Professor Jeong Ho-sang of the School of Biomedical Engineering at Korea University and Professor Park Jun-uk of the Department of Otorhinolaryngology at Seoul St. Mary's Hospital of the Catholic University of Korea said on the 4th that their team, in joint research with the Korea institute of Materials Science (KIMS), developed a sensor platform that uses artificial intelligence (AI) to diagnose head and neck cancer early with saliva.
Head and neck cancer is a cancer that occurs in areas of the head and neck, such as the mouth, pharynx, and larynx, and is often found late because early symptoms are rare. In particular, because biopsies that collect part of the tissue and endoscopic access are difficult, the need for noninvasive technology that does not require surgery or incision is increasing.
To that end, the research team analyzed metabolites in saliva and combined them with an AI technique that separates data in which multiple signals are mixed to distinguish each component. Through this, they newly proposed 15 biomarkers, indicators that show the presence and progression of disease, providing important clues not only for diagnosing disease but also for understanding the mechanism and process of disease occurrence.
The core technology that enabled this analysis is the "Plasmonic Nanocoral" structure. As gold particles grow at fine wrinkles and defect sites of graphene, a carbon material, a coral-shaped nanoparticle structure is formed. This structure strongly amplifies light and simultaneously collects volatile substances contained in saliva effectively. Using these properties, the research team completed a high-sensitivity sensor that can detect metabolites in saliva precisely and stably.
Using this sensor, the research team analyzed saliva samples from 50 people, including patients with head and neck cancer and healthy individuals, and confirmed that the accuracy of distinguishing the patient group from the normal group reached 98%. This demonstrated that metabolic changes in head and neck cancer can be precisely analyzed using only saliva. It shows the potential to evolve into a point-of-care diagnostic technology that can be deployed quickly in hospitals or on site.
Jeong said, "This study is the world's first case of presenting a point-of-care platform that can noninvasively diagnose head and neck cancer early using metabolic changes in saliva," and noted, "This technology could be used for the early diagnosis of various diseases and the discovery of new biomarkers in the future."
The results were published online on Oct. 7 (local time) in the international journal "Advanced Science."
References
Advanced Science (2025), DOI: https://doi.org/10.1002/advs.202517710