HMN 2025: What is the biosensor based mostly on gentle and AI for early analysis of cancer

Instant cancer diagnosis with light and AI!
Illustration of the high-sensitivity detection of hint quantities of cancer DNA in blood and its evaluation utilizing an AI-based methodology. Credit: Korea Institute of Materials Science (KIMS)

Dr. Ho Sang Jung and his analysis workforce from the Advanced Bio and Healthcare Materials Research Division on the Korea Institute of Materials Science (KIMS) have developed an optical biosensor able to detecting hint quantities of cancer cell DNA within the bloodstream with excessive sensitivity, enabling early cancer analysis.

The findings are published within the journal Advanced Science.

Compared to traditional liquid biopsy strategies, this provides superior sensitivity and specificity, and its standout function lies in its skill to diagnose cancer early utilizing solely and AI-based evaluation, with out the necessity for complicated testing procedures.

When develop, refined chemical modifications happen on the floor of DNA within the bloodstream—a phenomenon referred to as DNA methylation.

In the early phases of cancer, the focus of methylated DNA is extraordinarily low, making it tough to detect with typical biosensors at excessive sensitivity.

To overcome this problem, the analysis workforce developed a biosensor materials know-how that allows the detection of methylated DNA utilizing high-sensitivity optical alerts and synthetic intelligence evaluation, with out requiring extra processing or complicated evaluation steps.

The analysis workforce built-in high-sensitivity optical signaling and synthetic intelligence evaluation with plasmonic supplies. These supplies can amplify the optical alerts of DNA molecules by greater than 100 million occasions in response to gentle, enabling the detection of even extraordinarily small quantities of DNA.

As a end result, the biosensor is able to figuring out methylated DNA at concentrations as little as 25 fg/mL (femtograms per milliliter) within the early phases of cancer growth. To put this in perspective, 25 fg/mL is equal to dissolving 1/25,000th of a sugar grain in a single drop of water. This represents a 1,000-fold enchancment in sensitivity in comparison with typical biosensors.

The analysis workforce utilized the developed biosensor to from 60 colorectal cancer sufferers, reaching a 99% accuracy charge in detecting the presence of cancer. Moreover, the biosensor efficiently distinguished between cancer phases from Stage I to Stage IV with excessive precision.

The evaluation required solely 100 ?L (microliters) of blood and was accomplished inside 20 minutes, demonstrating a extremely easy and speedy diagnostic course of.

This know-how permits the high-sensitivity detection of hint quantities of cancer DNA with none pre-processing, making it extremely aggressive to be used within the early cancer analysis market.

Compared to traditional gear, it considerably reduces each evaluation time and price, permitting for potential software in hospitals, well being screening facilities, at-home diagnostic kits, and moveable diagnostic units. It shouldn’t be solely appropriate for precision medication, but additionally extremely efficient for Point-of-Care Testing (POCT) functions.

Dr. Ho Sang Jung, Senior Researcher at KIMS and lead of the mission, said, “This know-how serves as a next-generation diagnostic platform succesful not solely of early cancer detection, but additionally of predicting prognosis and monitoring therapy response.

“We plan to develop its software to a variety of ailments, together with autoimmune problems and neurological circumstances.”

More data:
Muhammad Shalahuddin Al Ja’farawy et al, Plasmonic Molecular Entrapment for Label?Free Methylated DNA Detection and Machine?Learning Assisted Quantification, Advanced Science (2025). DOI: 10.1002/advs.202503257

Citation:
Researchers develop biosensor based mostly on gentle and AI for early analysis of cancer ( 10)
13 July 2025
biosensor-based-ai-early-diagnosis.html

The content material is supplied for data functions solely.