New AI-powered biosensor provides instant cancer diagnoses with 99% accuracy

Cancer is often most dangerous when it hides in the early stages. Detecting it before symptoms appear is one of the biggest challenges in modern medicine. Now, researchers from the Korea Institute of Materials Science (KIMS) have developed a powerful new tool that could help tip the odds in favor of early diagnosis. Led by Dr. Ho Sang Jung, the team has created an advanced optical biosensor that can identify trace amounts of cancer-related DNA changes in a blood sample—using only light and artificial intelligence.

This sensor zeroes in on DNA methylation, a subtle chemical change that acts like a fingerprint of cancer. When cancer cells form, certain genes turn on or off due to the addition of a small chemical tag called a methyl group. These changes often silence genes that normally protect you from cancer, known as tumor suppressor genes. Spotting these DNA methylation patterns in the bloodstream offers a way to catch cancer early, even before tumors grow large enough to see.

Most current methods for analyzing DNA methylation rely on harsh chemical processes like sodium bisulfite conversion. While these techniques work, they are slow, complex, and often miss low levels of cancer DNA in early stages. Other newer methods—such as electrochemical and traditional optical biosensors—face problems too, like limited sensitivity and poor consistency.

To overcome these limits, Dr. Jung’s team combined several cutting-edge technologies into a single device. The core innovation lies in using light-enhancing materials called plasmonic nanoparticles. These particles can boost the optical signals of nearby molecules by over 100 million times, making even the smallest amounts of cancer DNA visible. The biosensor then uses artificial intelligence to sort through the light signals and find methylation patterns linked to cancer.

This design eliminates the need for chemical processing, making the system fast, clean, and low-cost. The device uses just 100 microliters of blood—less than a single drop—and delivers results in only 20 minutes.

The sensitivity of this sensor is extraordinary. It can detect DNA methylation at levels as low as 25 femtograms per milliliter—equivalent to dissolving one twenty-five-thousandth of a grain of sugar in a single drop of water. That’s about 1,000 times more sensitive than most standard biosensors.

To test its real-world performance, the team used the device to examine blood from 60 patients with colorectal cancer. The sensor detected cancer with 99% accuracy and even distinguished between early and late stages of the disease. This level of detail is crucial for both diagnosis and treatment planning.

At the heart of the biosensor is a breakthrough technique called plasmonic molecular entrapment, or PME. PME enables light signals to be tightly focused around target DNA molecules. Gold nanoparticles are grown directly around the DNA, forming “hotspots” where the light intensifies. This process creates a powerful signal that’s easy to read and hard to miss.

The light-based technique that powers this process is known as Surface Enhanced Raman Scattering (SERS). SERS works by shining light onto molecules and measuring how the light scatters. Each molecule scatters light in a unique way, like a fingerprint. For DNA methylation, even small structural changes can shift the scattering pattern. The PME method traps the DNA exactly where the light is strongest, creating a signal that reflects its methylation state.

This label-free system doesn’t require extra dyes or markers. Instead, it reads the natural vibrations of the DNA, enhanced by the gold particles. The team also designed a special porous gold nanosponge to act as a structure for forming these plasmonic hotspots, improving signal consistency and strength.

To make sense of the complex light data, the team trained an artificial intelligence model based on logistic regression. This machine learning tool analyzes patterns in the light signals and classifies whether a sample shows healthy or cancer-related methylation levels. In clinical tests, the system reached an accuracy of over 99%, with 100% sensitivity and 98.3% specificity.

These results suggest that PME-assisted SERS, combined with AI, can match or even outperform traditional methods. More importantly, it delivers faster results without the need for chemical preprocessing, making it suitable for hospitals, clinics, and even at-home testing.

Although this study focused on colorectal cancer, the technology could be applied to many other diseases. Since DNA methylation plays a role in autoimmune and neurological disorders as well, the biosensor has the potential to support a wide range of diagnostic tools.

Dr. Jung noted, “This technology serves as a next-generation diagnostic platform capable not only of early cancer detection, but also of predicting prognosis and monitoring treatment response.”

The ease of use and quick turnaround make it ideal for point-of-care testing. It can help personalize treatment plans, monitor disease progression, or even catch relapse after treatment. By eliminating the complexity and high cost of traditional lab tests, this approach could bring advanced molecular diagnostics into everyday healthcare.

While SERS has long been known for its sensitivity, one major challenge has been consistency. Many traditional SERS setups rely on random aggregation of nanoparticles, which can lead to unreliable signals. By building a structured gold nanosponge that localizes DNA and supports even signal enhancement, the KIMS team solved a problem that has held back clinical use of SERS for years.

Another difficulty in earlier systems was directing the DNA to the exact location of the light-enhancing hotspot. Previous methods used chemical ligands or electric fields to position the molecules, but these added steps made testing slower and more expensive. PME removes that step entirely by growing gold particles directly around the target DNA in its natural state, ensuring precise localization and better signal clarity.

The team hopes to scale the technology for broader clinical use. As a low-cost, label-free, and highly accurate method, it could be integrated into portable diagnostic devices or used in national cancer screening programs. Its ability to detect cancer early—when treatment is most effective—could save lives and reduce healthcare costs.

Looking forward, researchers also plan to expand its use to monitor treatment success and detect relapse. Because the system can track methylation changes over time, it may offer a powerful tool for long-term disease management.

Dr. Jung and his team continue refining the device, aiming to make it even more user-friendly and accessible. If successful, their innovation could mark a turning point in how diseases are diagnosed and managed worldwide.

Research findings are available online in the journal Advanced Science.

Note: The article above provided above by The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.