AI Sniffs Earwax to Detect Parkinson’s with 94% Accuracy

An AI-powered olfactory system has demonstrated a groundbreaking ability to detect Parkinson’s disease (PD) with 94% accuracy by analyzing scent-related compounds in earwax. This method could become a low-cost and effective first-line screening tool, offering new hope for early intervention in the progressive neurological disorder.

Most treatments for Parkinson’s only slow its progression, making early diagnosis crucial for improving patient outcomes. Current testing methods, such as clinical evaluations and brain imaging, are costly and can be subjective. However, researchers have discovered that certain volatile organic compounds (VOCs) present in sebum — the oily substance produced by skin — change as PD advances.

Because sebum on exposed skin can be affected by environmental factors like air pollution, it’s less reliable for testing. In contrast, earwax — primarily composed of sebum — is shielded within the ear canal and can be collected easily. Researchers from China, including Hao Dong and Danhua Zhu, focused on this unique biological material for their PD screening research.

The team collected earwax samples from 209 individuals, including 108 diagnosed with PD, and analyzed the VOCs using gas chromatography and mass spectrometry. They identified four key compounds — ethylbenzene, 4-ethyltoluene, pentanal, and 2-pentadecyl-1,3-dioxolane — that significantly differed between PD patients and healthy individuals. These were flagged as potential biomarkers for Parkinson’s.

Using this data, the researchers developed an artificial intelligence olfactory (AIO) model capable of distinguishing between the samples with 94% accuracy. They suggest that the AIO system could be integrated into early screening protocols, potentially revolutionizing how PD is diagnosed and treated.

Lead researcher Hao Dong noted that while the current study is a small-scale trial conducted at a single center in China, further research is planned across multiple stages of the disease, geographic regions, and ethnic groups. These studies aim to validate the system’s broader practical application in clinical settings.