precise diagnostic technologies that can quickly determine drug use on site has grown. Drug use is mainly tested by detecting drug components remaining in urine. 

While high-precision tests can provide accurate results, they have the drawback of requiring a long time to process samples. They also require specialized analytical equipment, making it difficult to obtain results immediately in the field. Rapid on-site test kits can deliver results quickly but have limitations in accuracy, such as cross-reacting incorrectly with other substances that have similar properties.

Molecules scatter light in their own unique patterns when illuminated. Scattering refers to the phenomenon in which light collides with particles and is dispersed in all directions. The research team developed a nano substrate that amplifies the unique scattering patterns emitted by drug component molecules by tens of thousands of times. 

 

The team built densely packed microscopic pillars, like a lawn, on a thin gold plate and roughened the surface. When drug molecules are placed on the substrate, their characteristic scattering patterns are amplified, allowing even trace amounts of drug molecules to be clearly distinguished. In addition, using the artificial intelligence (AI) model “LightGBM,” they were able to precisely differentiate only the scattering patterns generated by drug components from among the various chemicals present in urine. 

 

When tested using urine samples from 155 actual patients, the nano-AI fusion diagnostic platform combined with LightGBM detected 13 types of drug components in urine containing multiple medications such as painkillers, with a precision level of about 80%. In animal experiments on rats, the system detected drug components in urine collected 6 hours after drug administration.

 

This study is significant in that it combines the strengths of conventional large-scale precision analyzers and simple test kits to secure a core technology that can test for drugs on site using only a single drop of urine, without complex processing steps. 

 

Principal Researcher Lim said, “This study demonstrates that by combining nanotechnology with AI, we can accurately and rapidly classify drug components even in complex body fluids,” adding, “We expect that, once further advanced into a portable diagnostic device, it will transform the paradigm of drug enforcement and emergency diagnostics.”

 

The results were published in the September 24 issue of the international journal ACS Measurement Science.

Read the original article on dongascience.