Classification of root canal microorganisms using electronic-nose and discriminant analysis

Area of Science:

• Microbiology
• Analytical Chemistry
• Biotechnology

Background:

• Root canal treatment involves debridement to remove microorganisms.
• Identifying root canal bacteria aids treatment decisions but is challenging due to anaerobic complexity.
• Current practices often rely on empirical treatments rather than routine microbial identification.

Purpose of the Study:

• To classify seven common root canal infection microorganism strains using electronic nose odor data.
• To determine optimal electronic nose and discriminant analysis settings for enhanced classification performance.

Main Methods:

• Utilized odor data from 7 microorganism strains at 4 concentration levels.
• Employed 3 discriminant analysis methods and evaluated sensor response normalization with 3 baseline values.
• Investigated 2 dimension reduction methods (standard deviation, principal component analysis) to optimize classification.

Main Results:

• Quadratic discriminant analysis showed superior performance.
• Classification accuracy decreased with lower microorganism concentrations.
• Optimal performance was achieved using a difference model with standard deviation-based dimension reduction or a normalized fractional difference model with PCA.

Conclusions:

• Electronic nose technology offers a promising and convenient method for classifying root canal infection microorganisms.
• The study identified optimal parameters for electronic nose application and discriminant analysis, enhancing its utility in endodontics.