AI Saliva Sensor Enables Early Detection of Head and Neck Cancer

Now, researchers have developed an artificial intelligence (AI)–powered saliva sensor platform capable of identifying cancer-linked metabolic changes with 98% accuracy from a single saliva drop.

The system was created through a collaboration by Korea University (Seoul, South Korea) and Seoul St. Mary’s Hospital at the Catholic University of Korea (Seoul, South Korea), whose researchers combined expertise in biomedical engineering, spectroscopy, and clinical diagnostics to develop the platform. The technology works by detecting components of head and neck cancer within infiltrating metabolites found in saliva.

Researchers analyzed these metabolites with Raman spectroscopy, which identifies molecular structures based on how light scatters when it strikes a substance. AI algorithms then separated and classified the resulting mixed signals, enabling detection of cancer-specific signatures within saliva — a substance far more complex than blood. The platform leverages a specialized graphene-based nanoparticle structure in which gold particles grow along microscopic wrinkles, forming a coral-like surface. This configuration amplifies Raman signals and concentrates saliva components, further enhancing detection sensitivity.

In the study, saliva samples (described as acupuncture samples in the original text) from 50 individuals — including both healthy participants and patients with head and neck cancer — were analyzed. The platform presented 15 new biomarkers associated with the presence and progression of the disease. It distinguished cancer from non-cancer samples with 98% accuracy, demonstrating that metabolic changes linked to head and neck cancer can be reliably identified through a simple, non-invasive test.

The researchers note that the approach could evolve into a clinically deployable diagnostic tool usable in hospitals and screening centers. Potential applications extend beyond head and neck cancer, as saliva-based diagnostics are rapidly expanding globally. The platform’s ability to capture subtle metabolic changes suggests future possibilities for detecting various diseases and discovering new biomarkers across multiple clinical fields.

"This study is the world's first field-type platform to non-invasive early diagnosis of head and neck cancer using invasive metabolic changes," said Professor Ho-sang Jeong, corresponding author. "It can be used for early diagnosis of various diseases and discovery of new biomarkers in the future."

Related Links:
Korea University
Seoul St. Mary’s Hospital