Screening for Alzheimer's Disease Using Saliva: A New Approach Based on Machine Learning and Raman Hyperspectroscopy

Area of Science:

• Biomedical Engineering
• Spectroscopy
• Machine Learning

Background:

• Alzheimer's disease and related dementias (ADRDs) are increasing globally, posing a significant public health challenge.
• AD is a leading cause of death, with early diagnosis critical for effective management.
• Current diagnostic methods for ADRDs are often inefficient, costly, and fail to detect the disease in its nascent stages.

Purpose of the Study:

• To develop a novel, non-invasive diagnostic method for Alzheimer's disease (AD).
• To leverage Raman hyperspectroscopy and machine learning for analyzing saliva samples for AD diagnosis.

Main Methods:

• Saliva samples were collected from individuals with normal cognition, mild cognitive impairment (MCI), and AD.
• Raman hyperspectroscopy was employed to analyze the biochemical composition of the saliva samples.
• Genetic Algorithm and Artificial Neural Networks were utilized for spectral data analysis and diagnostic algorithm development.

Main Results:

• The diagnostic algorithm achieved 99% accuracy in internal cross-validation for distinguishing between the three groups.
• Independent external validation demonstrated 100% accuracy in diagnosing AD and MCI from saliva samples.
• The study highlights the efficacy of machine learning in interpreting complex spectral data.

Conclusions:

• Raman hyperspectroscopy of saliva presents a promising non-invasive approach for AD diagnosis.
• This method offers high efficiency and accuracy, potentially revolutionizing early ADRD detection.
• Further research can validate this technique for widespread clinical application.